Labyrinth Weirs: December 2010

See discussions, stats, and author proﬁles for this publication at: https://www.researchgate.
net/publication/272489237
Labyrinth Weirs
Thesis · December 2010
CITATIONS READS
0 2,130
1 author:
Brian Crookston
Utah State University
66 PUBLICATIONS 154 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
Dam Safety View project
Hydraulic Structures View project
All content following this page was uploaded by Brian Crookston on 19 February 2015.
The user has requested enhancement of the downloaded ﬁle.

DigitalCommons@USU
All Graduate Theses and Dissertations Graduate Studies, School of
12-1-2010
Labyrinth Weirs
Brian Mark Crookston
Recommended Citation
Crookston, Brian Mark, "Labyrinth Weirs" (2010). All Graduate Theses and Dissertations. Paper 802.
http://digitalcommons.usu.edu/etd/802
This Dissertation is brought to you for free and open access by the
Graduate Studies, School of at DigitalCommons@USU. It has been
accepted for inclusion in All Graduate Theses and Dissertations by an
authorized administrator of DigitalCommons@USU. For more
information, please contact becky.thoms@usu.edu.
LABYRINTH WEIRS
by
A dissertation submitted in partial fulfillment

of the requirements for the degree
of
DOCTOR OF PHILOSOPHY
in
Civil and Environmental Engineering
Approved:
_________________________ _________________________
Blake P. Tullis Michael C. Johnson
Major Professor Committee Member
_________________________ _________________________
Mac McKee Gary P. Merkley
Committee Member Committee Member
_________________________ _________________________
Steven L. Barfuss Byron R. Burnham
Committee Member Dean of Graduate Studies
UTAH STATE UNIVERSITY

Logan, Utah
2010
ii
Copyright © Brian Mark Crookston
All Rights Reserved

iii
ABSTRACT
Labyrinth Weirs
by
Brian Mark Crookston, Doctor of Philosophy
Utah State University, 2010
Major Professor: Dr. Blake P. Tullis

Department: Civil and Environmental Engineering
Labyrinth weirs are often a favorable design option to regulate upstream water
elevations and increase flow capacity; nevertheless, it can be difficult to engineer an
optimal design due to the complex flow characteristics and the many geometric design
variables of labyrinth weirs. This study was conducted to improve labyrinth weir design
and analyses techniques using physical-model-based data sets from this and previous
studies and by compiling published design methodologies and labyrinth weir information.
A method for the hydraulic design and analyses of labyrinth weirs is presented.
Discharge coefficient data for quarter-round and half-round labyrinth weirs are offered
for 6° ≤ sidewall angles ≤ 35°. Cycle efficiency is also introduced to aid in sidewall
angle selection. Parameters and hydraulic conditions that affect flow performance are
discussed. The validity of this method is presented by comparing predicted results to
data from previously published labyrinth weir studies.
A standard geometric design layout for arced labyrinth weirs is presented.
Insights and comparisons in hydraulic performance of half-round, trapezoidal, 6° and 12°

iv
sidewall angles, labyrinth weir spillways located in a reservoir with the following
orientations are presented: Normal, Inverse, Projecting, Flush, Rounded Inlet, and Arced
cycle configuration. Discharge coefficients and rating curves as a function of HT/P are
offered. Finally, approaching flow conditions and geometric similitude are discussed;
hydraulic design tools are recommended to be used in conjunction with the hydraulic
design and analysis method.
Nappe aeration conditions for trapezoidal labyrinth weirs on a horizontal apron
with quarter- and half-round crests (6° ≤ sidewall angle ≤ 35°) are presented as a design
tool. This includes specified HT/P ranges, associated hydraulic behaviors, and nappe
instability phenomena. The effects of artificial aeration (a vented nappe) and aeration
devices (vents and nappe breakers) on discharge capacity are also presented. Nappe
interference for labyrinth weirs is defined; the effects of nappe interference on the
discharge capacity of a labyrinth weir cycle are discussed, including the parameterization
of nappe interference regions to be used in labyrinth weir design. Finally, the
applicability of techniques developed for quantifying nappe interference of sharp-crested
corner weirs is examined.
(222 pages)
v
ACKNOWLEDGMENTS
This study was funded by the State of Utah; the support and efforts of Everett
Taylor and Mac McKee were indispensable, thank you both. I am also grateful for the
financial support of the Utah Water Research Laboratory, Utah State University, and the
United States Society on Dams (USSD).
I would like to express sincere gratitude to my advisor and friend, Dr. Blake P.
Tullis, for his instruction, support, encouragement, and guidance.
I also wish to thank the other members of my supervisory committee, Dr. Michael
C. Johnson, Dr. Mac McKee, Dr. Gary P. Merkley, and Steven L. Barfuss, for their
support, expertise, direction, and advice.
I wish to thank those at the Utah Water Research Laboratory who provided help
and assistance in a variety of ways: Alan Taylor, the “shop guys,” Zac Sharp (manager of
the Hydraulics Lab), and especially Ricky Anderson (fabrication and installation of the
weirs).
Most of all, I am thankful for the support of my loving wife, Kelsi, our children,
and to God for the desire and ability to perform and complete this work.

vi
CONTENTS
Page
ABSTRACT....................................................................................................................... iii
ACKNOWLEDGMENTS ...................................................................................................v
LIST OF TABLES ...............................................................................................................x
LIST OF FIGURES .......................................................................................................... xii
NOMENCLATURE ........................................................................................................ xix
CHAPTER
1. INTRODUCTION .......................................................................................1
Background and Motivation ............................................................1

Research Objectives .........................................................................3
Organization .....................................................................................4
2. BACKGROUND AND LITERATURE ......................................................5
Labyrinth Weirs ...............................................................................5

Labyrinth Weir Modeling ................................................................6
Analytical Approach ............................................................6

Derivation of Linear Weir Equation ....................................8
Similarity Relationships .....................................................10
Single Sample Uncertainty ................................................12
Labyrinth Weir Parameters ................................................13
Headwater Ratio (HT/P) .........................................13

Cycle Width Ratio (w/P)........................................13
Relative Thickness Ratio (P/tw) .............................14
Radius of Curvature (HT/R)....................................15
Sidewall Angle (α) / Magnification Ratio (M).......15
Apex Ratio (A/w) ...................................................16
Efficacy (ε) .............................................................16
Cycle Efficiency (ε’) ..............................................18
Crest Shape ............................................................18
Nappe Interference .................................................20
vii
Nappe Aeration ......................................................24
Cycle Configuration, Weir Orientation, and
Weir Placement ....................................................26
Labyrinth Weir Design Methods ...................................................27
Early Investigations ...........................................................27

Taylor (1968) and Hay and Taylor (1970) .........................28
Darvas (1971).....................................................................30
Hinchliff and Houston (1984) ............................................30
Lux and Hinchliff (1985) and Lux (1984, 1989) ...............31
Magalhães and Lorena (1989) ...........................................31
Tullis, Amanian, and Waldron (1995) ...............................32
Melo, Ramos, and Magalhães (2002) ................................33
Tullis, Young, and Chandler (2007) ..................................34
Emiroglu, Kava, and Agaccioglu (2010) ...........................34
Labyrinth Weir Case Studies .........................................................35

Labyrinth Weir Research Studies ..................................................37
Amanian M.S. Thesis (1987) .............................................37

Waldron M.S. Thesis (1994)..............................................39
Willmore M.S. Thesis (2004) ............................................41
Lopes, Matos, and Melo (2006, 2008) ...............................42
3. EXPERIMENTAL SETUP AND TESTING PROCEDURE....................44
Test Facilities .................................................................................44

Experimental Setup ........................................................................46
Rectangular Flume Facility ................................................46

Reservoir Facility ...............................................................47
Instrumentation ..................................................................49
Physical Models .............................................................................52
Materials, Fabrication, and Installation .............................52

Model Configurations ........................................................56
Test Procedure ...............................................................................57
4. HYDRAULIC DESIGN AND ANALYSIS

OF LABYRINTH WEIRS .........................................................................67
Abstract ..........................................................................................67
viii
Introduction ....................................................................................67
Flow Characteristics...........................................................69
Previous Studies .................................................................69
Experimental Method.....................................................................72
Experimental Results .....................................................................75
Discharge Rating Curves ...................................................75

Nappe Aeration Behavior and Stability .............................78
Nappe Ventilation ..............................................................81
Labyrinth Design and Analyses .........................................83
Data Verification ................................................................87
Summary and Conclusions ............................................................92
5. ARCED AND LINEAR LABYRINTH WEIRS

IN A RESERVOIR APPLICATION .........................................................95
Abstract ..........................................................................................95
Introduction ....................................................................................95
Previous Design Methods ..................................................99

Labyrinth Weirs Located in a Reservoir ..........................100
Experimental Method...................................................................102
Experimental Results ...................................................................104
Geometric Layout of Arced Labyrinth Weirs ..................104

Hydraulic Performance ....................................................105
Labyrinth Weir Orientation, Placement, and
Cycle Configuration .......................................................108
Flow Characteristics.........................................................113
Geometric Similitude Considerations for Arced
Labyrinth Weirs ...............................................................118
Design Example ...............................................................119
Summary and Conclusions ..........................................................120
6. NAPPE INTERFERENCE AND NAPPE INSTABILITY .....................123
Abstract ........................................................................................123
Introduction ..................................................................................123
Experimental Method...................................................................131
Experimental Results ...................................................................134
ix
Nappe Aeration Conditions..............................................134
Nappe Instability ..............................................................139
Artificial Nappe Aeration ................................................140
Nappe Interference ...........................................................140
Nappe Interference for Labyrinth Weirs ..............140

Application of Published Techniques
for Nappe Interference .......................................144
Summary and Conclusions ..........................................................152
7. SUMMARY AND CONCLUSIONS ......................................................156
Synopsis ...................................................................................................156
Chapter 2 – Background and Literature ...................................................156
Chapter 3 – Experimental Setup and Testing Procedures ........................156
Chapter 4 – Hydraulic Design and Analysis of Labyrinth Weirs ............157
Chapter 5 – Arced and Linear Labyrinth Weirs in a Reservoir
Application .............................................................................................159
Chapter 6 – Nappe Aeration, Nappe Instability, and Nappe
Interference for Labyrinth Weirs ...........................................................160
REFERENCES ................................................................................................................163
APPENDICES .................................................................................................................168
Appendix A: Schematics of Tested Labyrinth Weir Physical

Models in the Rectangular Flume Facility .............................................169
Appendix B: Schematics of Tested Labyrinth Weir Physical
Models in the Reservoir Facility ............................................................175
Appendix C: Visual Basic Code, Specific to Rectangular Flume
Facility, Used in Microsoft Excel ..........................................................182
Appendix D: Visual Basic Code, Specific to Reservoir Facility,
Used in Microsoft Excel ........................................................................188
CURRICULUM VITAE ..................................................................................................195

x
LIST OF TABLES
Table Page
2-1 Summary of labyrinth weir parameters from design methods ...............................29
2-2 Labyrinth weirs from across the globe...................................................................38
2-3 Summary of physical models tested by Amanian (1987) ......................................39
2-4 Summary of physical models tested by Waldron (1994) .......................................40
2-5 Summary of physical models tested by Willmore (2004) .....................................42
3-1 Physical models tested ...........................................................................................59
4-1 Physical model test program ..................................................................................73
4-2 Curve-fit coefficients for quarter-round labyrinth and linear weirs.......................78
4-3 Curve-fit coefficients for half-round labyrinth and linear weirs ............................78
4-4 Nappe aeration conditions and corresponding ranges of

HT/P for labyrinth weirs .........................................................................................81
4-5 Unstable nappe operation conditions for labyrinth weirs ......................................82
4-6 Recommended design procedure for labyrinth weirs ............................................84
4-7 Representative single sample uncertainties for the tested labyrinth

and linear weirs ......................................................................................................89
4-8 Comparison between the proposed design method and results obtained
from hydraulic model tests for labyrinth weir prototypes .....................................92
5-1 Labyrinth weir design methods ..............................................................................99
5-2 Physical model test program ................................................................................103
5-3 Trend line coefficients for half-round trapezoidal labyrinth weirs,

valid for 0.05 ≤ HT/P ≤ 0.2...................................................................................109
5-4 Trend line coefficients for half-round trapezoidal labyrinth weirs,

valid for 0.2 ≤ HT/P ≤ 0.7.....................................................................................109
xi
5-5 Cd representative single sample uncertainties for labyrinth
Weirs tested in this study, HT/P ≥ 0.05 ................................................................110
5-6 Discharge comparisons for arced projecting labyrinth weirs and

arced projecting linear weirs ................................................................................113
5-7 Predicted Cd to confirm calculated results ...........................................................120
6-1 Physical model test program ................................................................................132

xii
LIST OF FIGURES
Figure Page
1-1 Brazos dam, Texas, USA .........................................................................................2
2-1 General classifications of labyrinth weirs: triangular (A), trapezoidal (B),

and rectangular (C) ..................................................................................................6
2-2 Labyrinth weir geometric parameters ......................................................................7
2-3 Schematic for derivation of a standard weir equation .............................................9
2-4 Rounded apexes of Brazos Dam, Texas, USA ......................................................17
2-5 Efficacy (ε) vs. α for quarter-round trapezoidal labyrinth weirs, data set
from this study .......................................................................................................18
2-6 Flow efficiency (ε’) of half-round trapezoidal labyrinth weirs; data set
from Willmore (2004) ............................................................................................19
2-7 Examples of six weir crest shapes .........................................................................20
2-8 Actual and effective weir height ............................................................................21
2-9 Nappe interference occurring for trapezoidal, 15° quarter-round labyrinth

weir at HT/P = 0.20 ................................................................................................21
2-10 Nappe interference and cycle number....................................................................22
2-11 Nappe interference region and parameters as defined by Indlekofer and

Rouvé (1975) for sharp-crested corner weirs.........................................................23
2-12 Linear and arced (fully projecting) labyrinth weir cycle configurations, flush,
partially projecting, normal, and inverse orientations ...........................................27
2-13 Graphical solution for labyrinth weir submergence, modified from

Tullis et al. (2007) ..................................................................................................35
2-14 Full-width model of Lake Brazos labyrinth spillway in Waco, Texas, USA ........36
3-1 Outside view of the Utah Water Research laboratory main building and
the primary hydraulics testing bay within ..............................................................44
xiii
3-2 Rectangular flume test facility ...............................................................................45
3-3 Reservoir test facility .............................................................................................45
3-4 Schematic of the rectangular flume test facility ....................................................47
3-5 Schematic of the reservoir test facility...................................................................48
3-6 Supply piping to the rectangular flume test facility ...............................................49
3-7 4-in and 8-in supply piping and orifice plates for the reservoir test facility ..........50
3-8 20-in supply piping and orifice plate for the reservoir test facility........................50
3-9 Flow measurement equipment (power supply, pressure transducer,

data logger, and Hart communicator) ....................................................................51
3-10 Stilling well used for the rectangular flume facility ..............................................52
3-11 Stilling well used for the reservoir test facility ......................................................53
3-12 Carriage and point gauge system in the rectangular flume ....................................53
3-13 Straight and hooked point gauges for nappe profiling ...........................................54
3-14 Velocity field mapping with 2-D acoustic Doppler velocimeter ...........................54
3-15 Flow pattern and direction observations with the dye wand..................................55
3-16 The joint between the ½ apex (to be attached to flume wall) and the weir
sidewall of the 2-cycle, 6° half-round labyrinth weir ............................................57
3-17 Physical model cycle configurations, weir orientations and placements ...............58
3-18 Aeration tube apparatus for N = 2 (A) and nappe breakers located on the
downstream apex (B) and on the sidewall (C) .......................................................59
3-19 Example schematic of standardized layout for arced labyrinth weirs ...................60
3-20 Comparison of Cd for trapezoidal quarter-round labyrinth weirs

(based upon Lc) ......................................................................................................61
3-21 Tailwater submergence for the 10° half-round trapezoidal labyrinth weir ............63
3-22 Local submergence at an upstream labyrinth weir apex ........................................64

xiv
3-23 Standing wave in the downstream cycle ................................................................65
3-24 Physical representation of Bint in plan-view (A) and (C) and profile view (B) and
(D) for nappe interference regions, including reference grid ................................66
4-1 Labyrinth weir schematic including geometric parameters ...................................68
4-2 Aeration tube apparatus .........................................................................................74
4-3 Cd vs. HT/P for quarter-round trapezoidal labyrinth weirs ....................................76
4-4 Cd vs. HT/P for half-round trapezoidal labyrinth weirs..........................................77
4-5 Comparison of half-round and quarter-round crest shape on

labyrinth weir hydraulic performance ....................................................................79
4-6 Nappe flow conditions: clinging (A), aerated (B), partially aerated (C),
and drowned (D) ....................................................................................................80
4-7 Wedge-shaped nappe breakers placed on the downstream apex ...........................83
4-8 Cycle efficiency vs. HT/P for quarter-round labyrinth weirs .................................86
4-9 Cycle efficiency vs. HT/P for half-round labyrinth weirs ......................................87
4-10 Recommended procedure for labyrinth weir analyses ...........................................88
4-11 Dimensionless submerged head relationship for labyrinth weirs

(based on Tullis et al. 2007) ..................................................................................89
4-12 Comparison between Cd values obtained by Willmore (2004) and the

present study for non-vented, half-round labyrinth weirs ......................................90
4-13 Comparison between proposed Cd design curves by Tullis et al. (1995),

Willmore (2004), and the present study for non-vented,
quarter-round labyrinth weirs (based upon Lc) ......................................................91
5-1 Example of a labyrinth weir spillway ....................................................................96
5-2 Summary schematic of tested labyrinth weir orientations .....................................98
5-3 Standard geometric layout for an arced labyrinth weir ........................................104
5-4 Cd vs. HT/P for α = 6° half-round trapezoidal labyrinth weirs .............................106

xv
5-5 Cd vs. HT/P for α = 12° half-round trapezoidal labyrinth weirs ...........................107
5-6 Comparison of labyrinth weir orientations for α = 6°..........................................110
5-7 Comparison of labyrinth weir orientations for α = 12°........................................111
5-8 Example of flow passing from O1 to I1 and O5 to I4 .........................................112
5-9 Examples of surface turbulence (A) and (B), and local submergence (C) ..........115
5-10 A labyrinth weir with the Flush orientation (A) and a Rounded Inlet (B) ...........116
5-11 A 5-cycle trapezoidal labyrinth weir, Projecting, α = 6° at

HT/P = 0.604 (A) and α = 12° HT/P = 0.595 (B)..................................................116
5-12 A 5-cycle trapezoidal labyrinth weir, α = 12°, θ = 10° at

HT/P = 0.200 (A) and HT/P = 0.400 (B) ..............................................................117
5-13 A 5-cycle trapezoidal labyrinth weir, α = 12°, θ = 30° at

HT/P = 0.203 (A) and HT/P = 0.400 (B) .............................................................117
5-14 Two geometrically similar arced labyrinth weir spillways, N = 5 (A)

and a non-geometrically similar design at ½ scale, equivalent
crest length, and N = 10 (B) HT/P = 0.203 (A) and HT/P = 0.400 (B) .................118
6-1 Example of a labyrinth weir .................................................................................124
6-2 Collision of nappes from adjacent sidewalls and the apex ..................................126
6-3 The effects of nappe interference: standing waves (A), wakes and
air bulking (B), and local submergence (C) .........................................................127
6-4 Example of nappe interference regions for an aerated nappe at low HT/P ..........129
6-5 Nappe interference region and parameters as defined by Indlekofer and

Rouvé (1975) for sharp-crested corner weirs.......................................................129
6-6 Aeration tube apparatus for N = 2 (A) and nappe breakers located on the
downstream apex (B) and on the sidewall (C) .....................................................133
6-7 Clinging nappe aeration condition observed for trapezoidal labyrinth weir,
half-round crest shape, α = 12°, HT/P = 0.196 .....................................................134
6-8 Aerated nappe aeration condition observed for trapezoidal labyrinth weir,
quarter-round crest shape, α = 12°, HT/P = 0.202 ................................................135
xvi
6-9 Partially aerated nappe aeration condition observed for trapezoidal labyrinth
weir, half-round crest shape, α = 12°, HT/P = 0.296 ............................................135
6-10 Drowned nappe aeration condition observed for trapezoidal labyrinth weir,
quarter-round crest shape, α = 12°, HT/P = 0.604 ................................................136
6-11 Nappe aeration and instability conditions for labyrinth weirs with a
quarter-round crest ...............................................................................................137
6-12 Nappe aeration and instability conditions for labyrinth weirs with a
half-round crest ....................................................................................................138
6-13 Physical representation of Bint in plan-view (A) and profile view (B) for
nappe interference regions ...................................................................................142
6-14 Bint for quarter-round trapezoidal labyrinth weirs, 6°≤ α ≤ 35° ...........................143
6-15 Bint for half-round trapezoidal labyrinth weirs, 6°≤ α ≤ 35° ................................144
6-16 Bint/B specific to quarter-round trapezoidal labyrinth weirs tested in

this study (6°≤ α ≤ 35°) ........................................................................................145
6-17 Bint/B specific to quarter-round trapezoidal labyrinth weirs tested in

this study (6°≤ α ≤ 35°) ........................................................................................146
6-18 LD as a function of HT for quarter-round labyrinth weirs ....................................147
6-19 LD as a function of HT for half-round labyrinth weirs..........................................148
6-20 LD/HT as a function of α .......................................................................................149
6-21 LD-Falvey/LD vs. HT/P for quarter-round labyrinth weirs ........................................150
6-22 LD-Falvey/LD vs. HT/P for half-round labyrinth weirs .............................................150
6-23 Bd/Bint vs. HT/P for quarter-round labyrinth weirs ...............................................152
6-24 Bd/Bint vs. HT/P for half-round labyrinth weirs ....................................................153
A-1 Schematic of 2-cycle, trapezoidal 6° quarter- and half-round

labyrinth weirs, normal orientation ......................................................................170
A-2 Schematic of 2-cycle, trapezoidal 6° half-round labyrinth weir,

inverse orientation ................................................................................................170
xvii



A-7 Schematic of 4-cycle, trapezoidal 15° quarter-round labyrinth weirs,

normal orientation ................................................................................................173


B-1 Schematic of 5-cycle, trapezoidal 6° half-round labyrinth weir,

projecting orientation, linear cycle configuration (θ = 0°) ..................................176

projecting orientation, arced cycle configuration (θ = 10°) .................................176



flush orientation, linear cycle configuration ........................................................178

rounded inlet orientation, linear cycle configuration ...........................................178

projecting orientation, linear cycle configuration (θ = 0°) ..................................179

xviii


flush orientation, linear cycle configuration ........................................................181

rounded inlet orientation, linear cycle configuration ...........................................181
xix
NOMENCLATURE
A Inside apex width or Area
Ar Area scaling ratio
a Acceleration
ar Acceleration scaling ratio
A/w Apex ratio
Ac Apex center-line width
α Sidewall angle
α’ Upstream labyrinth weir sidewall angle
B Length of labyrinth weir (Apron) in flow direction
Bd Length of labyrinth weir in flow direction within the area disturbed by

nappe interference. Calculated from Ld developed by Indlekofer and
Rouvé (1975)
Bint Measured interference length in flow direction within the area

disturbed by nappe interference for labyrinth weir – specific to
physical models tested in this study
Bint-proto Bint scaled to a prototype structure
β Orientation angle of a linear labyrinth weir cycle configuration to the

approaching flow
Cair-avg Average air concentration
Cd Discharge coefficient, data from current study
Cd(α°) Discharge coefficient for labyrinth weir of sidewall angle α
Cd(90°) Discharge coefficient for linear weir
Cd-channel Discharge coefficient specific to a labyrinth weir located in a channel
Cd-corner Discharge coefficient for corner weir

xx
Cd-Darvas Darvas (1971) discharge coefficient
Cd-Lux Lux (1984, 1989) discharge coefficient
Cd-m Average discharge coefficient for the disturbed area of a corner weir
Cd-M&L Magalhães and Lorena (1989) discharge coefficient
Cd-res Discharge coefficient for a labyrinth weir spillway located in a

reservoir
Cd-side Discharge coefficient for a labyrinth weir located in a side-channel

application
Cd-sub Discharge coefficient for a labyrinth weir with tailwater submergence
Cd-Tullis Tullis et al. (1995) discharge coefficient
Cd-Waldron Waldron (1994) discharge coefficient
D Outside apex width
E Effectiveness (%)
EV Bulk modulus of elasticity
ε Efficacy
ε’ Cycle efficiency
FE Elastic force that acts on a fluid particle
Fg Gravitational force that acts on a fluid particle
FI Inertial force that acts on a fluid particle
Fp Pressure force that acts on a fluid particle
Fr Froude number
Fr-m Froude number specific to a model
Fr-p Froude number specific to a prototype
Fσ Surface tension force that acts on a fluid particle

xxi
Fν Viscous force that acts on a fluid particle
g Acceleration constant of gravity
γ Specific weight of water
H Design flow water surface elevation
h Depth of flow over the weir crest
Hapron Approach channel elevation
Hcrest Elevation of labyrinth weir crest
Hd Total head downstream of a labyrinth weir
Hd/HT Downstream/upstream ratio of total unsubmerged head
hm The head upstream of the weir as defined by Indlekofer and Rouvé

(1975) that is comprised of a specific upstream depth and two velocity
components
Hresidual Relative residual energy at the base of a labyrinth weir
HT Unsubmerged total upstream head on weir
HT/P Headwater ratio
HT/Rcrest Radius of curvature
HT-max Maximum total head
H* Submerged total upstream head on a labyrinth weir
H*/HT Submerged head discharge ratio
I(#) Inlet cycle number (e.g., I1, I2…I5)
k Apex constant
kθ-CW Converging sidewall adjustment constant
L Characteristic length or characteristic weir length
Lc Total centerline length of labyrinth weir

xxii
lc Centerline length of weir sidewall
Lc(α°) Total centerline length of labyrinth weir for a specific α
Lc-cycle Centerline length for a single labyrinth weir cycle
LD A theoretical disturbance length where Q and Cd = 0
LD-Falvey LD values calculated from Eq. (6-4) [proposed by Falvey (2003)]
Ld The length of the crest within the disturbed area as defined by

Indlekofer and Rouvé (1975)
Le Total effective length of labyrinth weir
Lm Length of model
Lp Length of prototype
Lr Length scaling ratio
M Magnification ratio (Lc-cycle/w) or Mass
Mr Mass scaling ratio
µ Dynamic viscosity
N Number of labyrinth weir cycles
ν Kinematic viscosity
O(#) Outlet cycle number (e.g., O1, O2…O5)
P Weir height
p Gauge pressure, point location denoted with subscript (e.g., p1, p2)
Peffective Effective increase in weir height caused by sharp or flat crested weirs
Pproto Weir height of a labyrinth weir prototype
P/tw Relative thickness ratio
Q Discharge over weir

xxiii
Qcycle Discharge over a single labyrinth weir cycle
Qdesign Design discharge over labyrinth weir
QLab Discharge over a labyrinth weir
QLin Discharge over a linear weir
Q/N Average labyrinth weir cycle discharge
R Arc radius for an arced labyrinth weir
r Arc center to channel width midpoint distance for an arced labyrinth

weir
r’ Segment height for an arced labyrinth weir
Rabutment Radius of rounded abutment wall
Rcrest Radius of rounded crest (e.g., Rcrest = tw / 2)
Re Reynolds number
Re-m Reynolds number specific to a model
Re-p Reynolds number specific to a prototype
ρ Density of water
ρr Density of water scaling ratio
S Submergence level
Sbed Longitudinal slope of rectangular flume floor
σ Surface tension
t Time
tm Time specific to a model
tp Time specific to a prototype
tr Time scaling ratio

xxiv
tw Thickness of weir wall
Θ Central weir arc angle for an arced labyrinth weir
θ Cycle arc angle for an arced labyrinth weir
θCW Converging channel wall angle
U Local flow velocity
V Average cross-sectional flow velocity upstream of weir
v Flow velocity, location denoted with subscript (e.g., v1, v2)
vm Flow velocity specific to a model
vp Flow velocity specific to a prototype
vr Flow velocity scaling ratio
V Volume
Vr Volume scaling ratio
W Width of channel
w Width of a single labyrinth weir cycle
W’ Width of the arced labyrinth weir spillway
w’ Cycle width for the arced labyrinth weir spillway
w/P Cycle width ratio
wCd Single sample uncertainty of Cd
We Weber number
wHT Single sample uncertainty of HT
wLc Single sample uncertainty of Lc
wQ Single sample uncertainty of Q
y Flow depth
xxv
y90 90% local air concentration characteristic flow depth
ymin Minimum local air concentration characteristic flow depth
ymax Maximum local air concentration characteristic flow depth
z Elevation above an arbitrary datum, location denoted with subscript

(e.g., z1, z2)
CHAPTER 1
INTRODUCTION
Background and Motivation
Water management and conveyance are a critical component of human
civilization. As infrastructure ages and development continues, the need for hydraulic
structures continues. With regards to spillways, many are found to require rehabilitation
or replacement due to a greater emphasis placed on dam safety and from revised and
increased probable maximum flood flows. Weirs are a common and useful hydraulic
structure for a wide range of applications (e.g., canals, ponds, rivers, reservoirs, and
others). Many existing spillways utilize a type of weir as the flow control structure.
The flow capacity of a weir is largely governed by the weir length and crest
shape. A labyrinth weir (see Fig. 1-1) is a linear weir folded in plan-view; these
structures offer several advantages when compared to linear weir structures. Labyrinth
weirs provide an increase in crest length for a given channel width, thereby increasing
flow capacity for a given upstream head. As a result of the increased flow capacity, these
weirs require less free board in the upstream reservoir than linear weirs, which facilitates
flood routing and increases reservoir storage capacity under base flow conditions (weir
height may be increased). In addition to spillways, labyrinth weirs are also effective drop
structures, energy dissipaters, and flow aeration control structures (Wormleaton and
Soufiani 1998; Wormleaton and Tsang 2000).
Labyrinth weirs are often a favorable design option to regulate upstream water
elevations and increase flow capacity (e.g., spillways); nevertheless, it can be difficult to
2
Fig. 1-1. Brazos Dam, Texas, USA
engineer an optimal design for a specific location because there are limited design data
for the many geometric design variables. The objective of this research was to improve
labyrinth weir design and analyses techniques using physical-model based data sets from
this and previous studies, and by compiling published design methodologies and weir
information.
There is a large amount of information that has been published on labyrinth weirs.
There are a number of studies that present a hydraulic design method or design curves
(e.g., Hay and Taylor 1970; Darvas 1971; Lux 1984, 1989; Lux and Hinchliff 1985;
Magalhães and Lorena 1989; and Tullis et al. 1995) and unique insights have been gained
from case studies [e.g., Avon Spillway (Darvas 1971), Brazos Spillway (Tullis and
3
Young, 2005), Dog River Dam (Savage et al. 2004), Hyrum Dam (Houston 1982), Lake
Townsend Dam (Tullis and Crookston 2008), Prado Spillway (Copeland and Fletcher
2000), Standley Lake (Tullis 1993), Weatherford Reservoir (Tullis 1992), and Ute Dam
(Houston 1982)] where physical models were used to design prototype labyrinth weirs.
However, after conducting a thorough review of literature and discussing this topic with
experts, the author identified numerous aspects of labyrinth weir behavior and design that
needed additional research.
Research Objectives
The main objectives of this dissertation were to:
• Provide a hydraulic design method for quarter-round and half-round labyrinth
weirs and include information regarding the following orientations: Normal,
Inverse, Flush, Rounded Inlet, and Projecting.
• Provide geometric and hydraulic design information regarding arced labyrinth
weir configurations (non-linear cycle configuration), including a standardized and
simple geometric design methodology.
• Present design information for nappe aeration conditions, nappe instability, and
artificial aeration with respect to labyrinth weir geometry and flow conditions.
• Examine the concept of nappe interference and how it influences the discharge
capacity of labyrinth weirs. The location and size of nappe interference regions
are to be quantified, including a discussion of observed hydraulic conditions
within this region.

4
• Provide a comprehensive review of published literature for labyrinth weirs, which
will document how an understanding of labyrinth weir hydraulics has evolved
over time.
• Provide an accurate and comprehensive list of geometric and hydraulic labyrinth
weir nomenclature and terminology. This includes new terms and definitions
developed in this study and a refinement of previously accepted nomenclature.
• Provide detailed documentation of the experimental setup, methods, and
procedures used in this study.
• Clearly present and make readily available the results of this study so that they
may be used in engineering practice.
Organization
This dissertation follows the multi-paper format, which means that the results are
written as stand-alone papers intended for publication in peer-reviewed journals
(Chapters 4-6). Because of the page limits associated with peer-reviewed journals,
additional chapters were added (Chapters 2 and 3) to allow for a more complete
accounting of the findings of this study. The final chapter of this dissertation (Chapter 7)
contains summaries of Chapters 2-3 and the contributions and conclusions of Chapters 4-
6.
5
CHAPTER 2
BACKGROUND AND LITERATURE
Labyrinth Weirs
A weir is a simple device that has been used for centuries to regulate discharge
and upstream water depths and to measure flow rates. Weirs have been implemented in
streams, canals, rivers, ponds, and reservoirs. There are many weir geometries and,
therefore, types of weirs; a labyrinth weir is a linear weir that is folded in plan-view.
This is done to increase the length of the weir relative to the channel or spillway width,
thereby increasing the flow capacity of the structure over a linear weir for a given driving
head. Other similar weirs or specific labyrinth-type weir designs are: skewed or oblique
weirs (Kabiri-Samani 2010; Noori and Chilmeran 2005), duck-bill weirs (Khatsuria et al.
1988), piano-key weirs (Ribeiro et al. 2007; Laugier 2007; Lempérière and Ouamane
2003), and fuse gates (developed by HydroPlus®, Falvey and Treille 1995).
There are an infinite number of possible geometric configurations of labyrinth
weirs; however, there are three general classifications based upon cycle shape: triangular,
trapezoidal, and rectangular (Fig. 2-1). Triangular and trapezoidal shaped labyrinth
cycles are more efficient than rectangular labyrinth weir cycles, based on a discharge per
unit length comparison. The geometric parameters associated with labyrinth weir
geometry are presented in Fig. 2-2.
Labyrinth weirs have been of interest to engineers and researchers for many years
because of their hydraulic behavior. A labyrinth weir provides an increase in crest length
for a given channel width, thereby increasing flow capacity for a given upstream water
6
(A) (B) (C)
Fig. 2-1. General classifications of labyrinth weirs:

Triangular (A), trapezoidal (B), and rectangular (C)
elevation. Therefore, labyrinth weirs maintain a more constant upstream depth and
require less free board than linear weirs. For example, a labyrinth spillway can satisfy
increased flood routing requirements and increase reservoir storage under base flow
conditions, relative to a linear weir structure, such as an ogee-crest spillway. In addition
to flow control structures, labyrinth weirs have also been found to be effective flow
aeration control structures, energy dissipaters, and drop structures.
Labyrinth Weir Modeling
Analytical Approach
Flow passing over a labyrinth weir is difficult to accurately describe
mathematically. Because the flow passing over a labyrinth weir is three-dimensional and
passes through a critical-flow section, a mathematical derivation must take into account:
energy, momentum, continuity, non-parallel streamlines, pressure under the nappe, the
7
Fig. 2-2. Labyrinth weir geometric parameters
dynamics of the air cavity behind the nappe (including the absence of one), nappe
interference or colliding nappe flows, local submergence, surface tension effects,

8
viscosity effects, weir geometry, and crest shape. Consequently, researchers typically
apply a weir discharge equation with empirically determined coefficients, which are
determined from experimental results obtained from physical modeling.
Derivation of Linear Weir Equation
Eq. (2-1) is a general equation for linear weirs, and was adopted by Tullis et al.
(1995) for labyrinth weirs.
2 32
Q = Cd L 2 g H T (2-1)
3
In Eq. (2-1), Q is the discharge over the weir, Cd is a dimensionless discharge
coefficient, L is a characteristic length (e.g., crest length), g is the acceleration due to
gravity, and HT is the total head on the crest. This equation is derived by assuming:
steady one-dimensional flow, an ideal fluid (non-compressible, non-viscous, no surface
tension, etc.), atmospheric pressures behind the nappe, assumes hydrostatic pressures,
and horizontal and parallel stream lines at the crest. With these assumptions, the energy
equation [Eq. (2-2)] and the continuity equation [Eq. (2-3)] are as follows:
2 2
p1 v p v
+ 1 + z1 = 2 + 2 + z 2 (2-2)
γ 2g γ 2g
H
Q = v1 y1 = ∫ v 2 Ldh (2-3)
0
In Eqs. (2-2) and (2-3), p is the gage pressure, γ is the unit weight of water, v is a
velocity, z is the elevation above an arbitrary datum, y is a depth, and h is the depth from
the streamline to the water surface. The subscripts refer to a point location along a
common streamline (see Fig. 2-3).

9
Fig. 2-3. Schematic for derivation of a standard weir equation
Applying Eq. (2-2) from point 1 to point 2 results in Eq. (2-4); simplifying and
rearranging to solve for v2 yields Eq. (2-5).
2 2
v v
h+P+ 1 =h+P− y+ 2 (2-4)
2g 2g
 v1 
2
v2 = 2 g  y +  (2-5)
 2 g 
Substituting Eq. (2-5) into Eq. (2-3) and integrating yields Eq. (2-6).
 3 3

2   v 
2 2
 v12  2 
Q = L 2 g   h + 1  −    (2-6)
3  2g   2g  
 
Due to the assumptions made in the derivation (ideal fluid, horizontal nappe flow,
etc.), a discharge coefficient is added to Eq. (2-6) to correct the flow rate to match
10
experimental results. Also, a slight simplification is commonly made, which results in
Eq. (2-1). Note that HT refers to h+V2/2g and V = v1.
Similarity Relationships
When a model is used to obtain information to predict the performance and
behavior of a prototype, the rules of similitude must be followed. Hydraulic Modeling:
Concepts and Practice (ASCE 2000) was referenced for the following discussion.
The first requirement of similitude is that the model be a scaled geometric replica
of the prototype. Geometric scaling of length (L), area (A), and volume (V) are presented
in Eqs. (2-7), (2-8), and (2-9), respectively. The subscripts r, m, and p denote scaling
ratio, model, and prototype, respectively.
Lp
Lr = (2-7)
Lm
2
Ar = Lr (2-8)
3
Vr = L r (2-9)
Kinematic similitude requires the scaling of velocity (v) and acceleration (a) at
corresponding points in the model and prototype. Eqs. (2-10) – (2-12) present the scaling
ratios for time (t), velocity (v), and acceleration (a).
tp
tr = (2-10)
tm
Lr
vr = (2-11)
tr
Lr vr
ar = = (2-12)
tr
2
tr
11
Dynamic similitude maintains a constant ratio of forces, which, for example, can
include inertia, pressure, gravity, friction, and surface tension. The mass (M) scaling
ratio, and six forces that act on a fluid particle are presented in Eqs. (2-13) – (2-19).
3
M r = ρ rVr = ρ r Lr (2-13)
FI = ρVa = ρL2V 2 Inertia (2-14)
F g = γV (2-15)
Gravity
Fp = pA Pressure (2-16)
Fσ = σL (2-17)
Surface Tension
dv
Fν = µ A Viscous (2-18)
dy
FE = EV A Elastic (2-19)
ρ is the density of the fluid (water), σ is surface tension, µ is the dynamic
viscosity, dv/dA is the velocity gradient, and EV is bulk modulus of elasticity.
From these relationships and the Buckingham Π-theorem, common dimensionless
parameters have been established for satisfying a condition of similitude in physical
modeling. Dimensionless parameters (commonly referred to as Π numbers or Π terms)
that are relevant to free-surface flows over labyrinth weirs (HT is used as the
characteristic length, L) are presented in Eqs. (2-20) – (2-22).
v
Fr = Froude number (2-20)
gL
vL
Re = Reynolds number (2-21)
ν
12
ρv L
2
We = = Weber number (2-22)
σ
In Eq. (2-21), ν is kinematic viscosity. It is not possible to use two dimensionless
parameters (e.g., Froude number and Reynolds number) simultaneously to scale a model
(e.g., if Fr-m = Fr-p then Re-m ≠ Re-p for a given location). However, the geometric scale
of a physical model may be determined to minimize the effects of a particular force. For
example, ASCE (2000) states that the effects of surface tension on spillways are
negligible for We ≥ 100. A physical model of a spillway could use Fr similitude, yet the
geometric scale should be sufficiently large so that We ≥ 100, making the effects of
surface tension negligible. However, in general We and Re limits are not well established
or understood for all Fr similitude applications.
Single Sample Uncertainty
The percent uncertainty for each calculated discharge coefficient (wCd) was
calculated following the procedure outlined by Kline and McClintock, (1953). After
determining individual parameter uncertainties and taking partial derivatives, Eq. (2-23)
was used to determine the uncertainty of Cd; the % difference is presented as wCd/Cd. The
VB code was used in an excel macro, for which the details are not shown here, and is
presented in Appendices C and D for the rectangular flume and reservoir facilities,
respectively.
1
  wQ  2  − wL 2
  − 27 wHT 
2
2
wCd =    + c
 +    2-23
  Q   Lc   8H T  
 
13
Labyrinth Weir Parameters
Published research studies have developed numerous design parameters to aid
engineers in the optimization and design of labyrinth weirs. The following section
discusses the influence each parameter has on the discharge capacity of a labyrinth weir,
including any key studies or conclusions found in published literature.
There are instances where researchers use different names to refer to a parameter,
or a parameter is given an obscure or misleading name; it is anticipated that the following
discussion will clarify and improve parameter designations.
Headwater Ratio (HT/P). The headwater ratio is the total head (HT = h+V2/2g),
measured relative to the weir crest elevation, immediately upstream of the weir over the
weir height (P). It is dimensionless and is commonly used on the abscissa of a plot that
presents the hydraulic performance of a labyrinth weir. However, a limitation associated
with HT/P becomes apparent when plotting data from two labyrinth weirs that have
identical discharge rating curves, but are of different P.
Several researchers have recommended an upper limit of HT/P for labyrinth weirs
(Hay and Taylor 1970; Lux 1989) based upon declining hydraulic efficiency noted in
their experimental results. However, the upper limit of 0.9 presented by Tullis et al.
(1995) is solely based upon the limit of the experimental results. Although labyrinth
weirs are typically design for HT/P ≤ 0.9, engineers may be interested in the hydraulic
performance of these weirs at higher headwater ratios.
Cycle Width Ratio (w/P). The cycle width ratio (previously referred to as the
vertical aspect ratio) was considered by Taylor (1968) to influence nappe interference.
He recommended that w/P should be greater than 2.0. Design recommendations were
14
also made by Tullis et al. (1995) (3.0 ≤ w/P ≤ 4.0), Magalhães and Lorena (1989) (w/P ≥
2.5) and Lux (1989) (w/P ≥ 2.0). Furthermore, Lux found from his experiments that the
discharge coefficient decreased as w/P decreased. To correlate these findings with Cd,
w/P was incorporated into the discharge equation Lux proposed for triangular and
trapezoidal labyrinth weirs.
Neither w nor P is a dominant influence in nappe interference or discharge. A
disturbance length that accurately describes the crest length affected by colliding nappes
would be a more direct parameter to evaluate nappe interference. Also, the influence of
P on Cd is directly linked with the tailwater elevation and additional geometric
parameters of the weir (R, tw, Lc-cycle/w or α).
Relative Thickness Ratio (P/tw). In practice, the minimum required wall thickness
would be determined from a structural design and analysis of the weir walls. Hydraulic
guidance has been given based upon the geometries of the physical models tested. For
example, Tullis et al. (1995) presents P/tw = 6, models tested by Willmore (2004)
correspond to P/tw = 8. However, Lake Townsend Labyrinth Spillway (Tullis and
Crookston 2008) was constructed with P/tw = 13.3. P/tw was previously designated as the
sidewall thickness ratio.
At the laboratory scale, sharp-crested weirs of varying P/tw have similar values of
Cd for a given HT/P. However, half-round and quarter-round crests have different Cd
values for corresponding HT/P. At low heads this may be due to scale effects and may be
more appropriately described by the Radius of Curvature (HT/Rcrest). Thus, additional
research is needed to quantify the influence of P/tw.

15
Radius of Curvature (HT/Rcrest). The discharge coefficient Cd is influenced by
HT/Rcrest. Matthews (1963) studied the effects of curvature on weirs with a round-crest
and concluded that weirs with a small radius of curvature would have a larger Cd than
weirs with a large radius of curvature, at a given head. A discharge rating curve for half-
rounded weirs was presented as HT/Rcrest vs. Cd by Rouvé and Indlekofer (1974).
Currently, Cd values provided with labyrinth weir design methods for round-crested (e.g.,
quarter-round, half-round, Ogee, WES or truncated Ogee, etc.) weirs include the effects
of HT/Rcrest inherent with the physical models tested; however, no method for labyrinth
weirs was found in published literature to adjust Cd for larger or smaller values of Rcrest.
The flow pattern of the nappe and the presence/behavior of the air cavity behind
the nappe are influenced by HT/Rcrest. Babb (1976) explored this relationship when
conducting model studies for Boardman Labyrinth Spillway; however, more research is
needed in this area.
Sidewall Angle (α) / Magnification Ratio (M). α refers to the angle (in degrees)
formed by the sidewall of a labyrinth relative to the cycle center line, see Fig. 2-1. M is
defined as Lc-cycle/w, or the ratio of the center-line length of a weir crest for a single cycle
(Lc-cycle) to the cycle-width (w). A variation in nomenclature that is commonly seen is M
= L/W. M and α are related geometrically by Eq. (2-24) for trapezoidal labyrinth weirs
and Eq. (2-25) for triangular labyrinth weirs.
( w − 2 Ac )
sin (α ) = trapezoidal (2-24)
(Lc−cycle − 2 Ac )
w
sin(α ) = triangular (2-25)
Lc−cycle
16
In Eq. (2-24), α is in degrees, Ac is the centerline length of the apex, and Lc-cycle =
2(lc+Ac), where lc is the centerline length of the weir sidewall.
Apex Ratio (A/w). A refers to the inside apex length of a labyrinth weir, as shown
in Fig. 2-2. Apexes are commonly used to facilitate constructability of concrete labyrinth
weirs (formwork and placement of steel reinforcing). From a hydraulic perspective,
structures with a smooth transition at the upstream apex (e.g., triangular labyrinth weirs,
piano-key weirs) are slightly more efficient than the abrupt transition typically found on
trapezoidal labyrinth weirs. Conversely, there is little performance difference regarding
downstream apexes due to the presence of a recirculating eddy or stagnation zone;
evidence of this zone is easily detectable with dye, fine sediment, or simply observing the
rise in the water surface profile in this area.
Two recent labyrinth weir installations feature atypical apexes. The efficiency of
Brazos Dam (Tullis and Young 2005) was increased by creating a relatively smooth
transition by rounding the apexes, as shown in Fig. 2-4. Also, the apexes of Boyd Lake
Spillway (Loveland, Colorado, USA) were notched to confine base-flow discharges and
facilitate flood routing (Brinker 2005). Base-flow discharges can also be confined by
slightly decreasing P for one or more labyrinth weir cycles.
A/w is useful for characterizing and comparing labyrinth weir geometries.
However, it is does not play a critical role in design optimization. The Tullis et al. (1995)
design method sizes A as tw ≤ A ≤ 2tw. A/w is the result of the optimizing L, α, and N in
an available spillway footprint and the structural requirements of the weir (e.g., tw).
Efficacy (ε). ε is a method for comparing the hydraulic performance of a labyrinth
weir to that of a linear weir. It incorporates sidewall angle and magnification, and is
17
Fig. 2-4. Rounded apexes of Brazos Dam, Texas, USA
presented as Eq. (2-26).
C d (α o )
ε= M (2-26)
C d (90o )
Families of HT/P curves are plotted as α vs. ε (see Fig. 2-5) to aid in sidewall
angle selection for a particular design head. According to Falvey (2003), ε is greatest for
all values of HT/P for an 8° quarter-round trapezoidal labyrinth. He based his analysis on
data from Tullis et al. (1995), which contains incorrect and less efficient α = 6° for HT/P
≤ 0.60. In Fig. 2-5, this conclusion is corrected. ε is plotted from quarter-round
trapezoidal labyrinth weir data from this study, which clearly shows an increasing trend
in ε with decreasing values of α. Due to the requirement of linear weir data (Cd = 90°),
which unnecessarily encumbers and complicates the procedure, it is proposed that ε be
replaced with ε’.

18
8
HT/P = 0.05
HT/P = 0.1
7 HT/P = 0.2
HT/P = 0.3
HT/P = 0.4
6
HT/P = 0.5
HT/P = 0.6
5 HT/P = 0.7
HT/P = 0.8
HT/P = 0.9
4
ε
HT/P = 1.0
0
0 10 20 30 40 50 60 70 80 90
α (°)
Fig. 2-5. Efficacy (ε) vs. α for quarter-round trapezoidal

labyrinth weirs, data set from this study
Cycle Efficiency (ε’). Cycle efficiency, ε’, was developed by Willmore (2004) as
a simple method for optimizing a labyrinth weir design, which is particularly useful for
low-head applications, and is presented as Eq. (2-27).
ε ′ = Cd (α )M
o (2-27)
Families of α curves are plotted as ε’ vs. HT/P to quickly view the hydraulic
performance of labyrinth weirs of different sidewall angles (see Fig. 2-6). As shown, ε’
appears to converge to a value of ~1.0 for all sidewall angles with increasing head.
Crest Shape. The shape of a weir crest can have a significant influence on the
hydraulic efficiency of a labyrinth weir. Examples of six crest shape definitions are
presented in Fig. 2-7. The most hydraulically efficient crest shape that has been
19
4.5
7-degree HR
8-degree HR
4.0
10-degree HR
3.5 12-degree HR
15-degree HR
3.0 20-degree HR
ε' = Cd*Lc-ycle/w
35-degree HR
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
HT/P
Fig. 2-6. Flow efficiency (ε’) of half-round trapezoidal

labyrinth weirs; data set from Willmore (2004)
constructed was an ogee-type crest (Willmore 2004); the leading radius is 1/3tw, and the
trailing radius is 2/3tw. The improved hydraulic efficiency is due to the structure
approximating the underside of the nappe profile. Half-round and ogee-type crest shapes
are also more efficient because these geometries allow the nappe to cling to the
downstream face of the weir at low heads, resulting in sub-atmospheric pressures
between the weir wall and nappe. An abrupt or sharp leading edge of a crest is less
efficient than a rounded (fillet) or chamfered leading edge.
Using HT/P to compare different crest shape data requires special consideration
when comparing sharp-crested or flat-crested data to round-crested weirs. Even though

20
the weir structures may have the same physical height, for sharp-crested and flat-crested
weirs the nappe springs from the leading edge of the crest and causes an effective
increase in P, resulting in Peffective. This is illustrated in Fig. 2-8.
Sharp Flat Quarter Half Ogee WES

Round Round
Fig. 2-7. Six examples of weir crest shapes
Nappe Interference. Nappe interference refers to the interaction of flow passing
over a weir in a converging flow situation (e.g., in the vicinity of the upstream apex of a
labyrinth weir cycle). The discharge over one weir wall interacts with and potentially
impacts the discharge efficiency of an adjacent weir wall by creating localized
submergence effects. For a trapezoidal labyrinth weir, the nappes from the sidewall not
only collide, but also interact with the nappe of the apex. An example of colliding nappes
near an upstream apex is shown in Fig. 2-9. Nappe collision is also dependent on the
nappe aeration condition and therefore the area of collision does not increase linearly
with increasing HT.

21
Fig. 2-8. Actual and effective weir height
Fig. 2-9. Nappe interference occurring for trapezoidal,

15° quarter-round labyrinth at HT/P = 0.20
22
The influence of nappe interference corresponds with the selection of N for a
labyrinth weir. Maintaining a constant length, the spillway footprint can be reduced by
increasing N; however, a 2-cycle labyrinth should be more efficient than a 20-cycle
labyrinth of equal length due to the increase in the number of apexes and consequently
the length of weir crest being affected by colliding nappes. This concept is demonstrated
in Fig. 2-10.
Indlekofer and Rouvé (1975) explored the concept of nappe interference by
studying sharp-crested corner weirs (α = 23.4°, 31°, 44.8°, 61.7°). A corner weir can be
characterized as a single triangular labyrinth weir cycle with channel boundaries
perpendicular to each sidewall. Indlekofer and Rouvé divided the corner weir into two
flow regions: a disturbed region where the flow from each sidewall converges (colliding
nappes) and a second region where the flow streamlines are perpendicular to the sidewall
(i.e., linear weir flow) (see Fig. 2-11).
The length of the crest within the disturbed area was defined as Ld. By comparing
Fig. 2-10. Nappe interference and cycle number

23
Fig. 2-11. Nappe interference as defined by Indlekofer and Rouvé (1975)

for sharp-crested corner weirs
the efficiency of a corner weir to a linear weir, an average discharge coefficient for the
disturbed area, Cd-m; a theoretical disturbance length, LD; and an empirical discharge
relationship were developed [Eq. (6-2)]. Cd-m represents the efficiency of a corner weir
relative to a linear weir (Cd-m = Cd-corner / Cd(90°)). Applying the linear weir discharge
coefficient, Cd(90°), to the corner weir, LD represents the theoretical portion of crest length
where Q and Cd = 0 (see Fig. 2-11).
 3Q  1 1
Ld =  Lc(α ° ) −  = LD
 2Cd( 90° ) 2 g hm  1 − Cd − m
3 2
1 − Cd − m

(2-28)
In Eq. (2-28) hm is the head upstream of the weir as defined by Indlekofer and
Rouvé (1975); hm represents a specific upstream depth and includes two velocity
components [see Indlekofer and Rouvé (1975) for details].
Falvey (2003) applied this approach to the experimental results of several
labyrinth weir models. Using corner weir data, Falvey developed an empirical LD
relationship [Eq. (2-29)] as an alternative to polynomial relationships developed by

24
Indlekofer and Rouvé. Falvey also developed Eq. (2-30) based upon an analysis of
available labyrinth weir experimental data. Falvey does not, however, give a
recommendation with regard to which LD equation is most appropriate or accurate.
Based on an analysis of Tullis et al. (1995) labyrinth weir discharge rating curves, Falvey
proposes a design limit of LD / lc ≤ 0.35 (35% or less of weir length is ineffective), where
lc is the weir sidewall length. For corner weirs and triangular labyrinth weirs, lc = Lc-cycle /
2; for trapezoidal labyrinth weirs, Lc-cycle / 2 = lc+Ac. Falvey also states that additional
research is needed, including ascertaining the validity of Eq. (2-30). In Eq. (2-30), HT/P
is the headwater ratio (total upstream head over the weir height).
LD
= 6.1e− 0.052α ° α ≥ 10° (2-29)
h
 H  
LD = lc  0.224 ln T  + (0.94 − 0.03α °) α ≤ 20° and HT/P ≥ 0.1 (2-30)
  P  
The work of Indlekofer and Rouvé (1975) provides some insights for labyrinth
weir nappe interference; however, flow efficiency is also influenced by the approach flow
streamlines orientation as they pass over the weir sidewall and local submergence. The
streamlines are generally not perpendicular to a labyrinth weir crest, except at very low
heads, as the streamline trajectory deviates more and more from perpendicular as HT
increases. Falvey (2003) expressed the need for additional labyrinth weir nappe
interference research. Crookston and Tullis (2010) conducted preliminary investigations
into this concept, as applied to half-round labyrinth weirs, and concluded that a more
accurate method to describe nappe interference is needed.
Nappe Aeration. Nappe aeration refers to the presence or absence of an air cavity
behind the nappe; structures can be used that ‘artificially’ aerate the nappe, creating a
25
‘vented’ condition. This study defines the nappe aeration conditions of labyrinth weirs
as: Clinging, Aerated, Partially Aerated, and Drowned; however, other terms can be
found in literature. For example, Falvey (2003) refers to four nappe aeration conditions
(termed crest flow conditions) and are: Pressure, Atmospheric, Cavity, and
Subatmospheric. Lux (1989) refers to aerated, transitional (unstable air cavity), and
suppressed (solid water flow at high head) aeration condition. This study also identifies
an unstable nappe condition, which refers to a nappe with an oscillating trajectory that is
often accompanied by shifting nappe aeration conditions.
Nappe aeration conditions are a function of crest shape, velocity head, turbulence,
and tailwater elevation adjacent to the labyrinth sidewalls. Venting the nappe to the
atmosphere, or artificial aeration, can stabilize the pressures behind the nappe and
therefore may aid in stabilizing an unstable or oscillating nappe and may decrease
vibrations and noise (Naudascher and Rockwell 1994). It should be noted that nappe
vibration is not generally caused or remedied by nappe aeration conditions (Falvey 1980).
Artificial aeration can be accomplished with nappe breakers (also called splitter
piers) that are placed on top of the crest, or with vents (e.g., circular conduits). Hinchliff
and Houston (1984) recommend that nappe breakers be located a distance of
approximately 10% of the sidewall length (lc) from the downstream apex, based upon
research conducted for Ute and Hyrum Dams. However, there was no other information
found in published literature to design, configure, locate, or size nappe breakers and vents
for neither labyrinth weirs nor the conditions for which they are effective with respect to
labyrinth weir geometry and flow conditions.

26
Cycle Configuration, Weir Orientation, and Weir Placement. Traditionally,
labyrinth weir cycles follow a linear configuration [e.g., Lake Townsend (Tullis and
Crookston 2008), Bartletts Ferry (Mayer 1980)]; however, curved or arced labyrinth
weirs have also been constructed [e.g., Avon (Darvas 1971), Kizilcapinar (Yildiz and
Uzecek 1996), and Weatherford (Tullis 1992). Arced labyrinth configurations increase
efficiency by orienting the cycle to take advantage of the converging nature of the
reservoir approach flow.
Houston (1983) conducted a study of Hyrum Dam where the test program
included various weir orientations and placements of the labyrinth weir relative to the
reservoir discharge channel (normal, inverse, flush, and partially projecting) of the two-
cycle labyrinth weir. Examples of linear and arced cycle configurations, and four general
labyrinth weir orientations and placements are presented in Fig. 2-12.
Houston (1983) found that for channelized approach flow conditions, the normal
orientation had 3.5% greater discharge than the inverse orientation, and partially
projecting increased discharge by 10.4% when compared to flush with intake. It should
be noted that curved guide walls or a rounded inlet were used immediately upstream of
the labyrinth, and that the results of this study may be limited because the weir was
comprised of only two cycles. Additional research is needed to provide design guidance
for labyrinth orientations and placements (including N ≥ 2), primarily because current
design methodologies have been developed in channelized flow conditions (laboratory
flumes). At present, there is no design information available for arced labyrinth weirs.
27
Fig. 2-12. Linear and arced (fully projecting) labyrinth cycle configurations, flush,
rounded inlet, inverse, and normal orientations
Labyrinth Weir Design Methods
Early Investigations
Two studies were conducted that provided initial insights into labyrinth weir
behavior. However, due to the limited scope of each study, there were insufficient data
28
for general labyrinth weir design.
Gentilini (1940) published a study based upon previous work on oblique weirs by
placing multiple oblique weirs together to form triangular labyrinth weirs. The sharp-
crested weirs were tested at three sidewall angles (α=30°, 45°, and 60°) and relatively
small w/P ratios. Due to the large operating head (compared to cycle width), Gentilini’s
results were found to be dependent on w/P and were presented as a function of h/w.
Kozák and Sváb (1961) tested eleven different trapezoidal labyrinth weirs (tw=6-
mm) with a flat-topped crest with both edges chamfered. The tested weirs had the
following parameter ranges: 0.05 ≤ h/P ≤ 0.25, 5.7° ≤ α ≤ 20.6°, 1.23 ≤ Lc-cycle/w ≤ 4.35,
1.15 ≤ w/P ≤ 4.61. Kozák and Sváb concluded that the discharge capacity of labyrinth
weirs is appreciably greater than a linear weir operating under the same head. They also
concluded that a larger number of small cycles are more efficient and economical than a
labyrinth weir of equivalent length composed of fewer cycles. It is important to note that
this study was conducted for small operating heads where discharge capacity is not
significantly reduced by sidewall angle and nappe interference.
A summary of the tested labyrinth weir parameters from the design methods
presented in the following discussion are summarized in Table 2-1.
Taylor (1968) and Hay and Taylor (1970)
Geoffrey Taylor conducted a large study (24 models) primarily on triangular
labyrinth weirs along with a limited number of trapezoidal and rectangular weirs. Two
crest shapes were investigated, sharp-crested and half-round, and Taylor also explored
four sloped apron configurations. The weirs were tested for 0.05 ≤ h/P ≤ 0.55. Hay and
29
Table 2-1. Summary of labyrinth weir parameters from design methods
Study Crest Type

Triangular
Sh
1 Hay and Taylor (1970) Trapezoidal
HR
Rectangular
2 Darvas (1971) LQR Trapezoidal
Sh Triangular
3 Hinchliff and Houston (1984)
QR Trapezoidal
Lux and Hinchliff (1985) Triangular
4 QR
Lux (1984, 1989) Trapezoidal
5 Magalhães and Lorena (1989) WES Trapezoidal
6 Tullis et al. (1995) QR Trapezoidal
7 Melo et al. (2002) LQR Trapezoidal
8 Tullis et al. (2007) HR Trapezoidal
9 Emiroglu et al. (2010) Sh Triangular
† QR – Quarter-round (Rcrest=tw/2), LQR – Large Quarter-round (Rcrest=tw),
HR – Half-round, Sh – Sharp, WES – Truncated Ogee
Taylor (1970) defined the hydraulic performance in terms of flow magnification,
QLab/QLin (Labyrinth weir discharge / Linear weir discharge) vs. h/P. They present
Effectiveness (E) to determine the advantages gained from an increase in crest length,
shown in Eq. (2-31).
QLab QLin
E (%) = *100 (2-31)
Lw
In addition to two discharge relationship charts specific to sharp-crested labyrinth
weirs, this design method gives recommendations regarding Lc-cycle/w, submergence,
channel-bed elevation, aprons, and general nappe interference. However, the authors
neglected the velocity component in the driving head (results limited to channels and not
including V2/2g) and concluded that discharge is relatively independent of w/P. They
suggest using maximum possible values for α and recommend triangular labyrinth weirs.
Hay and Taylor (1970) discouraged the use of labyrinth weirs where they would operate
30
under submerged conditions or with a high tailwater that would remove the aeration
cavity behind the nappe (based upon hydraulic efficiency).
Darvas (1971)
Darvas (1971) introduced an empirical discharge equation, Eq. (2-32), to
accompany a design chart. His approach utilizes HT, and introduced Cd-Darvas, a
dimensional labyrinth weir discharge coefficient (ft1/2/s).
Q
C d − Darvas = 3
(2-32)
WH T 2
Results are presented as Cd-Darvas vs. Lc-cycle/w, and include a family of HT/P design
curves (0.2 ≤ HT/P ≤ 0.6) for trapezoidal labyrinth weirs without aprons, and w/P ≥ 2.
The supporting data for this design method are limited to a large quarter-round (Rcrest =
tw) crest shape and are based upon physical model studies of Avon Dam (α = 22.8°) and
Woronora Dam (α = 27.5°).
Hinchliff and Houston (1984)
The U.S. Bureau of Reclamation conducted flume studies of labyrinth weirs to aid
in the design of Ute Dam; the design was beyond the scope of Hay and Taylor (1970) and
it was important to confirm their results. Discrepancies between investigations were
attributed to variation in upstream head definition [h, Hay and Taylor (1970), HT, USBR]
in his flume investigations.
Labyrinth spillway design guidelines (Hinchliff and Houston 1984) were
developed based on the results of the Ute Dam and Hyrum Dam model studies; including
rating curve data presented in a form consistent with Hay and Taylor (1970). As
31
previously mentioned, the information regarding weir placement provided new insights in
labyrinth weir design, despite scope limitations (N = 2).
Lux and Hinchliff (1985), Lux (1984, 1989)
Lux and Hinchliff (1985) and Lux (1984, 1989) presented a different discharge
coefficient Cd-Lux, which included the vertical aspect ratio (w/P) and a shape constant (k)
to determine the discharge of a single labyrinth cycle (Qcycle), presented as Eq. (2-33).
Qcycle
Cd − Lux = (2-33)
wP 32
w g HT
w P+k
Although this non-dimensional equation applies to trapezoidal and triangular
weirs, the inclusion of w/P complicates the weir equation, especially with the design
limitation of w/P ≥ 2.0. Similar parameter limits have been prescribed by other design
methods that do not explicitly include w/P in the head-discharge equation.
Magalhães and Lorena (1989)
Magalhães and Lorena (1989) developed curves similar to Darvas (1970) for a
truncated ogee crest-shaped labyrinth weir crest (referred to as a “WES”) and present a
dimensionless discharge coefficient, Cd-M&L, as shown in Eq. (2-34).
Q
Cd −M &L = 32
(2-34)
W 2g HT
Eq. (2-34) is similar to the standard weir equation [Eq. (2-1)] without the 2/3 term
and the channel width (W) instead of crest length (Lc) is the selected characteristic length.
Experimental results obtained in this study were systematically lower than those of
Darvas (1971). This design method includes a comparison of predicted Cd-M&L from their
32
study to values predicted by Darvas (1971) and computed Cd-M&L from six other hydraulic
model studies conducted at the Laboratório Nacional de Engenharia Civil (LNEC),
Lisbon, Portugal (Harrezza, Keddara, Dungo, São Domingos, Alijó, and Gema). This
additional information gives confidence in a design method, and this validation technique
has since been used, for example, by Tullis et al. (1995), Falvey (2003), and by the author
of this dissertation.
Tullis, Amanian, and Waldron (1995)
Tullis et al. (1995) made a minor adjustment to the conventional weir equation to
define the discharge coefficient, Cd-Tullis, as presented in Eq. (2-35).
3Q
C d −Tullis = 32
(2-35)
2 2 g Le H T
In Eq. (2-35), Cd-Tullis is dimensionless, and the characteristic length is an effective
weir length, Le. This method is based upon research conducted by Amanian (1987),
Waldron (1994), and a model study for Standley Lake (Tullis 1993). Labyrinth weir
discharge coefficient data are presented as Cd-Tullis vs. HT/P, with the data segregated by
weir sidewall angle (α). The discharge coefficient of a linear weir was also included for
comparison. The data were fit with eight regression equations (quartic polynomials).
Also, the findings of Amanian (1987) for labyrinth weirs oriented at an angle (β) to the
approaching flow are noted.
A significant and unique contribution of this study is the presentation of the
design method as a spreadsheet program used to optimize a labyrinth weir design. This
approach may be partially responsible for the widespread use of this design method in the
USA. For example, the Tullis et al. (1995) design method was used (as presented by
33
Falvey 2003) to design the emergency labyrinth spillway (59-cycles, α = 8°) for Boyd
Lake, located in Loveland, Colorado, USA (Brinker 2005). The spillway width is nearly
400-m, the labyrinth weir length is ~2.3-km, and has a maximum discharge capacity of
1,200 cms.
The method’s support data are, however, limited to labyrinth weirs with quarter-
round crest shapes (Rcrest=tw/2), α ≤ 18°, and 3 ≤ w/P ≤ 4. Willmore (2004) corrected a
minor error in the Tullis et al. (1995) method associated with computing Le. Willmore
also found the α = 8° data to be in error. A closer examination of the discharge data for
HT/P ≤ 0.2 reveals disorderly Cd-Tullis values. Also, the Cd-Tullis values for α = 6° are
significantly lower than the other curves; additional investigations at the UWRL found
higher Cd values that are much closer to the other labyrinth weir coefficient curves.
Melo, Ramos, and Magalhães (2002)
Based upon their study of a single-cycle labyrinth weir located in a channel with
converging walls, Melo et al. (2002) further developed the methodology of Magalhães
and Lorena (1989) by adding an adjustment parameter, kθ-CW, shown in Eq. (2-36). This
design method presents kθ-CW as a function of θCW (0° – 90°) to include the effect of
converging channel walls (1.0 ≤ kθ-CW ≤ 1.4), which increase labyrinth weir efficiency by
directing a larger upstream flow area into a labyrinth weir cycle (converging flow) and
improving the orientation of the flow lines to the labyrinth weir sidewall (closer to
perpendicular).
Q
Cd −M &L = 32
(2-36)
kθ −CW W 2 g H T
34
Tullis, Young, and Chandler (2007)
Previous to the Tullis et al. (2007) study, the linear weir submergence method
developed by Villemonte (1947) was commonly applied to labyrinth weirs for lack of a
more appropriate alternative. Tullis et al. (2007) developed a dimensionless submerged
head relationship for labyrinth weirs that is simple to solve and has an average predictive
error of 0.9%, shown as Eqs. (2-37)-(2-39). The procedure is iterative; the author of this
dissertation has modified the presentation to facilitate graphical solutions of this method,
shown in Fig. 2-13.
4 2
H* H  H  H 
= 0.0322 d  + 0.2008 d  + 1 0 ≤  d  ≤ 1.53 (2-37)
HT  HT   HT   HT 
2
H* H  H 
= 0.9379 d  + 0.2174 1.53 ≤  d  ≤ 3.5 (2-38)
HT  HT   HT 
H 
H* = Hd 3.5 ≤  d  (2-39)
 HT 
In Eqs. (2-37) – (2-39), H* is the total upstream head on a submerged labyrinth
weir, HT is the total upstream head on an unsubmerged labyrinth weir (same Q associated
with HT), and Hd is the total head downstream of the labyrinth.
This method for evaluating labyrinth submergence has been verified by Lopes et
al. (2009), who studied α = 12° and 30° labyrinth weirs in a sloped and horizontal
channel. Their experimental results were reported to be within 6% of those presented by
Tullis et al. (2007).
Emiroglu, Kaya and Agaccioglu (2010)
Emiroglu et al. (2010) studied the discharge capacity of a single-cycle labyrinth

35
S 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 1.0
5.0
4.5
4.0
3.5
3.0
H*/HT
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Hd/HT
Fig. 2-13. Graphical solution for labyrinth weir submergence,

modified from Tullis et al. (2007)
weir (22.5° ≤ α ≤ 75°) used as a side or lateral weir in straight channels (side-channel
application). The water surface profiles, velocity profiles, upstream Froude numbers (Fr)
and discharge coefficients (Cd-side) were presented as Cd-side vs. Fr and Cd-side vs. P/h. The
empirical discharge equation is presented as Eq. (2-40).
−1.431
 W 
0.012
W 
0.112

18.6 − 23.535  + 6.769  − 
w L
C d − side =  (2-40)
 4.024 
0.502 P  + 0.094 sin (2α ) − 0.393Fr 2.155 
 h 
Labyrinth Weir Case Studies
Design methods are useful tools for predicting and extrapolating the hydraulic
36
performance of case-specific labyrinth weir models and prototype structures. A physical
model study is particularly useful when a specific labyrinth weir configuration, operating
condition, or prototype condition is not included in available labyrinth weir information.
Also, a physical model study is a useful tool for refining and finalizing a labyrinth weir
design, and has the potential to significantly reduce estimated construction costs. An
example is Lake Brazos Dam; a photo of the full-width model (1:15 scale) is presented in
Fig. 2-14. This innovative spillway (designed by Freese and Nichols, Inc.) is the result of
using available labyrinth information and physical modeling to design a unique, efficient,
and cost effective structure [estimated savings of ~$14 million, Vasquez et al. (2007)].
Due to the hydraulic characteristics and a highly adaptable geometric design,
labyrinth weirs are a favored design option for spillway rehabilitation, replacement, and
new spillways. For example, Schnabel Engineering (www.schnabel-eng.com/) has
Fig. 2-14. Full-width model of Lake Brazos labyrinth spillway in Waco, Texas, USA
37
recently designed 39 labyrinth weir spillways; 15 have been built and 8 are in various
stages of completion (personal communication, June 22, 2010). Also, there are labyrinth
spillways located throughout the globe; Table 2-2 presents a list of labyrinth weir
structures (including reference citation) that have been built and/or a physical model
study was conducted. For details concerning labyrinth weir prototype geometries, site
conditions, design flow rates, downstream hydraulic conditions, etc. please refer to the
referenced reports and publications.
Labyrinth Weir Research Studies
Amanian M.S. Thesis (1987)
Amanian (1987) tested linear weirs and half-round triangular labyrinth weirs in a
channel, including oblique labyrinth weirs (the labyrinth cycles oriented at an angle β to
the approaching flow, shown in Fig. 2-3). The weirs were fabricated from plywood, with
tw~19.05-mm. Although there are very few data points associated with each physical
model, Amanian did test eight labyrinth weirs and eleven linear weirs. A summary of the
tested geometries are presented in Table 2-3.
Trends appear to have been difficult to discern due to the small number of data
points; however, Amanian states that good agreement was found between the sharp-
crested experimental results and the results of previous studies. Limited information is
provided regarding nappe aeration conditions during testing. Amanian concluded that the
discharge efficiency of labyrinth weirs declines as HT increases (due to submergence and
nappe interference), and efficiency can be increased with a half-round crest shape
(relative to quarter-round, flat, or sharp crest shapes).

38
Table 2-2. Labyrinth weirs from across the globe
Name Location Source

Agua Branca Portugal Quintela et al. (2000)
Alfaiates Portugal Quintela et al. (2000)
Alijó Portugal Magalhães & Lorena (1989)
Arcossó Portugal Quintela et al. (2000)
Avon Australia Darvas (1971)
Bartletts Ferry USA Mayer (1980)
Belia Zaire Magalhães & Lorena (1989)
Beni Bahdel Algeria Afshar (1988)
Boardman USA Babb (1976), Lux (1985)
Boyde Lake USA Brinker (2005)
Brazos USA Tullis and Young (2005)
Calde Portugal Quintela et al. (2000)
Carty USA Afshar (1988)
Castelletto-Nerv. Canal Italy Magalhães & Lorena (1989)
Cimia Italy Lux & Hinchliff (1985)
Dog River USA Savage et al. (2004)
Dungo Angola Magalhães & Lorena (1989)
East Park USA Magalhães & Lorena (1989)
Estancia Venezuela Magalhães & Lorena (1989)
Forestport USA Lux (1989)
Garland Canal USA Lux & Hinchliff (1985)
Gema Portugal Magalhães & Lorena (1989)
Harrezza Algeria Lux (1989)
Hyrum USA Houston (1983)
Infulene Canal Mozambique Magalhães & Lorena (1989)
Juturnaiba Brazil Afshar (1988)
Keddara Algeria Magalhães & Lorena (1989)
Kizilcapinar Turkey Yildiz (1996)
Lake Townsend USA Tullis & Crookston (2008)
María Cristina Dam Spain Page et al. (2007)
Mercer USA CH2M-Hill (1976)
Navet Pumped Storage Trinidad Phelps (1974)
Ohau C Canal New Zealand Walsh (1980)
Pacoti Brazil Magalhães & Lorena (1989)
Pisão Portugal Quintela et al. (2000)
Prado USA Copeland and Fletcher (2000)
Quincy USA Magalhães & Lorena (1989)
Ritschard USA Vermeyen (1991)
Rollins Dam USA Tullis (1986)
Saco Brazil Quintela et al. (1988)
São Domingos Portugal Magalhães & Lorena (1989)
Sam Rayburn Lake USA USACE (1991)
Santa Justa Portugal Magalhães & Lorena (1989)
Sarioglan Turkey Yildiz (1996)
Sarno Algeria Afshar (1988)
Skelton Grange Canal England Magalhães & Lorena (1989)
Standley Lake USA Tullis (1993)
Teja Portugal Quintela et al. (2000)
Ute USA Houston (1982)
Weatherford USA Tullis (1992)
Woronora Australia Darvas (1971)
39
Table 2-3. Summary of physical models tested by Amanian (1987)
Model α P Lcycle β Lc-cycle/w w/P N Crest Type

(°) (cm) (cm) (°)
1 10.5 15.48 505.45 0 5.53 3.94 1.5 HR Triangular
2 16 15.09 315.29 0 3.45 4.04 1.5 HR Triangular
3 21 15.30 257.25 0 2.81 3.98 1.5 HR Triangular
4 32.12 15.42 163.98 0 1.79 3.95 1.5 HR Triangular
5 49.04 15.67 116.13 0 1.271 3.89 1.5 HR Triangular
6 24.5 15.42 162.46 0 2.55 2.39 2 HR Triangular
7 24.5 15.42 162.46 30 2.55 2.39 2 HR Triangular
8 24.5 15.42 162.46 45 2.55 2.39 2 HR Triangular
9 90 .335ft 91.44 0 1 8.96 - HR Linear
10 90 0.668 91.44 0 1 4.49 - HR Linear
11 90 .500 91.44 0 1 6.00 - HR Linear
12 90 .363 91.44 0 1 8.26 - HR Linear
13 90 .309 91.44 0 1 9.71 - HR Linear
14 90 0.674 91.44 0 1 4.45 - Sh Linear
15 90 0.339 91.44 0 1 8.85 - Sh Linear
16 90 0.675 91.44 0 1 4.44 - QR Linear
17 90 0.343 91.44 0 1 8.75 - QR Linear
18 90 0.673 91.44 0 1 4.46 - Flat Linear
19 90 0.343 91.44 0 1 8.75 - Flat Linear
Waldron M.S. Thesis (1994)
Waldron (1994) also conducted physical modeling of linear weirs and trapezoidal
labyrinth weirs in a channel. All labyrinth weirs featured a quarter-round crest shape and
were oriented perpendicular to the approaching flow (β = 0°). The weirs were fabricated
from plywood, with tw~25.4-mm. The apron length B was held constant for all tested
models by varying N, which produced partial cycles. A summary of the labyrinth weirs
tested is presented in Table 2-4.
From the experimental results, Waldron concluded that Cd is independent of N
(based upon 12° data). Nappe performance (springing, clinging, drowning) and the
corresponding HT/P values were noted. Waldron stated that the peak Cd values signal the
40
Table 2-4. Summary of physical models tested by Waldron (1994)
α P Lc-cycle Lc-cycle/w w/P N Crest Type

(°) (cm) (cm)
1 6 17.22 149.48 6.47 5.36 3.99 QR Trapezoidal
2 9 17.19 146.48 4.78 5.37 3.01 QR Trapezoidal
3 12 17.22 146.67 3.94 5.36 2.48 QR Trapezoidal
4 12 15.82 117.38 3.82 5.83 3.00 QR Trapezoidal
5 12 7.86 120.03 4.35 11.73 3.34 QR Trapezoidal
6 12 16.06 153.95 6.68 5.74 4.00 QR Trapezoidal
7 15 16.70 153.52 3.33 5.52 2.00 QR Trapezoidal
8 18 17.01 149.63 2.79 5.42 1.72 QR Trapezoidal
9 21 17.65 147.31 2.48 5.23 1.55 QR Trapezoidal
10 90 24.201 92.23 1 3.81 - Flat Linear
11 90 23.957 92.35 1 3.85 - QR Linear
point where the weir is no longer self-aerating, which is not correct. Polynomial design
curves are presented for Cd-Waldron (similar to Amanian 1987), and not a dimensionless Cd.
Waldron does acknowledge the discrepancies between Cd-Waldron values at low HT/P
(crossing α curves), stating that the accuracy and precision of the experimental setup was
insufficient in this range. As a solution, Waldron suggests an average Cd-Waldron value be
used in design; Waldron also states that these errors are irrelevant because labyrinth weirs
are typically designed for HT/P > 0.3. To determine an optimum α, Waldron calculated a
unit discharge (Qcycle/w); from the computed results, a 12° labyrinth weir is recommended
as the most efficient cycle configuration.
Waldron compared his experimental results the experimental results of physical
model studies of Avon, Bartletts Ferry, Boardman, Hyrum, Ritschard, South Heart, and
Ute labyrinth spillways. There were varying degrees of agreement, which were attributed
to differences in labyrinth weir geometry, approach conditions, tw and P, and studies that
used h instead of HT.

41
The data from Waldron (1994) is limited by the tested range of geometric
configurations, crest shape, and it requires the use of Le instead of Lc. In addition, the
defined discharge coefficient combined a portion of the weir equation
Cd −Waldron = 2 3 Cd 2 g and is not dimensionless (ft1/2/s). The experimental results of this
study are used in the Tullis et al. (1995) design method. However, the 21° and 9° data
sets are not included and the 9° head-discharge data lies below the 8° degree data set
presented by Tullis et al. (1995).
Willmore M.S. Thesis (2004)
Willmore tested 2-cycle (B was not restricted), trapezoidal labyrinth weirs (tw =
36.96 mm) in a rectangular laboratory flume. The majority of the tested models featured
a half-round crest shape (7° ≤ α ≤ 35°); however, quarter-round and a new ‘ogee’ crest
shape were also tested for α = 7° and 8° (see Table 2-5). Models were tested with and
without a vented nappe, and the aeration or clinging nappe aeration condition HT/P
ranges were documented. The influence uniform sediment deposits (a false floor placed
upstream and within the labyrinth weir upstream cycles) and the influence of a ramp
located immediately upstream of a labyrinth weir physical model were also examined.
Willmore developed polynomial curve-fit equations for Cd vs. HT/P for all tested
models, based upon Lc. Willmore also developed new polynomial curves (also based
upon Lc) for the Tullis et al. (1995) quarter-round data and corrected a trigonometric error
in the calculations of B. Willmore found the effects of an upstream ramp to be negligible,
and Cd was not influenced by the installation of false flooring (uniform sediment deposit).
Finally, Willmore reports that the ‘ogee’ crest shape is more efficient than the half-round
42
Table 2-5. Summary of physical models tested by Willmore (2004)
α P Lcycle Lc-cycle/w w/P N Crest Type

(cm) (cm)
1 7 28.56 398.65 6.56 2.13 2 QR Trapezoidal
2 7 30.30 398.65 6.56 2.01 2 HR Trapezoidal
3 7 28.38 398.65 6.56 2.14 2 Ogee Trapezoidal
4 8 30.45 351.43 5.78 2.00 2 QR Trapezoidal
5 8 30.42 351.43 5.78 2.00 2 HR Trapezoidal
6 8 30.11 351.43 5.78 2.02 2 Ogee Trapezoidal
7 10 30.42 285.29 4.69 2.00 2 HR Trapezoidal
8 12 30.42 241.63 3.97 2.00 2 HR Trapezoidal
9 15 30.42 197.54 3.25 2.00 2 HR Trapezoidal
10 20 30.42 153.57 2.53 2.00 2 HR Trapezoidal
11 35 30.51 97.64 1.61 1.99 2 HR Trapezoidal
crest shape, which is more efficient than a quarter-round crest shape. Flow efficiency
[referred to as cycle efficiency (ε’) in this study] was proposed as a new parameter to
compare discharge capacities of labyrinth weirs of different α.
Lopes, Matos, and Melo (2006, 2008)
Lopes (Ph.D. dissertation, forthcoming) and Lopes et al. (2006, 2008)
investigated flow patterns, air entrainment, characteristic depths, flow bulking, and
residual energy immediately downstream of a trapezoidal labyrinth weir. Physical
modeling of a trapezoidal labyrinth weir with a quarter-round crest shape (α = 30°, w/P =
2) and a horizontal apron and downstream chute was conducted at the Laboratório
Nacional de Engenharia Civil (LNEC) Flow pattern observations noted the impact
locations of the nappe to the downstream water surface, spray regions, maximum
tailwater depths, shockwaves, and shockwave intersection locations. Air concentrations
were measured with a conductivity probe (Matos and Frizell 1997, 2000) aligned with the
main downstream flow direction. Based upon their experimental results and the results of
43
Magalhães and Lorena (1994) for labyrinth weirs with a ‘WES’ crest shape, an empirical
relationship was developed to predict the relative residual energy (Hresidual) at the base of
a labyrinth weir [Eq. (2-41)].
H residual H 
= 0.709 + 0.254 ln T  (2-41)
HT + P  P 
Mean air concentration distributions, shockwave intersection locations, and
minimum, maximum, and 90% local air concentration characteristic flow depths [ymin,
ymax, y90] are also presented graphically. Average air concentration (Cair-avg) compared
favorably with the advective diffusion model developed by Chanson (1995, 1997) for
self-aerated chute flows.

44
CHAPTER 3
EXPERIMENTAL SETUP AND TESTING PROCEDURE
Test Facilities
All Research for this study was performed in the primary hydraulics testing bay at
the Utah Water Research Laboratory (UWRL), located on the Utah State University
Campus in Logan, Utah (Fig. 3-1). Two facilities were used for physical modeling: a
rectangular flume (Fig. 3-2) for channelized applications and a large headbox for
reservoir applications (Fig. 3-3). Water to the UWRL is supplied from 1st Dam, located
on the Logan River; gravity fed flow rates from the dam can exceed 7 cms.
In this study, great care was taken to minimize random and systemic errors. An
extensive review of published literature was conducted before the formulation of the
physical model test program. In particular, the physical model facilities, construction
materials, test procedures, and data accuracy from Amanian (1987), Waldron (1994),
Willmore (2004), and Young (2005) were examined to enhance the accuracy of the
experimental results. Further experimentations and observations during testing refined
Fig. 3-1. Outside view of the Utah Water Research Laboratory main building
and the primary hydraulics testing bay within
45
Fig. 3-2. Rectangular flume test facility
Fig. 3-3. Reservoir test facility

46
the testing procedures adopted for this study. The results are a highly accurate,
controlled, and repeatable experimental method, which has proved satisfactory for this
research study.
Experimental Setup
Rectangular Flume Facility
The tilting rectangular laboratory flume (1.2 m wide x 14.6 m long x 1.0 m deep)
is composed of a steel framework and acrylic panels for the walls and floor. The slope of
the flume is adjusted by four large mechanical jacks; for this study the longitudinal slope
of the flume floor, Sbed, was set to zero. The labyrinth weir models were installed upon a
horizontal platform (2.44 m long x 30.5 cm tall) made of High Density Polyethylene
Plastic (HDPE) that featured adjustable steel supports every 15 cm. After installation the
platform was adjusted until horizontally level (±0.4 mm). A 2.44-m long ramp installed
at ~7° upstream of the platform allowed for a smooth transition between the flume floor
and the platform. Based upon the findings of Willmore (2004), who tested the effects of
ramps upstream of a labyrinth weir, the placement and geometry of this ramp had no
discernable effects (relative to a horizontal approach) on the hydraulic performance of the
physical models tested in this study.
Two supply lines convey water to a steel headbox that contains a baffle structure
to establish tranquil flows and uniform approach conditions to the flume. The diameters
of the small and large supply lines are approximately 20.3-cm (8 in) and 50.8-cm (20 in).
Maximum flow conveyed by the 50.8-cm pipeline is approximately 0.68 cms (24 cfs). At
47
the downstream exit, the flume features a sluice gate and a stop-log structure to control
tailwater elevations. A schematic of the test facility is presented as Fig. 3-4.
Reservoir Facility
Reservoir simulations were conducted in an elevated headbox (7.3-m x 6.7-m x
1.5-m deep). Similar to the rectangular flume, a large platform was constructed from
10.2-cm (4-in) steel box beams and 19-mm thick HDPE sheeting. The horizontal
platform was surveyed to within ±0.4-mm of level. A false floor was installed over the
remaining portion of the headbox to maintain a constant depth and uniform approach
flow conditions. The apron downstream of the labyrinth weir was the same elevation as
the upstream floor in the reservoir.
Three pipelines (10.2-cm, 20.3-cm, and 50.8-cm) supply flows to a diffuser that is
located along three sides of the headbox, behind a baffle wall made of fine synthetic
mesh (such as those commonly used in swamp-coolers). This setup conveys flows to the
labyrinth models from 180°. After passing over the weirs, the flow drops ~2.3 m to a
Fig. 3-4. Schematic of the rectangular flume test facility

48
collection channel; there was no structure to control tailwater depths. A schematic of the
reservoir test facility is presented as Fig. 3-5.
Fig. 3-5. Schematic of the reservoir test facility

49
Instrumentation
Flow rates for each test facility were metered using calibrated orifice flow meters
(located in the supply piping). Images of the supply piping and the orifice flow meters
are presented in Figs. 3-6 to 3-8. Differential pressures were measured using pressure
transducers and an electronic data logger (Fig. 3-9). The data logger recorded the
average differential pressure used to calculate the flow rate for a particular hydraulic
condition.
Water temperatures were taken with a Traceable® Thermometer (SN 9146829 –
9/29/2011) with a range of -58°F to 302°F and readable to ±0.05°F. Stilling wells
equipped with point gages (readable to ±0.15 mm) were used to determine the
approaching flow depths, shown in Figs. 3-10 and 3-11. The hydraulic connection
location or ‘pressure tap’ in the rectangular flume was ~ 1-m upstream (3.3HT-max) of the
labyrinth weir models (Fig. 3-4). The tap location in the reservoir (Fig. 3-5) was located
at a tranquil section on the facility floor, between the baffle and where the platform began
and along the centerline of the test facility. When properly located, the stilling well can
Fig. 3-6. Supply piping to the rectangular flume test facility

50
Fig. 3-7. 4-in and 8-in supply piping and orifice plates for the reservoir test facility
Fig. 3-8. 20-in Supply piping and orifice plate for the reservoir test facility
give a highly accurate depth measurement, even when the water surface is uneven (e.g.,
small waves, surface turbulence).
The rectangular flume featured a rolling carriage that rested upon guiderails
mounted to the top of the flume walls, Fig. 3-12. The rails were surveyed prior to testing
51
Fig. 3-9. Flow measurement equipment (power supply, pressure transducer,

data logger, and Hart communicator)
to ±0.4-mm of level. A point gage with interchangeable tips (straight, hooked) was fixed
to a machined rail system that was bolted to the upstream face of the carriage (see Fig. 3-
13). This point gage was used for nappe profiling. The downstream face of the carriage
featured a second point gage fixed to a worm-gear assembly. This gage was used to
determine the crest elevation and weir height of the labyrinth models.
A 2-D Sontek Flowtracker Handheld ADV (Acoustic Doppler Velocimeter) (Fig.
3-14), mounted to a wading rod, was used for 1-point and 3-point velocity profiling and
field mapping. Sontek reports that this Flowtracker unit has a velocity measurement
range of 0.0009 m/s to 4.572 m/s (0.003 ft/s to 15 ft/s). Also, still and video photography
were used to document weir flow behavior. A dye injection device (dye tank and
“wand”) and particles of various sizes and densities (including particles coated with Zinc
52
Fig. 3-10. Stilling well used for the rectangular flume facility
Sulfide) were used to observe the flow directions and complex flow patterns of labyrinth
weirs during testing. Fig. 3-15 is an image of the dye wand being used during testing.
Geometric measurements of the test facilities and physical models were made with a steel
measuring tape (±0.75 mm) and digital calipers (±0.0013 mm).
Physical Models
Materials, Fabrication, and Installation
A high-performance sonolastic sealant (NP1) was used to seal all joints, ensuring
53
Fig. 3-11. Stilling well used for the reservoir test facility
Fig. 3-12. Carriage and point gage system in the rectangular flume
54
Fig. 3-13. Straight and hooked point gages for nappe profiling
Fig. 3-14. Velocity field mapping with 2-D acoustic doppler velocimeter
55
Fig. 3-15. Flow pattern and direction observations with the dye wand
that each physical model and test facility was water tight. This sealant is grey in color;
for small, highly visible locations that required sealing (e.g., screw holes, etc.) a clear,
high-grade silicon sealant was used in conjunction with 102-mm wide clear tape
(wrestling mat tape) to improve aesthetics for visual documentation.
All labyrinth weir models were fabricated in-house, using high density
polyethylene (HDPE). The stock material was purchased as 1.2 m x 2.4 m (4 ft x 8 ft)
sheets in 19.1 mm, 25.4 mm, and 38.1 mm thicknesses. The thermal contraction of this
material was tested and documented to maintain consistency (details found under the
section on Test Procedure) during experimental testing (water temperatures from the
Logan River generally range from ~0.5°C to 10.5°C).
Fabrication took place at the UWRL machine shop. The material was first cut
into sections with a table saw. The material was next planed to thickness by a high speed
56
industrial planer and checked with digital calipers. The bottom and top edges of the
HDPE sections were sent through a jointer and a shaper table to smooth, square, parallel
edges. The crest was machined with a shaper table; the apexes were machined using an
industrial mill. The angles for the sidewalls were cut with an industrial compound miter
saw. Grooves were machined into each weir joint to accommodate extra NP1 sealant to
ensure a watertight seal. Drilling for fasteners was accomplished using a drill press.
UWRL cranes and/or forklifts were used to carefully transport and install the
labyrinth weirs. The fabrication, assembly, and installation of each labyrinth weir model
were strictly monitored to minimize fabrication errors that would be greatly magnified at
a prototype scale. Assembly tolerances were ±0.4 mm. Specific attentions were given to
the alignment of the machined crests (Fig. 3-16) and to the levelness of the crest after
installation. As previously mentioned, the crest of each labyrinth model was surveyed to
±0.4 mm.
Model Configurations
Data from 32 lab-scale trapezoidal labyrinth weir models were analyzed in this
research study. Testing included reservoir and channelized approach conditions, linear
and arced cycle configurations, normal, inverse, flush, rounded inlet, and projecting
placement scenarios (see Fig. 3-17), quarter-round (Rcrest = tw/2) and half-round crest
shapes, and nappe breakers and aeration vents (see Fig. 3-18, placement and quantity
were varied). A summary of the labyrinth weir physical models are presented in Table 3-
1. Detailed schematics are presented in Appendices A and B.

57
Fig. 3-16. The joint between the ½ apex (to be attached to flume wall) and
the weir sidewall of the 2-cycle, 6° half-round labyrinth
A new standard geometric layout for arced labyrinth weirs projecting into a
reservoir was developed; a sample schematic of an arced labyrinth weir is presented in
Fig. 3-19. It is simple to design geometrically; the centerline length for one labyrinth
cycle is kept constant between the linear and arced geometries. Also, the arc follows the
curvature of a circle, and cycles are spaced at the desired angle, θ. It allows for any
variation in sidewall angle, apex width, and cycle number.
Test Procedure
Experimental data were collected by setting a flow rate, allowing the upstream
water level to stabilize, and measuring Q and h. This is a common modeling procedure;
however, differences in Cd (some exceeding 10%) have been noted between experimental
58
Fig. 3-17. Physical model cycle configurations, weir orientations and placements
data sets from this study and Amanian (1987), Tullis, (1993), Waldron, (1994), and
Willmore (2004), shown in Fig. 3-20.
Fig. 3-20 is based upon center-line length of the crest, Lc, instead of effective
length Le, which is used in the Tullis et al. (1995) design method. Good agreement exists
between the current study and experimental data from Willmore (2004). Differences
59
(A) (B) (C)
Fig. 3-18. Aeration tube apparatus for N = 2 (A) and nappe breakers located on the
downstream apex (B) and on the sidewall (C)
Table 3-1. Physical models tested
Model α θ P Lcycle Lc-cycle/w w/P N Crest Type Orientation

() (°) (°) (mm) (cm) () () () () () ()
1 6 0 304.8 465.457 7.607 2.008 2 HR Trap Inverse
2-3 6 0 304.8 465.457 7.607 2.008 2 QR, HR Trap Normal
4-5 8 0 304.8 354.492 5.793 2.008 2 QR, HR Trap Normal
6-7 10 0 304.8 287.905 4.705 2.008 2 QR, HR Trap Normal
8-9 12 0 304.8 243.514 3.980 2.008 2 QR, HR Trap Normal
10-11 15 0 304.8 199.135 3.254 2.008 2 QR, HR Trap Normal
12 15 0 152.4 199.135 3.254 4.015 2 QR Trap Normal
13 15 0 152.4 99.567 3.254 2.008 4 QR Trap Normal
14 15 0 304.8 99.567 3.254 1.019 4 QR Trap Normal
15-16 20 0 304.8 154.810 2.530 2.008 2 QR, HR Trap Normal
17-18 35 0 304.8 98.352 1.607 2.008 2 QR, HR Trap Normal
19 6 0 203.2 307.547 7.607 2.008 5 HR Trap
Projecting
20 12 0 203.2 63.455 4.705 2.008 5 HR Trap
21-23 6 10, 20, 30 203.2 307.547 7.607 2.008 5 HR Trap Arced &
24-26 12 10, 20, 30 203.2 63.455 4.705 2.008 5 HR Trap Projecting
27 6 0 203.2 307.547 7.607 2.008 5 HR Trap
Flush
28 12 0 203.2 63.455 4.705 2.008 5 HR Trap
29 6 0 203.2 307.547 7.607 2.008 5 HR Trap Rounded
30 12 0 203.2 63.455 4.705 2.008 5 HR Trap Inlet
31-32 90 - 304.8 122.377 1.000 4.015 - QR, HR - -
between experimental data sets may be associated with model size, model construction
quality (levelness of crest, uniformity of crest profile, etc.), uniformity and degree of
60
Fig 3-19. Example schematic of standardi zed layout for arced labyrinth weirs
standardized
turbulence in the approach flow, techniques used for measuring HT and Q, the accuracy
of the crest reference, the degree to which a given flow condition has reached steady state
prior to data collection, and the accuracy of the instrumentation. A more detailed data set
comparison between the experimental results of this study and those of Willmore (2004)
and Tullis et al. (1995) is presented in Chapter 4.
To ensure that the time period for collecting a single flow measurement [steady-
state conditions were established (time period required for the upstream water level to
stabilize)] was accurate

accurate,, the time interval for the previously described modeling
procedure was extended to 60 minutes for a specific Q for each tested model; in addition
to monitoring the establishment of steady-state conditions, this extended observation
period made it possible to observe any harmonic or low frequency flow phenomena (e.g.,
61
0.7
6-degree QR Crookston
8-degree QR Crookston
7-degree QR Willmore
0.6
8-degree QR Willmore
6 degree QR Eq (Tullis 1995)
8-degree QR Eq (Tullis 1995)

0.5
8.51-degree QR Standley Lake (Tullis, 1993)
8.51-degree QR Standley Lake Corr.
9 degree QR Waldron
Cd 0.4
0.3
0.2
0.1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT /P
Fig. 3-20. Comparison of Cd for trapezoidal quarter-round

labyrinth weirs (based upon Lc)
oscillating or recurring hydraulic behaviors of the nappe). It was found that
approximately 5-7 minutes were needed to make an accurate measurement of Q and HT.
A unique component of this study is the large number of data points for each weir.
A total of 2,606 flow measurements were taken for this study. On average,
approximately 80 reliable data points comprise the Cd vs. HT/P data set for each weir
configuration, with the uncertainty of each point calculated as prescribed by Kline and
McClintock (1953). A system of checks was established, where at least 10% of the data
were checked twice to ensure accuracy and determine measurement repeatability. A
large experimental data set is obviously preferred for data analyses and deriving
representative trends and empirical equations.

62
In this study, digital photography and high-definition (HD) digital video recording
were used extensively to document the hydraulic behaviors of the tested labyrinth weirs
(including HT/P = 0.1 to 0.7). This provided: documentation, a visual resource for data
analyses, and careful observations encouraged research exploration that can be described
as ‘looking for something new’. Concise and meticulous notes for each experimental
data point also provide further insights and qualitative information (e.g., aeration,
hydrodynamics of the nappe).
For certain test conditions, profiling of the top and underside of the nappe was
conducted perpendicular to the labyrinth sidewall for the quarter-round, α = 15° labyrinth
weirs. The measurement location was not influenced by the labyrinth apexes or the
flume wall. A three-axis metal square with a veneer (±0.5 mm) was used to position the
tip of the point gage used for profiling. Profiling of the upper nappe surface began over
the labyrinth weir crest, and advanced in 10 mm increments. Underside profiling began
at the downstream face of the weir sidewall.
The dye tank and wand were used to explore the flow patterns in the reservoir and
flume facilities. The wand diameter is 9.5 mm, which caused very minor flow
disturbances. When dye explorations included wading and standing in the reservoir
facility, care was taken not to disturb the flows patterns at the location of interest.
Velocity measurements in the rectangular flume were made upstream and within
the cycles of the labyrinth weirs using the traditional 3-point method (0.2y, 0.6y, and
0.8y, measured from the water surface) for open-channel current metering. A 304-mm
geometric grid was developed for velocity measurements (six-tenths method, 0.6y) in the
reservoir facility to document the velocity field within the labyrinth weir cycles and the
63
approaching flow conditions. Velocity data were collected at a point, and the velocity
data at that point were averaged over a 30-s sample period.
The tailwater control structures located in the rectangular flume were only used to
explore the differences between local and tailwater submergence. Tailwater
submergence (see Fig. 3-21) refers to a tailwater depth (downstream of the structure) that
Fig. 3-21. Tailwater submergence for the 10° half-round trapezoidal labyrinth weir
64
is greater than the weir height P. A tailwater that exceeds the crest height but does not
increase the headwater elevation upstream of the weir or shift the flow control or critical
section is referred to as modular submergence. Local submergence differs from the
traditional tailwater-induced submergence in that local submergence is independent of the
downstream tailwater conditions. Local submergence is caused by the inflow exceeding
the local outflow capacity of the outlet cycle, resulting in a local increase in tailwater,
often above the crest elevation. The local submergence region (see Fig. 3-22) develops
downstream of the upstream apex and increases in size as weir discharge increases.
During this study, observations noted that local submergence occurred for quarter- and
half-round crest shapes. Observations also noted standing waves that exceeded the crest
elevation at high values of HT/P in the downstream labyrinth weir cycles (Fig. 3-23).
Fig. 3-22. Local submergence at an upstream labyrinth apex

65
Fig. 3-23. A labyrinth standing wave in the downstream cycle
Submergence investigations commenced with the labyrinth freely discharging.
Point gages on the rolling carriage and the stilling well were used to mark the upstream
water depth. After noting the depth, the tailwater was slowly increased until the upstream
depth was observed to change. After making visual documentation, the tailwater was
once again increased until the upstream depth surpassed the flume capacity; Visual
documentation was made and the process was repeated in reverse, concluding with the
weir freely discharging.
In order to characterize the size of nappe interference regions, Bint was developed
and physical measured using still images and 25.4-mm reference grid cells; this
interference length is illustrated in Fig. 3-24. It describes the interference region length
originating at and perpendicular to the upstream apex wall to the point where the nappe
66
Fig. 3-24. Physical representation of Bint in plan-view (A) and (C) and profile view (B)
and (D) for nappe interference regions, including reference grid
region intersects the weir crest. Depending upon the labyrinth weir geometry and the
flow conditions, the nappe interference region may include a turbulent flow region [Fig.
3-24 (D)], a local submergence region [Figs. 3-24 (C), or both [Fig. 3-24 (A)].
67
CHAPTER 4
HYDRAULIC DESIGN AND ANALYSIS OF LABYRINTH WEIRS
Abstract
A method for the hydraulic design and analyses of labyrinth weirs is presented
based upon the experimental results of physical modeling. Discharge coefficient data for
quarter-round and half-round labyrinth weirs are presented for 6° ≤ sidewall angles ≤ 35°.
Cycle efficiency is also introduced to aid in sidewall angle selection. Parameters and
hydraulic conditions that affect flow performance are discussed, including weir geometry,
nappe flow regimes, artificial aeration (vents, nappe breakers), and nappe stability.
Finally, the validity of this method is presented by comparing predicted results to data
from previously published labyrinth weir studies.
Introduction
A labyrinth weir is a linear weir that is “folded” in plan-view to increase the crest
length for a given channel or spillway width. An example of a labyrinth weir is presented
in Fig. 4-1.
There are infinite possible labyrinth weir configurations and design variations;
however, labyrinth cycles are typically placed in a linear fashion (i.e., upstream apexes
align at a common channel cross section as shown in Fig. 4-1), have a sidewall angle, α,
less than 30°, and are oriented towards the approaching flow.
A labyrinth weir is able to pass large discharges at relatively low heads compared
to traditional linear weir structures. As a result of their hydraulic performance and

68
Fig. 4-1. Labyrinth weir schematic including geometric parameters
versatility, labyrinth weirs have been placed in streams, canals, rivers, ponds, and
reservoirs as headwater control structures, energy dissipaters, flow aerators, and
spillways. Labyrinth weirs are especially well suited for spillway rehabilitation where
dam safety concerns, freeboard limitations, and a revised and larger probable maximum
flow have required modification or replacement of the spillway. The recently constructed
Lake Brazos spillway, Texas, USA, is such an example (Vasquez et al. 2007).
69
Flow Characteristics
The geometry of a labyrinth weir causes complex 3-dimensional flow patterns.
At very low heads, a labyrinth weir behaves similar to a linear weir (α = 90°) of
equivalent length oriented normal to the flow direction. However, as the driving head
increases, flow efficiency begins to decline, nappe interference appears, local
submergence regions develop, the air cavities under the nappe become very dynamic, and
for certain flow conditions and geometries, the nappe itself can become unstable. In the
past, physical models have proven to be highly useful for designing and analyzing
specific labyrinth weir designs.
Previous Studies
Labyrinth weir head-discharge relationships have been described by several
different empirical equations. These relationships vary based on different characteristic
weir lengths and driving head definitions (e.g., the inclusion of the velocity component
V2/2g, where g is the acceleration constant of gravity). However, the basic equation
developed for linear weirs is proposed, which includes total head upstream measured
relative to the crest, HT, and utilizes centerline length of the crest as the characteristic
length [Eq. (4-1)].
2 32
Q= C d Lc 2 g H T (4-1)
3
In Eq. (4-1), Q is the discharge of a labyrinth weir, Cd is a dimensionless discharge
coefficient, g is the acceleration constant of gravity, and HT is defined as HT = V2/2g + h
(V is the average cross-sectional velocity at the gauging location, h is the piezometric
head upstream of the weir).

70
Several earlier labyrinth weir studies resulted in published design methods; a
selection is presented here and discussed. Hay and Taylor (1970) presented parameter
guidelines for sharp-crested triangular and trapezoidal labyrinth weirs. Discharge rating
curves for h/P < 0.6 were presented in terms of a labyrinth-to-linear weir discharge ratio
(based on a common channel width and h), requiring discharge information for a linear
weir (α = 90°) of equivalent height (P), wall thickness (tw), and crest shape. The Bureau
of Reclamation (USBR) conducted model studies to aid in the design of Ute Dam
(Houston 1982). Discrepancies found between the experimental results and the
recommendations by Hay and Taylor (1970) were attributed to different definitions of
upstream head (h, Hay and Taylor (1970); HT, USBR) and limited scope of geometric
variation. From the physical model studies of Ute Dam and Hyrum Dam, Hinchliff and
Houston (1984) developed new design guidelines. Despite scope limitations, they
provided valuable insights regarding labyrinth weir orientation and placement in reservoir
and channel applications.
Based upon model studies of Avon and Woronora Dam, Darvas (1971) simplified
labyrinth weir design by introducing an empirical discharge equation and a discharge
coefficient to accompany the discharge rating curves for labyrinth weirs. However,
Magalhães and Lorena (1989) juxtaposed this method with their own experimental results
for a truncated ogee or WES crest shape, and reported their curves to be systematically
lower than Darvas’ (1971).
Lux and Hinchliff (1985) and Lux (1984, 1989) developed a new empirical
equation, which includes the cycle width ratio, w/P, and an apex shape constant, k, to
determine the discharge of a single labyrinth cycle. Although this dimensionless

71
equation applies to trapezoidal and triangular weirs, the inclusion of w/P complicates the
weir equation and was limited to w/P ≥ 2.0. Similar parameter limits have been set by
other design methods that do not explicitly include w/P in the head-discharge equation.
Tullis et al. (1995) developed a design method based upon the standard weir
equation [Eq. (4-1)] and research conducted at the Utah Water Research Laboratory
(UWRL) by Waldron (1994), Tullis (1993), and Amanian (1987). Tullis et al. (1995)
introduced an effective weir length, Le, as the characteristic weir length (instead of
channel width, W, or Lc) to define the discharge coefficient for trapezoidal, quarter-round
labyrinth weirs; Le was intended to account for apex influences on discharge efficiency.
Two significant contributions of this study were: the design method is presented as a
table to be used in a spreadsheet program, and the design curves include a linear weir
discharge curve that is useful for determining the hydraulic benefits of a labyrinth weir
relative to a linear weir. This design method is favored by Falvey (2003); however, the α
= 6° data are significantly lower than the adjacent curves and Willmore (2004) has noted
the following discrepancies: the α = 8° data falls above the α = 9° presented by Waldron
(1994), and a minor mathematical error was found in the geometric calculations. The
supporting data for this method (quarter-round crest shape) is limited to 6° ≤ α ≤18° and
provides linearly interpolated curves for α = 25° and 35°.
Recently, Melo et al. (2002) expanded the work of Magalhães and Lorena (1989)
by adding an adjustment parameter, kθ-CW, for labyrinth weirs located in a channel with
converging sidewalls. Tullis et al. (2007) developed a dimensionless submerged head-
discharge relationship (tailwater submergence) for labyrinth weirs that was verified by
Lopes et al. (2009).

72
Design methods are useful tools for determining the hydraulic performance of
labyrinth weirs and can be used to estimate and extrapolate the performance for labyrinth
weir geometric configurations or operating conditions not included in available labyrinth
information. For example, the Tullis et al. (1995) design method was recently used (a
spreadsheet presented by Falvey 2003) to design the emergency spillway for Boyd Lake,
located in Loveland, Colorado, USA. The spillway width is nearly 400 m; the labyrinth
weir length is ~2.30 km, features 59 cycles (N), α = 8°, and has a maximum discharge
capacity of 1,200 cms. This labyrinth weir features notched apexes for passing base-
flows, which is not included in the design method (Brinker 2005).
The purpose of this study is to provide new insights into the performance and
operation of labyrinth weirs and to improve the design and evaluation tools currently
available. This is to be accomplished by utilizing the experimental results from physical
modeling to provide a design optimization program, an analysis program, and additional
hydraulic information (e.g., nappe behavior and nappe aeration). The design program, as
developed during this study, is similar to the design table presented by Tullis et al. (1995)
with the addition of a user-specified footprint size (channel width, W, and apron length,
B); it contains new data sets for quarter-round and half-round crests, utilizes Lc of the
crest as the characteristic length, and includes new and previously-published design tools,
parameters, and ratios, such as cycle efficiency (ε’), nappe behavior, aeration conditions,
aeration device placement, and tailwater submergence.
Experimental Method
Physical modeling of labyrinth weirs was conducted at the Utah Water Research
73
Laboratory (UWRL). Labyrinth weirs were fabricated from High Density Polyethylene
Plastic (HDPE) and tested in a rectangular flume (1.2 m x 14.6 m x 1.0 m). The
influence of sidewall effects is considered to be negligible, based upon the findings of
Johnson (1996). Details of the tests performed are summarized in Table 4-1. When the
outside apexes of a labyrinth weir attach to the training wall at the upstream or beginning
region of the apron, it is termed a “normal orientation” (e.g., Fig. 4-1). When said apexes
attach to the training wall at the downstream end of the apron, it is termed an “inverse
orientation.”
Model test flow rates were metered using calibrated orifice meters in the flume
supply piping, differential pressure transducers, and a data logger. The flume was
equipped with a headbox and baffle to create uniform and tranquil approach conditions, a
stilling well, and a rolling instrument carriage. The point gauge instrumentation was
carefully referenced to the crest of the labyrinth. The labyrinth weirs were installed on an
Table 4-1. Physical model test program
Model α P Lc-cycle Lc-cycle/w w/P N Crest Type Orientation†

() (°) (mm) (cm) () () () () () ()
1 6 304.8 465.457 7.607 2.008 2 HR Trap Inverse
2-3 6 304.8 465.457 7.607 2.008 2 QR, HR Trap Normal
4-5 8 304.8 354.492 5.793 2.008 2 QR, HR Trap Normal
6-7 10 304.8 287.905 4.705 2.008 2 QR, HR Trap Normal
8-9 12 304.8 243.514 3.980 2.008 2 QR, HR Trap Normal
10-11 15 304.8 199.135 3.254 2.008 2 QR, HR Trap Normal
12 15 152.4 199.135 3.254 4.015 2 QR Trap Normal
13 15 152.4 99.567 3.254 2.008 4 QR Trap Normal
14 15 304.8 99.567 3.254 1.019 4 QR Trap Normal
15-16 20 304.8 154.810 2.530 2.008 2 QR, HR Trap Normal
17-18 35 304.8 98.352 1.607 2.008 2 QR, HR Trap Normal
19-20 90 304.8 122.377 1.000 4.015 - QR, HR - -
†Linear configuration was used for all model orientations
74
elevated horizontal apron with a ramped upstream floor transition. Weirs were tested
with and without a nappe aeration apparatus consisting of an aeration tube for each
labyrinth sidewall. Several different apparatus’ were required during testing to
accommodate the range of labyrinth weir geometries, an example is presented in Fig. 4-2.
The test program also evaluated the performance of wedge-shaped nappe breakers in a
variety of locations (upstream apex, weir sidewall, downstream apex).
Experimental data were collected under steady-state conditions. Q measurements
were recorded for 5 to 7 minutes with the data logger to determine an average flow rate,
and h was determined with the stilling well equipped with a point gage accurate to ±0.15
mm. Velocity data were measured inside the weir cycles with a 2-dimensional acoustic
doppler velocity probe. Also, a dye wand was used to observe the unique and complex
local flow patterns associated with labyrinth weir flow. Digital photography and high-
definition (HD) digital video recording were used extensively to document the hydraulic
behaviors of the tested labyrinth weirs. Observations also noted nappe aeration
conditions and behavior, nappe stability, nappe separation point, nappe interference, areas
Fig. 4-2. Aeration tube apparatus

75
of local submergence, and any harmonic or recurring hydraulic behaviors for all α tested.
In an effort to accurately characterize the labyrinth weir behavior, a large number
of head-discharge data points were collected for all tested weir geometry. Also, a system
of checks was established wherein at least 10% of the data were repeated to ensure
accuracy and determine measurement repeatability.
Experimental Results
Discharge Rating Curves
The general weir equation [Eq. (4-1)] was selected to determine the discharge of
labyrinth weirs; the required characteristic length is Lc. Cd is influenced by weir
geometry (e.g., P, tw, A, α, crest shape), flow conditions (HT, approaching flow angle,
local submergence, nappe interference), and aeration conditions of the nappe (clinging,
aerated, partially aerated, drowned). Accurate Cd values and corresponding hydraulic
conditions are critical for accurate labyrinth weir analyses and design. Cd data are
presented in terms of HT/P for non-vented trapezoidal labyrinth weirs (normal or inverse
weir orientations) for 6° ≤ α ≤ 35° in Fig. 4-3 (quarter-round crest shape) and Fig. 4-4
(half-round crest shape). Data for α = 90° (linear weirs) are included for comparison.
In Fig. 4-3, the α = 12° discharge coefficient data was slightly more efficient than
the adjacent curves (α ≥ 15°) for HT/P ≤ 0.085. A similar phenomenon can be seen in the
experimental data for Tullis et al. (1995) at very low heads. Also, the air cavity behind
the nappe abruptly disappeared at ~ 0.25 HT/P for the α = 35° and α = 20° (to a lesser
extent) causing a slight increase in efficiency.
In Fig. 4-4, the sharp decrease in weir efficiency, caused by the weirs shifting out
76
Fig. 4-3. Cd vs. HT/P for quarter-round trapezoidal labyrinth weirs
of the clinging nappe aeration regime, is clearly visible for 12° ≤ α ≤ 20° (e.g., HT/P ~
0.38 for α = 20°). The decrease in Cd is less dramatic but nevertheless present for all
tested half-round weirs.
For convenience in applications, the labyrinth weir Cd data in Figs. 4-3 and 4-4
were curve-fit per Eq. (4-2), and the corresponding coefficients are presented in Tables 4-
2 and 4-3. Eq. (4-3) was used for α = 90° data, and the corresponding coefficients are
presented in the aforementioned tables. The curves have been validated for 0.05 ≤ HT/P
< 0.9; however, the data are well behaved and the curves have been extrapolated. This
extrapolation has only been verified for the α ≤ 15°; model 17 (Table 4-1) was tested to
HT/P = 1.993.
77
Fig. 4-4. Cd vs. HT/P for half-round trapezoidal labyrinth weirs
 HT C 
 B* 
H 
 P 
Cd = A * T +D
P Labyrinth Weirs (4-2)
1
Cd(90 o ) = +D
HT C Linear Weirs (4-3)
A+ B* +
P HT P
A comparison between the half- and quarter-round experimental data is presented
as the ratio of the half-round over the quarter-round Cd values (Cd-HR/Cd-QR) versus HT/P
in Fig. 4-5.
A crest that is rounded on the downstream face helps the flow stay attached
(clinging flow) to the weir wall, thus increasing flow efficiency. If the flow detaches
(momentum, debris, etc.), the gains in efficiency are lost. Further gains in efficiency can
78
Table 4-2. Curve-fit coefficients for quarter-round labyrinth and linear weirs,
validated for 0.05 ≤ HT/P < 0.9
α A B C D
6° 0.02623 -2.681 0.3669 0.1572
8° 0.03612 -2.576 0.4104 0.1936
10° 0.06151 -2.113 0.4210 0.2030
12° 0.09303 -1.711 0.4278 0.2047
15° 0.10890 -1.723 0.5042 0.2257
20° 0.11130 -1.889 0.5982 0.2719
35° 0.03571 -3.760 0.7996 0.4759
90° -2.3800 6.476 1.3710 0.5300
Table 4-3. Curve-fit coefficients for half-round labyrinth and linear weirs,
validated for 0.05 ≤ HT/P < 0.9
α A B C D
6° 0.009447 -4.039 0.3955 0.1870
8° 0.017090 -3.497 0.4048 0.2286
10° 0.029900 -2.978 0.4107 0.2520
12° 0.030390 -3.102 0.4393 0.2912
15° 0.031600 -3.270 0.4849 0.3349
20° 0.033610 -3.500 0.5536 0.3923
35° 0.018550 -4.904 0.6697 0.5062
90° -8.60900 22.650 1.8120 0.6375
be obtained by using an ogee-type crest [modified half-round crest with an upstream
radius of tw = 1/3 and a downstream radius of tw = 2/3 (Willmore 2004)]. Brazos Dam
features this crest geometry; however, after construction was completed, algae growth on
the crest caused the nappe to detach, thereby reducing the hydraulic benefits of the crest
shape. The curves are not perfectly smooth due to slight variations in the experimental
data. As HT/P increases, the advantages of the improved half-round crest begin to
diminish; all Cd-HR/Cd-QR curves should eventually converge to 1.0.
Nappe Aeration Behavior and Stability
The behavior of the nappe and the air cavity behind the nappe influences the
79
1.20
6 degree
8 degree
10 degree
12 degree
1.15
15 degree
20 degree
35 degree
90 degree
Cd-HR / Cd-QR
1.10
1.05
1.00
0.95
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 4-5. Comparison of half-round and quarter-round crest shape

on labyrinth weir hydraulic performance
discharge efficiency. Four different aeration conditions (shown in Fig. 4-6) were
observed during labyrinth weir testing: clinging, aerated, partially aerated, and drowned.
The aeration condition is influenced by the crest shape, HT, the depth and turbulence of
flow behind the nappe, the momentum and trajectory of the flow passing over the crest,
and the pressure behind the nappe (sub-atmospheric for non-vented or atmospheric for
vented nappes). As HT increases, a labyrinth weir will transition from clinging to aerated,
to partially aerated, and finally to drowned. All four aeration conditions do not
necessarily occur for all labyrinth weir geometries.

80
Fig. 4-6. Nappe aeration conditions: clinging (A), aerated (B),

partially aerated (C), and drowned (D)
Sub-atmospheric pressures develop on the downstream face of the weir when the
nappe is clinging, and can exist for an aerated nappe that is not vented. An aerated nappe
will transition to a partially aerated nappe when the air cavity behind the nappe becomes
unstable; the air cavity varies spatially and temporally. The behavior of a partially
aerated nappe can be characterized as follows: the air cavity oscillates between labyrinth
weir apexes, increasing or decreasing the length of sidewall that is aerated; the air cavity
may repeatedly be completely removed and then replaced as the turbulent water surface
behind the nappe fluctuates. An unstable air cavity also causes fluctuating pressures at
the downstream face of the weir. Finally, a drowned nappe will occur at relatively larger
81
HT/P and can be characterized by a large, thick nappe with no air cavity. Table 4-4
presents the range of HT/P that was observed for each nappe aeration condition for
quarter-round and half-round labyrinth weirs.
Table 4-4. Nappe aeration conditions and corresponding

ranges of HT/P for labyrinth weirs
Quarter-Round (HT/P) Half-Round (HT/P)

α Partially Partially
Clinging Aerated Drowned Clinging Aerated Drowned
(°) Aerated Aerated
6 <0.050 0.051-0.256 0.256-0.319 >0.319 <0.165 0.165-0.298 0.298-0.405 >0.405
8 <0.050 0.057-0.288 0.288-0.364 >0.364 <0.165 0.165-0.312 0.312-0.465 >0.465
10 <0.050 0.061-0.293 0.293-0.479 >0.479 <0.219 0.219-0.283 0.283-0.505 >0.505
12 <0.050 0.061-0.275 0.275-0.510 >0.510 <0.250 - 0.250-0.530 >0.530
15 <0.050 0.052-0.256 0.256-0.508 >0.508 <0.306 - 0.306-0.560 >0.560
20 <0.050 0.053-0.240 0.240-0.515 >0.515 <0.363 - 0.363-0.599 >0.599
35 <0.050 0.059-0.232 0.232-0.515 >0.515 <0.411 0.140-0.185 0.411-0.460 >0.460
A phenomenon, which is referred to herein as nappe instability, was also observed
during testing. Nappe instability is characterized by a nappe whose trajectory oscillates
(temporal variations) and is often accompanied by changes in the nappe aeration
condition (e.g., clinging, aerated, partially aerated, drowned) at a given weir flow rate.
Such instabilities are low frequency phenomena that are accompanied by an audible
flushing noise caused by the formation and removal of air behind the nappe. At higher
flow rates, air cavity formation and nappe instability diminish. The ranges of HT/P where
instability occurred are provided in Table 4-5. It is suggested that these ranges be
avoided, as vibrations, pressure fluctuations, and noise may reach sufficient levels as to
be undesirable or harmful.
Nappe Ventilation
Artificial aeration was found to greatly improve nappe instability and decrease
82
Table 4-5. Unstable nappe operation conditions for labyrinth weirs
α Quarter-Round Half-Round
(°) (HT/P) (HT/P)
6 none none
8 none none
10 none 0.325-0.326
12 0.300-0.350 0.329-0.385
15 0.271-0.468 0.332-0.577
20 0.223-0.530 0.363-0.599
35 0.215-0.700 0.411-0.460
noise; however, the phenomenon was still observed (to a lesser degree) for α ≥ 20°. Also,
the colliding nappes at the upstream apex did not allow air to be passed from one sidewall
to the adjacent sidewall, which has a direct influence on the placement of aeration
devices.
Aeration vents were found to have little to no effect on the discharge capacity of
quarter-round labyrinth weirs, but they decreased flow capacity for half-round labyrinth
weirs at lower HT/P values by reducing the range over which the clinging nappe is
present, effectively undermining the purpose of a half-round crest. Aeration vents should
be provided for each labyrinth weir sidewall. Aeration vents placed near the downstream
apex were found to be less effective.
It is suggested that nappe breakers with a triangular cross-section be placed on the
downstream apexes with the point oriented into the flow, as shown in Fig. 4-7. The
leading edge should be protected to minimize the potential damage from debris impact.
This orientation produced no measurable reduction in the labyrinth weir discharge
capacity as was seen with aeration vents. Also, the number of required breakers is
minimized when placed on the downstream apexes, which also minimizes the number of
locations where debris may be collected. The orientation of the streamlines passing over
83
Fig. 4-7. Wedge-shaped nappe breakers placed on the downstream apex
the weir sidewall change with flow rate; therefore, nappe breakers placed on the weir
sidewall can only be oriented into the flow for a specific flow condition. When the nappe
breaker is not oriented into the flow, it acts as an obstruction on the crest and decreases
the efficiency of the weir, even though it continues to aerate the nappe.
Labyrinth Design and Analyses
The recommended procedure for designing a labyrinth weir is presented as Table
4-6. The top section of the design table includes the user-defined hydraulic conditions or
requirements for the labyrinth weir. For example, Qdesign may be a flood event
determined from a hydrologic analysis that the labyrinth spillway must pass, HT will be
based upstream flood plain constraints and Hd might be determined by a backwater curve
flow profile analysis for Qdesign. Weir geometric parameters are entered into the next
section of the design table to begin optimizing the labyrinth weir layout for a given
footprint size and weir height. Though not tied specifically to any calculations in the
design method, a place is provided in the design table to specify a nappe aeration device
if desired. In the third section of the table, the weir geometry and hydraulic performance
84
Table 4-6. Recommended design procedure for labyrinth weirs
Parameter Symbol Value Units Notes

Hydraulic Conditions – Input Data
Design Flow Qdesign 1,500.00 (m3/s) Input g = 9.81 m/s2
Design Flow Water Surface Elevation H 1,680.00 (m) Input
Approach Channel Elevation Hapron 1,675.00 (m) Input
Crest Elevation Hcrest 1,678.00 (m) Input
Unsubmerged Total Upstream Head HT 2.00 (m) Input (Piesometric Head + Velocity Head - Losses
Downstream Total Head Hd 0.50 (m) Input
Labyrinth Weir Geometry – Input Data
Angle of Side Legs α 12 (°) α ~ 6° - 35°
Width of Labyrinth (Normal to Flow) W 99.87 (m) Input or W = Nw
Length of Apron (Parallel to Flow) B 22.35 (m) Input or B = [Lc/(2N)-(A+D)/2]cos(α)+tw
Crest Height P 4 (m) P ~1.0HT
Thickness of Weir Wall at the Crest tw 0.50 (m) tw ~ P/8
Inside Apex Width A 0.50 (m) A ~ tw
Crest Shape Crest Shape Quarter - Quarter- or Half-Round
Aeration Device (Nappe Breakers, Vents) - Breakers - Breakers, Vents, or None
Calculated Data
Headwater Ratio HT/P 0.50 -
Labyrinth Weir Discharge Crest Coefficient Cd(α°) 0.43 - Cd(α°) = f(HT/P, α, Crest Shape)
Total Centerline Length of Weir Lc 418.24 (m) Lc = 3/2Qdesign/[(Cd(α°)HT3/2)(2g)1/2]
Centerline Length of Sidewall lc 2.33 (m) lc = (B-tw)/cos(α)
Number of Cycles N 9 - W/w or Input
Cycle Width w 11.10 (m) w = 2lcsin(α)+A+D
Outside Apex Width D 1.30 (m) D = A+2twtan(45-α/2)
Magnification Ratio M 4.19 - M = Lc/(wN)
Cycle Width Ratio w/P 2.77 - Normally 2 ≤ w/P ≤ 4
Relative Thickness Ratio P/tw 0.13 - ~8
Apex Ratio A/w 0.05 - <0.08
Cycle Efficiency ε’ 1.80 - ε’ = Cd(α°)M
Efficacy ε 2.23 - ε = Cd(α°)M/Cd(90°)
# of Nappe Breakers or Vents - 9 - Breaker on ds Apex, 1 Vent per Sidewall
Linear Weir Discharge Coefficient Cd(90°) 0.81 - Cd(90°) = f(HT/P, α, Crest Shape)
Length of Linear Weir for same Flow Lc(90°) 222.33 (m) Lc(90°) = 3/2Qdesign/[(Cd(90°)HT3/2)(2g)1/2]
Submergence (Tullis et al. 2007)
Downstream/Upstream Ratio of Unsubmerged Head Hd/HT 0.25 (m)
Submerged Upstream Total Head H* 1.014 (m) Piecewise function Tullis et al. (2007)
Submergence Level S 0.49 -
Submerged Head Discharge Ratio H*/HT 0.51 -
Submerged Weir Discharge Coefficient Cd-sub 0.22 - Cd(α°)(H*/HT)
†Design limited to extent of experimental data; designs that exceed these limits may warrant a physical model study
are calculated based on previously defined geometric parameters and the head-discharge
requirements. If desired, N and B can be switched between independent and dependent
variables from the equations provided (a minor adjustment to the table). For comparison,
85
Cd(90°) and the required weir length to match the design head-discharge condition are
reported. The last section of the design method includes the submerged head-discharge
relationships developed by Tullis et al. (2007). This design method can be conveniently
implemented in a spreadsheet-based computer program.
Per Figs. 4-3 and 4-4, Cd decreases with decreasing α. For a given footprint size
(W and B held constant); however, labyrinth weir crest length increases with decreasing
α. Both of these factors should be considered when trying to optimize a labyrinth weir
design based on discharge capacity, as increasing the weir length compensates for
reduction in flow efficiency with decreasing α. To aid in the selection of α, cycle
efficiency, ε’ (ε’=Cd Lc-cycle/w), which is representative of the discharge per cycle, is
presented in Figs. 4-8 (quarter-round) and 4-9 (half-round) as a function of HT/P. These
figures show that the maximum ε’ values occur at relatively low HT/P (as delineated by
the dashed line); discharge per cycle or ε’ increases as α decreases; and the benefits of
smaller α angles decrease with increasing HT/P. Cycle efficiency maintains a constant
cycle width, w, and does not consider additional factors that influence cycle geometry
such as apron length and construction costs associated with an increase in weir length.
Beyond the ability to design a labyrinth weir for a particular flow rate, per Table
4-6, the ability to determine the head-discharge characteristics for a specified labyrinth
weir geometry (e.g., an existing structure) is also important. Such a procedure, which
also easily adapts to a spreadsheet computer program, is outlined in Fig. 4-10. The
known labyrinth weir geometries are entered. Missing geometric parameters and
labyrinth weir ratios are calculated, and a head-discharge rating curve is produced. The
effects of tailwater submergence may be determined by solving for Q or H* by iteration.

86
4.5
6 degree QR 8 degree QR 10 degree QR 12 degree QR
4.0 15 degree QR 20 degree QR 35 degree QR 90 degree QR
3.5
3.0
ε'=CdLc-cycle/w
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 4-8. Cycle efficiency vs. HT/P for quarter-round labyrinth weirs
The dimensionless submerged head relationship, based on Tullis et al. (2007), is
presented in Fig. 4-11, which includes tailwater submergence levels (S).
The design method, design charts, and corresponding curves are limited to the
geometries (Tables 4-1 and 4-6) and hydraulic conditions tested in this study (e.g., 0.05 ≤
HT/P ≤ 0.9). However, these results can be conservatively applied (with sound
engineering judgment) to other labyrinth weir geometries and flow conditions
(differences may merit a hydraulic model study). For example, there was no discernable
performance difference between the normal and inverse oriented α = 6° labyrinth weirs
(data not presented); therefore, these results may be applied to either weir orientation.
Models 16 and 17 were tested to HT/P of 1.5 and 2.0, respectively, and were found to
87
4.5
6 degree HR 8 degree HR 10 degree HR 12 degree HR
4.0 15 degree HR 20 degree HR 35 degree HR 90 degree HR
3.5
3.0
ε'=CdLc-cycle/w
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 4-9. Cycle efficiency vs. HT/P for half-round labyrinth weirs
closely agree with Eq. (4-2). Consequently, these results may be extrapolated to larger
flow rates (HT/P ≤ 2.0). Finally, linear interpolation is recommended to determine Cd for
α values other than those presented.
Data Verification
An uncertainty analysis was performed as outlined by Kline and McClintock
(1953) for single-point experimental data. The resulting maximum and average (%)
uncertainties in the Cd data are presented in Table 4-7. Single sample uncertainties were
largest for very small values of Q and h (instrument readability, HT/P ≤ 0.075) and
smallest for large values of Q and h.
Differences may exist between the experimental data sets of different researchers.
88
Fig. 4-10. Recommended procedure for labyrinth weir analyses
To verify the experimental results obtained in this study, several comparisons were made.
First, a comparison was made with the experimental results of non-vented, half-round
labyrinth weirs from Willmore (2004), shown in Fig 4-12. There is good agreement for
all weir geometries, with the largest discrepancy appearing for large α weirs at HT/P ≤
0.2.
The second comparison (shown in Fig. 4-13, and in terms of Lc) was made with
the Tullis et al. (1995) design method. There appears to be relatively good agreement at
large values of HT/P; however, large differences are visible for HT/P ≤ 0.4. This may be
attributed to the smaller sized labyrinth weir models used by the Tullis et al. (1995)
method (potential size scale effects and different values of P), a higher level of
89
S 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 0.95 1.0
5.0
4.5
4.0
3.5
3.0
H*/HT
2.5
2.0
1.5
1.0
0.5
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
Hd/HT
Fig. 4-11. Dimensionless submerged head relationship for labyrinth weirs

(based on Tullis et al. 2007)
Table 4-7. Representative single sample uncertainties for the

tested labyrinth and linear weirs
α Quarter-round α Half-round
(°) Min (%) Avg (%) Max (%) (°) Min (%) Avg (%) Max (%)
6 1.399 2.099 4.487 6 1.354 2.148 7.064
8 1.143 1.657 3.940 8 1.099 1.494 3.731
10 0.979 1.656 5.678 10 0.947 1.407 4.453
12 0.875 1.552 4.375 12 0.843 1.301 3.963
15 0.773 1.357 4.121 15 0.750 1.153 4.494
20 0.675 1.207 4.015 20 0.643 1.046 3.558
35 0.544 0.980 3.564 35 0.529 0.945 4.147
90 0.398 0.906 4.484 90 0.413 0.717 2.513
uncertainty in flow measurement, and maintaining a constant B instead of N, resulting in
cycle fragments. The linearly interpolated α = 25° and α = 35° curves based on the
experimental data (α = 18° and α = 90°) also do not agree. Willmore (2004) found the α
90
0.9
0.8
0.7
0.6
0.5
Cd
0.4
0.3
0.2
6 degree HR Crookston 8 degree HR Crookston 10 degree HR Crookston 12 degree HR Crookston
15 degree HR Crookston 20 degree HR Crookston 35 degree HR Crookston 0.05

0.1
7 degree HR Willmore 8 degree HR Willmore 10 degree HR Willmore 12 degree HR Willmore
15 degree HR Willmore 20 degree HR Willmore 35 degree HR Willmore

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 4-12. Comparison between Cd values obtained by Willmore (2004) and the present
study for non-vented, half-round labyrinth weirs
= 8° data from Tullis et al. (1995) to be in error; therefore, it is replaced with
experimental data by Willmore, which shows excellent agreement to the present study.
Consequently, the author proposes that quarter-round Cd curves presented herein replace
the Tullis et al. (1995) quarter-round design curves.
The final comparison is made with data from the 13 physical model studies for
prototype labyrinth weir structures, compiled from a variety of sources, given in Table 4-
8. Where possible, the original document was consulted for the most accurately reported
weir geometries and tested hydraulic conditions, as some of this information can be found
in multiple sources. The agreement between Cd values calculated from this design
91
0.9
0.8
0.7
0.6
0.5
Cd
0.4
0.3
0.2
6 degree QR Crookston 8 degree QR Crookston 10 degree QR Crookston 12 degree QR Crookston
0.1 15 degree QR Crookston 20 degree QR Crookston 35 degree QR Crookston 90 degree QR Crookston
6 degree QR Tullis et al. 1995 8 degree QR Willmore 12 degree QR Tullis et al. 1995 15 degree QR Tullis et al. 1995
18 degree QR Tullis et al. 1995 25 degree QR Tullis et al. 1995 35 degree QR Tullis et al. 1995 90 degree QR Tullis et al. 1995
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 4-13. Comparison between proposed Cd design curves by

Tullis et al. (1995), Willmore (2004), and the present study for
non-vented, quarter-round labyrinth weirs (based upon Lc)
method [Table 4-6 with Eq. (4-2)] and predicted Cd values [Eq. (4-1)] from the prototype
structures are also presented in Table 4-8. From the results, it is clear that there is good
agreement between the proposed design method (Table 4-6) and the reported model
studies. However, there are varying levels of agreement for multiple HT/P values for a
single structure, indicating the presence of sources of uncertainty associated with physical
modeling, such as different model sizes, experimental methods, entrance configurations,
and others. For Table 4-8, an average difference of 2.7% with a standard deviation of
6.12% was calculated.

92
Table 4-8. Comparison between the proposed design method and results obtained from
hydraulic model tests for labyrinth weir prototypes
Q HT/P α N Cd Cd Diff.
Name Source (m^3/s) () (°) () Eq. (4-1) T. 4-6 (%)
Magalhães & 14.4 0.200 18.00 1 0.6571 0.6159 6.48%
1 Alijó
Lorena (1989) 39.7 0.400 18.00 1 0.6386 0.5528 14.39%
1790.0 0.932 27.50 10 0.4867 0.4590 5.88%
2 Avon Darvas (1971)
1415.8 0.720 27.50 10 0.5645 0.5119 9.77%
387.0 0.652 19.44 2 0.4995 0.4937 1.16%
3 Boardman Babb (1976)
386.8 0.507 18.21 2 0.5129 0.5381 -4.80%
576.0 0.686 15.20 4 0.4542 0.4144 9.18%
120.7 0.200 15.20 4 0.6041 0.6001 0.67%
Magalhães &
4 Dungo 303.1 0.400 15.20 4 0.5364 0.5223 2.66%
Lorena (1989)
491.9 0.600 15.20 4 0.4739 0.4430 6.74%
576.0 0.558 15.20 4 0.4542 0.4583 -0.89%
148.0 0.528 19.00 2 0.5508 0.5127 7.17%
Magalhães & 148.0 0.440 19.00 2 0.5508 0.5470 0.69%
5 Gema
Lorena (1989) 41.5 0.200 19.00 2 0.6625 0.6210 6.46%
114.1 0.400 19.00 2 0.6438 0.5626 13.45%
350.0 0.543 15.20 3 0.5208 0.4641 11.52%
Magalhães &
6 Harrezza 350.0 0.442 15.20 3 0.5208 0.5046 3.16%
Lorena (1989)
220.8 0.400 15.20 3 0.5195 0.5223 -0.54%
Houston 256.3 0.458 9.85 2 0.4097 0.3990 2.63%
7 Hyrum
(1983) 254.0 0.500 9.73 2 0.3564 0.3785 -6.02%
Magalhães & 250.0 0.703 14.90 2 0.4078 0.4053 0.63%
8 Keddara
Lorena (1989) 250.0 0.586 14.90 2 0.4078 0.4442 -8.54%
CH2M-Hill 239.0 0.400 13.00 4 0.4649 0.4887 -5.01%
9 Mercer
(1973) 135.4 0.233 13.37 4 0.5892 0.5716 3.04%
94.3 0.400 13.30 2 0.5066 0.4935 2.63%
São Magalhães &
10 157.9 0.600 13.30 2 0.4525 0.4134 9.02%
Domingos Lorena (1989)
160.0 0.511 13.30 2 0.4726 0.4462 5.75%
Standley
11 Tullis (1993) 1539.4 0.648 8.51 13 0.3155 0.2980 5.71%
Lake
Tullis & 959.2 0.208 11.40 7 0.6021 0.5635 6.62%
12 Townsend Crookston
2717.2 0.554 11.40 7 0.3917 0.3956 -1.01%
(2008)
15574.0 0.633 12.90 14 0.3696 0.3958 -6.85%
Houston
13 Ute 15574.3 0.650 12.15 14 0.3552 0.3780 -6.21%
(1982)
2830.7 0.147 12.15 15 0.5622 0.5996 -6.42%
Summary and Conclusions
A labyrinth weir design and analysis procedure is presented (Table 4-6) based
upon the results of physical modeling in a laboratory flume. Q is calculated based on the
traditional weir equation [Eq. (4-1)], utilizing HT and selecting the centerline length of the
93
weir, Lc, as the characteristic length. Tailwater submergence for labyrinth weirs, as
presented by Tullis et al. (2007), is included. The proposed design and analysis method
is validated by juxtaposing the experimental results of this study with other physical
model studies presented in Figs. 4-12, 4-13, and Table 4-8.
Figs. 4-3 and 4-4 present a dimensionless discharge coefficient, Cd, as a function
of HT/P for quarter-round and half-round labyrinth weirs (6° ≤ α ≤ 35°) and for linear
weirs. The test results indicate that the increase in efficiency provided by a half-round
crest shape (relative to a quarter-round crest) is more significant for HT/P ≤ 0.4.
Cycle efficiency, ε’, is a tool for examining the discharge capacity of different
labyrinth weir geometries (Figs. 4-8 and 4-9). The results of ε’ indicate how the increase
in crest length compensates for the decline in discharge efficiency associated with
decreasing α.
The experimental results indicate that nappe aeration conditions and nappe
stability should not be overlooked in the hydraulic and structural design of labyrinth
weirs. The results presented in Tables 4-4 and 4-5 indicate flow behaviors that may
include negative or fluctuating pressures at the weir wall, noise, and vibrations. These
tables also aid in the selection of a crest shape. Finally, the effects of nappe ventilation
by means of aeration vents or nappe breakers are presented, including recommended
placements of vents (one per sidewall) and breakers (one centered on each downstream
apex).
Although the methods and tools presented herein will accurately design and
analyze a labyrinth spillway, a physical model study is recommended to verify hydraulic
performance. A model study would include site-specific conditions that may be outside
94
the scope of this study and may provide valuable insights into the performance and
operation of the labyrinth weir.
Additional components of this study not presented here include arced labyrinth
weirs and various labyrinth weir orientations and placements in a reservoir, a detailed
look at nappe behavior (including local submergence, nappe interference, and nappe
stability), scale effects, and other labyrinth weir flow phenomena.

95
CHAPTER 5
ARCED AND LINEAR LABYRINTH WEIRS
IN A RESERVOIR APPLICATION
Abstract
A standard geometric design layout for arced labyrinth weirs is presented.
Insights and comparisons in hydraulic performance of half-round, trapezoidal, 6° and 12°
sidewall angles, labyrinth weir spillways are presented with the following orientations:
Normal, Inverse, Projecting, Flush, Rounded Inlet, and Arced cycle configuration.
Discharge coefficients (specific to the experimental results) as a function of HT/P,
including rating curves, are presented. Finally, approaching flow conditions and
geometric similitude are discussed and hydraulic design tools are recommended to be
used in conjunction with the hydraulic design and analysis method presented in Chapter
4.
Introduction
Many spillways utilize a type of weir as the flow control structure. The flow
capacity of a weir is largely governed by the weir length, Lc, shape of the crest, and the
conditions of the approaching flow. A labyrinth weir spillway (see Fig. 5-1) is a linear
weir folded in plan-view; these structures offer several advantages when compared to
linear weir structures. Labyrinth weirs provide an increase in crest length for a given
channel width, thereby increasing the flow capacity for a given upstream flow depth
(labyrinth weirs are typically designed for HT/P ≤ 1.0). As a result of the increased flow
capacity, these weirs require less free board than linear weirs, which better facilitates
96
Fig. 5-1. Example of a labyrinth weir spillway

97
flood routing and can also allow for higher reservoir pool elevations under base-flow
conditions (i.e., the amount of reservoir storage volume above normal pool reserved for
flood routing storage can be reduced).
The hydraulic design of a labyrinth weir spillway requires the optimization of
many geometric parameters. For example, the sidewall angle (α), total crest length (Lc),
number of cycles (N), and crest shape must be determined for a given footprint size. The
configuration of the labyrinth cycles and the orientation and placement of the weir can
also influence discharge efficiency. If the spillway is located in a chute or channel, the
labyrinth weir can have a normal or inverse orientation [Fig. 5-2 (E) and (F),
respectively]. For reservoir spillway application, the labyrinth weir may be Flush [Fig 5-
2 (C)], have rounded abutments [Rounded Inlet, Fig. 5-2 (D)], or be partially or fully
projecting [Fig. 5-2 (A)] into the reservoir. The cycle configuration may also be arced
[Fig. 2 (B)] to improve the cycle orientations to the approach flow conditions of the
reservoir and further increase the weir crest length. Arced labyrinth cycles are
characterized by the downstream cycle geometry.
The weir discharge capacity can generally be improved by optimizing the inlet
section [e.g., rounded abutment walls (Rounded Inlet)]. The relative increase in
hydraulic efficiency associated with improved abutments, however, diminishes as N
increases. The details of the downstream spillway channel must also be considered,
which include the downstream apron elevation, apron slope, tailwater elevation and
possible submergence effects, supercritical waves, and energy dissipation. For arced,
projecting labyrinth weir spillways (reservoir applications), it is possible to oversize the
labyrinth weir (i.e., weir length), relative to the discharge capacity of the spillway chute
98
Fig. 5-2. Summary schematic of tested labyrinth weir orientations
inlet. In such cases, as the spillway discharge increases, the point of flow control will
eventually shift from the labyrinth weir to the chute inlet or other possible control point
downstream, limiting the spillway discharge capacity.

99
Previous Design Methods
A selection of notable research studies that have provided hydraulic design
guidance for labyrinth weir spillways are presented in Table 5-1. The design method
presented in Chapter 4, for example, is based upon the general weir equation [Eq. (5-1)]
and presents discharge coefficient data for quarter-round and half-round labyrinth weirs
for 6° ≤ α ≤ 35°. It also includes cycle efficiency (ε’), nappe flow regimes, artificial
aeration (vents, nappe breakers), and nappe stability.
2 32
Q= C d Lc 2 g H T (5-1)
3
In Eq. (5-1), Q is the weir discharge, Cd is a dimensionless discharge coefficient,
g is the acceleration constant of gravity, and HT is the total upstream head defined as HT =
V2/2g + h (V is the average cross-sectional velocity at the upstream gauging location, and
h is the piezometric head measured relative to the weir crest elevation).
Table 5-1. Labyrinth weir design methods
Design Methods
Labyrinth Crest
() Authors
Cycle Type Shape
Triangular
1 Hay and Taylor (1970) Trapezoidal Sh, HR
Rectangular
2 Darvas (1971) Trapezoidal LQR
Triangular
3 Hinchliff and Houston (1984) Sh, QR
Trapezoidal
Lux and Hinchliff (1985) Triangular
4 QR
Lux (1989) Trapezoidal
5 Magalhães and Lorena (1989) Trapezoidal WES
6 Tullis et al. (1995) Trapezoidal QR
7 Melo et al. (2002) Trapezoidal LQR
8 Tullis et al. (2007) Trapezoidal HR
9 Lopes et al. (2008) Trapezoidal LQR
10 Chapter 4 (Crookston) Trapezoidal QR, HR
100
Labyrinth Weirs Located in a Reservoir
Many labyrinth weir design methods are based upon physical modeling conducted
in laboratory flumes, where the approaching flow field is relatively uniform and
perpendicular to the weir (e.g., Tullis et al. 1995, Magalhães and Lorena 1989, Hay and
Taylor 1970). The approaching flow for labyrinth weirs located in a reservoir, however,
may not be uniform or perpendicular to the weir; varying angles of the approach flow and
flow convergence may result in appreciable differences in weir efficiency [e.g. Prado
Spillway, Copeland and Fletcher (2000)]. There is useful but limited information
regarding the inlet section, labyrinth weir placement and orientation, non-uniform
approach conditions, and non-linear cycle configurations (curved and arced labyrinth
weirs). For example, Melo et al. (2002) presents an adjustment parameter for a labyrinth
weir with converging channel sidewalls. Also, case studies for Boardman Dam (Babb
1976) and Hyrum Dam (Houston 1983) reported that placing curved abutment walls
upstream of the labyrinth weir minimized the loss of efficiency caused by flow
separation.
The test program for Hyrum Dam (Houston 1983) included various weir
orientations and placements (Normal, Inverted, Flush, and Partially Projecting, see Fig. 5-
2) for the two-cycle labyrinth weir. For similar entrance conditions, it was reported that
the Normal orientation had a 3.5% greater discharge than the Inverted orientation, and the
Partially Projecting orientation increased discharge by 10.4% when compared to the
Flush orientation. It should be noted that these orientations featured rounded abutment
walls and that the results of this study are limited because the weir was composed of only
two cycles. A comparison of the hydraulic performance of a normal and inverse oriented
101
α = 6° labyrinth weir in a channel application is included in Chapter 4 found no change in
hydraulic performance.
Traditionally, labyrinth weir cycles follow a linear configuration [e.g., Lake
Townsend (Greensboro, North Carolina, USA)]; however, an arced cycle configuration
can increase discharge efficiency if it improves the orientation of the cycle to the
approaching flow (~90° is desirable). Falvey (2003) commented that the efficiency of
Prado Spillway could have been increased if the cycle configuration was curved to
improve alignment to the approaching flow. Avon (Darvas 1971), Kizilcapinar (Yildiz
and Uzecek 1996), and Weatherford (Tullis 1992) are examples of curved or arced
labyrinth weir spillways (physical model studies were conducted for these structures).
Recently, Page et al. (2007) conducted a study for María Cristina Dam (Castellón,
Spain). Following preliminary investigations, two labyrinth weir geometric designs for
the emergency spillway were examined: a 9-cycle labyrinth weir with 4 cycles following
an arced configuration, and a 7-cycle labyrinth weir that featured 5 arced cycles. The
physical models (1/50th scale, P~140-mm) were found to be less efficient than predicted
discharges from the Magalhães and Lorena (1989), Lux and Hinchliff (1985), and Tullis
et al. (1995) design methods. However, the 7-cycle arced configuration provided the
greatest improvement of cycle orientation alignment to the approaching flow; as a result
it was found to be the more efficient design.
The purpose of this study is to provide new insights and design information
regarding the performance and operation of arced labyrinth weirs and labyrinth weirs
located in a reservoir. Because geometric similitude requires more than geometrically
similar cycles, a layout for arced labyrinth weirs projecting into a reservoir is also
102
presented. This information is to be used in conjunction with “Hydraulic Design of
Labyrinth Weirs” (Chapter 4).
Experimental Method
Physical modeling of several labyrinth weir configurations was conducted at the
Utah Water Research Laboratory (UWRL), located in Logan, Utah, USA. Labyrinth
weirs were fabricated from High Density Polyethylene (HDPE) sheeting, featured a half-
round crest shape, and were tested in an elevated headbox (7.3 m x 6.7 m x 1.5 m deep)
and in a laboratory flume (1.2 m x 14.6 m x 1.0 m). The labyrinth weirs were installed
on an elevated horizontal platform (level to ±0.4-mm). The flume facility also featured a
horizontal elevated platform upon which the test weirs were installed and a ramped
upstream floor transition, which was reported by Willmore (2004) to have no influence
on the discharge capacity. Sidewall effects in the rectangular flume were considered to be
negligible based upon the finding of Johnson (1996). In the headbox, the discharge
channel downstream of the weir was relatively short (~10 cm) and terminated with a free
overfall to minimize any spillway chute specific tailwater effects. The radius for the
rounded inlet was set to the cycle width (Rabutment = w). Details of the labyrinth weir
spillway configurations modeled in a reservoir and channel are summarized in Table 5-2
and Fig. 5-2.
Model test flow rates were determined using calibrated orifice meters in the
supply piping, differential pressure transducers, and a data logger. Point velocity
measurements (U) were made using a 2-dimensional acoustic Doppler velocity probe.
The headbox and flume were each equipped with a plenum and a baffle located between
103
Model α‡ θ P Lc-cycle Lc-cycle/w w/P N Crest Type Orientation†

() (°) (°) (mm) (cm) () () () () () ()
1 6 0 304.8 465.457 7.607 2.008 2 HR Trap Inverse
2 6 0 304.8 465.457 7.607 2.008 2 HR Trap Normal
3 6 0 203.2 307.547 7.607 2.008 5‡ HR Trap Projecting
Arced &
4-6 6 10, 20, 30 203.2 307.547 7.607 2.008 5‡ HR Trap
Projecting
7 6 0 203.2 307.547 7.607 2.008 5 HR Trap Flush
Rounded
8 6 0 203.2 307.547 7.607 2.008 5 HR Trap
Inlet
9 12 0 304.8 243.514 3.980 2.008 2 HR Trap Normal
10 12 0 203.2 63.455 4.705 2.008 5‡ HR Trap Projecting
Arced &
11-13 12 10, 20, 30 203.2 63.455 4.705 2.008 5‡ HR Trap
Projecting
14 12 0 203.2 63.455 4.705 2.008 5 HR Trap Flush
Rounded
15 12 0 203.2 63.455 4.705 2.008 5 HR Trap
Inlet
16-17 90 - 304.8 122.377 1.000 4.015 - QR, HR - -
†Linear cycle configuration was used for all model orientations unless ‘Arced’ is specified. Normal and Inverse orientations are
specific to channel application
‡Based upon the outlet labyrinth cycles
the water supply and the test section to create relatively uniform and tranquil flow
conditions. The point gauge instrumentation was carefully referenced to the crest of the
labyrinth weir. Models were tested without any artificial nappe aeration.
Experimental data were collected under steady-state conditions. Flow
measurements were recorded for 5 to 7 minutes with the data logger to determine an
average flow rate, and h was determined using a stilling well equipped with a point gage
accurate to ±0.15 mm. A system of checks was established wherein at least 10% of the
data were repeated to ensure accuracy and determine measurement repeatability.
Velocity data measurements followed a ~30 cm grid (1 ft) and were time averaged for 30
s. A dye wand was used to make qualitative observations of the approaching flow field
and the flow passing over the labyrinth weir. The hydraulic behavior of the tested
104
labyrinth weirs was extensively documented with digital still and high-definition (HD)
video photography. Observations also noted nappe aeration conditions, nappe stability,
areas of local submergence, areas of flow convergence, wakes, and the general hydraulic
performance of each cycle.
Geometric Layout of Arced Labyrinth Weirs
The following discussion presents a standard layout and important geometric
parameters for arced labyrinth weir spillways, developed and tested in this study. The
geometric design process begins by selecting the geometry of a single labyrinth cycle.
The cycle is then repeated by following the arc of a circle, as shown in Fig. 5-3.
Fig. 5-3. Standard geometric layout for an arced labyrinth weir

W’/W Width Ratio, specific to aarced
rced labyrinth weir spillways
105
Important geometric parameters are:
W Downstream channel width
W’ Width of the arced labyrinth weir spillway, W’ = R θ
w’ Cycle width for the arced labyrinth weir spillway, w’ = W’ / N
R Arc radius, R = (W2/4 + r’2)1/2
r’ Segment height, r’ = R - r
r Arc center to channel width midpoint distance, r = R - r’
Θ Central weir arc angle, Θ = W’ / R
θ Cycle arc angle, θ = Θ / N
α Sidewall angle for labyrinth weir cycle, used for linear or arced
configurations.
α’ Upstream sidewall angle, α’ = α + θ / 2
N Number of labyrinth cycles
A Inside apex width
lc Center-line length of the sidewall
tw Weir wall thickness at crest
Hydraulic Performance
Physical modeling determined Q and HT for the half-round crested labyrinth weirs
installed in the reservoir. The discharge coefficients, Cd, were determined using Eq. (5-1).
Cd is dimensionless and is influenced by weir geometry, approach flow conditions,
aeration conditions of the nappe, and local submergence. Local submergence refers to a
location where the water surface elevation immediately downstream of the weir wall is
higher than the weir crest (e.g., the upstream apexes of a labyrinth weir at a high
106
discharge). It is caused by flow convergence, wakes, and standing waves. Local
submergence is location specific and therefore distinct from tailwater submergence,
where the tailwater elevation downstream of the weir exceeds the weir height and the
entire spillway becomes submerged. Cd data are presented in terms of the headwater
ratio, HT/P, for α = 6° (Fig 5-4) and α = 12° (Fig. 5-5); θ = 0 denote a linear cycle
configuration. Data for α = 90° (half-round crest shape) from Chapter 4 is included for
comparison.
As shown in Fig. 5-4 and Fig. 5-5, the ‘Flush’ orientation was found to be the
least efficient labyrinth weirs tested in this study, and the arced configurations were
found to be the most efficient. The increased efficiency of the arced labyrinth weirs is
attributed to the improved orientation of the cycles to the approaching flow. However,
0.9
0.8
0.7
0.6
0.5
Cd
0.4
0.3
0.2
α=6° Normal in Channel α=6° Projecting (Linear,θ=0°)
α=6° Flush α=6° Arced Projecting, θ=10°

0.1
α=6° Rounded Inlet α=6° Arced Projecting, θ=20°
α=90° (Linear Weir) in Channel α=6° Arced Projecting, θ=30°

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
HT/P
Fig. 5-4. Cd vs. HT/P for α = 6° half-round trapezoidal labyrinth weirs

107
0.9
0.8
0.7
0.6
0.5
Cd
0.4
0.3
0.2
α=12° Normal in Channel α=12° Projecting (Linear,θ=0°)
α=12° Flush α=12° Arced Projecting, θ=10°

0.1
α=12° Rounded Inlet α=12° Arced Projecting, θ=20°
α=90° (Linear Weir) in Channel α=12° Arced Projecting, θ=30°

0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8
HT/P
Fig. 5-5. Cd vs. HT/P for α = 12° half-round trapezoidal labyrinth weirs
local submergence limits the gains in discharge efficiency from an arced labyrinth cycle
configuration. Local submergence develops sooner for arced labyrinth weirs because
these geometries discharge more flow into the downstream cycles and channel than a
linear cycle configuration for a given HT. As HT increased, the portion of the labyrinth
weir cycle that was submerged also increased, resulting in a shift of the flow control
section that began at the crest and moved down the weir cycle. With sufficient HT, this
control region will eventually move to a control point in the downstream channel (e.g.,
spillway chute inlet). The limiting influence of local submergence was observed for the θ
= 30° arced labyrinth weirs at HT/P ~ 0.15 (α = 6°) and HT/P ~ 0.30 (α = 12°). The
decline in efficiency was gradual for α = 12° (Fig. 5-5) but more rapid for α = 6° (shown
in Fig. 5-4) where the upstream cycle flow area (α’ = 21°) was significantly larger.
108
Visual observations noted local submergence regions that originated at the upstream
apexes and an increase in tailwater elevation where the flows exiting the labyrinth cycles
converged.
Trend lines were fit to the Cd data in Figs. 5-4 and 5-5 per Eq. (5-2) for
convenience of use. Corresponding coefficients for 0.05 ≤ HT/P ≤ 0.2 are presented in
Table 5-3, and coefficients for 0.2 ≤ HT/P ≤ 0.7 are presented in Table 5-4.
3 2
H H H
Cd = A * T + B * T + C * T + D
P P P (5-2)
Uncertainty of the experimental Cd data was quantified from a single sample
uncertainty analysis adapted from Kline and McClintock (1953). Maximum errors
occurred at the lowest values of HT, with the error decreasing as HT increased. The
minimum, maximum, and average uncertainties (%) determined for each tested physical
model are presented in Table 5-5.
Labyrinth Weir Orientation, Placement,

and Cycle Configuration
The labyrinth weir orientations and cycle configurations tested in this study are
summarized in Fig. 5-2 and Table 5-2. Cd values from each model were juxtaposed to the
Cd values from a labyrinth weir located in a channel with a Normal orientation
(orthogonal to the stream-wise direction) to quantify differences in hydraulic efficiency.
The ratio of Cd-res (spillway models tested in the reservoir) to Cd-Channel (Normal
orientation located in a channel) vs. HT/P for α = 6° and α = 12° are presented in Figs. 5-6
and 5-7, respectively.
No difference in hydraulic performance was observed between the Normal and

109
Table 5-3. Trend line coefficients for half-round trapezoidal
labyrinth weirs, valid for 0.05 ≤ HT/P ≤ 0.2
α Coefficients
Orientation
(°) A B C D
Projecting, θ = 30° -10.072 -13.85 3.4033 0.5238
Arced
Projecting, θ = 20° -15.86 -6.7336 2.1836 0.5647
Projecting, θ = 10° 25.031 -22.061 3.8631 0.488
6 Linear
Projecting, θ = 0° 98.599 -47.272 6.0173 0.3819
Flush 166.004 -68.1254 7.4922 0.3373
Rounded Inlet 112.61 -47.638 5.2119 0.441
Projecting, θ = 30° 89.891 -44.348 6.9154 0.4284
Arced
Projecting, θ = 20° 31.087 -20.732 3.8441 0.546

Projecting, θ = 10° 35.244 -21.308 3.4392 0.5719
12
Projecting, θ = 0° 8.8398 -10.593 1.8034 0.6258
Linear
Flush 83.586 -41.581 5.5661 0.4719

Rounded Inlet 79.276 -37.17 4.8114 0.5168
Table 5-4. Trend line coefficients for half-round trapezoidal

labyrinth weirs, valid for 0.2 ≤ HT/P ≤ 0.7
α Coefficients
Orientation
(°) A B C D
Projecting, θ = 30° -4.1930 7.3673 -4.6092 1.2327
Arced
Projecting, θ = 20° -3.3019 5.9622 -3.9526 1.1798

Projecting, θ = 10° -3.2392 5.709 -3.7124 1.1178
6
Projecting, θ = 0° -1.8936 3.5802 -2.5204 0.8605
Linear
Flush -1.8381 3.2521 -2.2005 0.762

Rounded Inlet -2.0028 3.5671 -2.4166 0.833
Projecting, θ = 30° 1.5198 -1.3712 -0.5984 0.9124
Arced
Projecting, θ = 20° 1.4404 -1.3929 -0.4088 0.8606

Projecting, θ = 10° 1.2107 -1.0806 -0.4449 0.8128
12
Projecting, θ = 0° -0.1153 0.7162 -1.1144 0.8163
Linear
Flush -0.7374 1.5114 -1.3966 0.8162

Rounded Inlet -1.1832 2.1713 -1.7164 0.8916
Inverse spillway orientations. The abrupt increase in efficiency seen in Fig. 5-7 at HT/P ~
0.25 is due to a sudden decrease in the reference Cd data (Normal orientation in a
channel) caused by the nappe shifting from the clinging to the aerated nappe condition
(Chapter 4). This abrupt shift in discharge efficiency was not observed for the α = 6°
normally oriented weir (channel application), nor in the models tested in the reservoir,
110
Table 5-5. Cd representative single sample uncertainties for
labyrinth weirs tested in this study, HT/P ≥ 0.05
α Cd Single Sample Uncertainty

Orientation
(°) Min (%) Avg (%) Max (%)
Inverse 1.350 1.925 4.215
Normal 1.354 2.148 7.064
Projecting, θ = 0° 1.676 2.389 5.798
Projecting, θ = 10° 1.504 2.122 5.637
6
Projecting, θ = 20° 1.469 2.072 4.945
Projecting, θ = 30° 1.610 2.121 5.308
Flush 1.782 2.481 6.284
Rounded Inlet 1.640 2.329 5.778
Normal 0.843 1.301 3.963
Projecting, θ = 0° 1.025 1.841 5.014
Projecting, θ = 10° 0.923 1.579 5.141
12 Projecting, θ = 20° 0.895 1.555 4.943
Projecting, θ = 30° 0.934 1.642 5.448
Flush 1.075 1.712 5.043
Rounded Inlet 1.011 1.863 4.957
1.4
α=6° Normal in Channel
α=6° Projecting (Linear,θ=0°)
α=6° Arced Projecting, θ=10°
1.3 α=6° Arced Projecting, θ=20°
α=6° Flush
α=6° Rounded Inlet
1.2
Cd-Res / Cd-Channel
1.1
0.9
0.8
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
HT/P
Fig. 5-6. Comparison of labyrinth weir orientations for α = 6°

111
1.4
α=12° Normal in Channel
α=12° Projecting (Linear,θ=0°)
1.3 α=12° Arced Projecting, θ=20°
α=12° Flush
α=12° Rounded Inlet
Cd-Res / Cd-Channel 1.2
1.1
0.9
0.8
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7
HT/P
Fig. 5-7. Comparison of labyrinth weir orientations for α = 12°
because labyrinth cycles did not all transition from clinging to aerated flow conditions
under identical discharges (attributed to the angle variation in approach flows).
The Flush orientation proved to be the least efficient orientation with Cd
decreasing by ~10%. The Rounded Inlet orientation (Rabutment = w) and the Projecting
orientation behaved similarly for α = 12° (~2%-5% less efficient) yet for the α = 6° weirs,
the behavior of the Cd data for the Projecting orientation and Rounded Inlet were
distinctly different from one another. The Projecting orientation became ~5% more
efficient at ~ 0.15 HT/P because the flow passing over the outside side legs of O1 and O5
(see Fig. 5-8) was approximately perpendicular. However, at ~0.28 HT/P, local
submergence and the downstream channel for α = 6° caused the hydraulic efficiency of
112
Fig. 5-8. Example of flow passing from O1 to I1 and O5 to I4
the Projecting orientation to decline. As the flow was required to abruptly change
direction, the outlet labyrinth weir cycles O1 and O5 became submerged and contributed
some flow to the adjacent inlet cycles I1 and I4, producing a noticeable wake (see Fig. 5-
8). The overall effects of O1 and O5 on discharge efficiency become less significant as N
increases (e.g., the effects shown in Fig. 5-8 are relatively insignificant at N = 30).
The arced cycle configurations provided efficiency gains ranging from 10% to
over 25% for the α = 12°; however, these gains in efficiency were limited by local
submergence. As shown in Figs. 5-6 and 5-7, this submergence effect and the control
shifting downstream greatly limited the efficiency of the θ = 20° and θ = 30° for HT/P ≥
0.5. Therefore, it is important to verify that local submergence and the discharge
capacity of the downstream channel does not limit the discharge capacity of the arced
labyrinth weir spillway.
A comparison of the discharge capacities of two arced projecting labyrinth weirs

113
(α = 6° and 12°, θ = 30°) and two arced projecting linear weirs is presented in Table 5-6.
The arc radius, R, was unique for each arced labyrinth weir geometry; consequently, two
separate arced linear weirs, one for each arced labyrinth weir R, were evaluated. The
arced linear weirs overlay the downstream apexes of the corresponding arced labyrinth
weirs and the location of the endpoints of the arced labyrinth and linear projecting weirs
were common (the contact points between the weir walls and the reservoir headwall). At
HT/P = 0.1, the discharge capacities of the α = 6° and 12° arced projecting labyrinth weirs
are ~ 690% and 380% greater (due to the significantly longer crest lengths) than the arced
projecting linear weirs. Even at a relatively high HT/P value of 0.6, where a large portion
of the labyrinth weir crest length experiences local submergence and Cd is significantly
less than a linear weir, the α = 6° and 12° labyrinth weirs have ~ 270% and 180%
(respectively) greater discharge capacity than an arced linear weir.
Flow Characteristics
In order to optimize the orientation of a labyrinth weir, the site conditions, permit
restrictions, approaching flow field, upstream pool elevation limitations, required
discharge capacity, and construction costs should be considered. The following
discussion presents general flow characteristics and design considerations associated with
each labyrinth weir orientation.
Table 5-6. Discharge comparison for arced projecting labyrinth weirs

and arced projecting linear weirs
QLab / QLin
HT/P α = 6°, θ = 30° α = 12°, θ = 30°
0.1 693% 381%
0.3 319% 267%
0.6 192% 183%
114
The flow passing over a labyrinth weir with cycles that are poorly aligned to the
approach flow direction will make significant changes in flow direction at the upstream
apexes, shown in Fig. 5-9. As the head on the weir increases, surface turbulence
increases, vortices can develop, the discharge per cycle becomes unbalanced, pressure
waves can form, and areas of local submergence occur [Fig. 5-9 (A) and (B)]. Further
increases in HT will expand the regions of local submergence and will eventually engulf
nearly the entire weir crest [Fig. 5-9 (C)], which will greatly diminish the hydraulic
efficiency of the labyrinth weir spillway. Fig. 5-9 (C) also illustrates a high local
submergence condition where the flow control region has shifted toward the downstream
end of the labyrinth weir.
Flow separation that occurs in labyrinth weirs with a Flush orientation is
presented Fig. 5-10 (A). At higher discharges, the flow that normally enters I1 and I5
partially enters I2 and I4, which results in flow separation at the guide walls and a less
efficient spillway design. The surface waves and the wake associated with the flow
separation at the abutments extended into I2 and I4. For labyrinth weir spillways with
many cycles, the reduction in spillway capacity associated with the abutment wall will be
less significant. However, for spillways with fewer cycles, it is suggested that the inlet be
modified to prevent flow separation and maintain an equal flow distribution to each
labyrinth cycle. A rounded inlet (Rabutment = w) is presented as an example in Fig. 5-10
(B).
For cases where it may not be feasible to add guide walls or move the spillway
into the downstream channel, the discharge capacity can be increased by projecting the
spillway into the reservoir, as shown in Fig. 5-11 (A) and (B). The hydraulic efficiency
115
Fig 5-9. Examples of surface turbulence (A) and (B), and local submergence (B) and (C)
116
Fig. 5-10. A labyrinth weir with the Flush orientation (A) and a Rounded Inlet (B)
Fig. 5-11. A 5-cycle trapezoidal labyrinth weir, Projecting, with α = 6°

at HT/P = 0.604 (A) and α = 12° HT/P = 0.595 (B)
of this orientation is also limited by the outside labyrinth cycles (I1 and I4, O1 and O5).
The reservoir regions that flow over the outside sidewalls of O1 and O5 [see arrows Fig.
5-11(A)] are significantly larger than the regions that contribute flow to O2, O3, and O4.
At increased Q and HT, flow that normally entered O1 and O5 was observed to spill into
the adjacent labyrinth cycles, I1 and I4, creating wakes and an increase in local
submergence. Fig. 5-11 (A) and (B) also presents observable differences in local
submergence for α = 6° and 12° at HT/P ~ 0.6.

117
An arced cycle configuration better orients each cycle to the available approach
flow area (reservoir application) relative to a projecting labyrinth weir, and reduces the
size inequality of the reservoir regions that flow into each labyrinth weir cycle. α = 12°,
θ = 10° arced labyrinth weirs are presented in Fig. 5-12. Further increases in θ improve
cycle orientation and spillway efficiency, as shown in Fig. 5-13 (A). Nevertheless, if the
downstream cycle discharge capacity is inadequate, local submergence will, at higher
heads, [shown in Fig. 5-13 (B)] increase the pool elevation for a given discharge.
Fig. 5-12. A 5-cycle trapezoidal labyrinth weir, α = 12°, θ = 10°

at HT/P = 0.200 (A) and HT/P = 0.400 (B)
Fig. 5-13. A 5-cycle trapezoidal labyrinth weir, α = 12°, θ = 30°

at HT/P = 0.203 (A) and HT/P = 0.400 (B)
118
Geometric Similitude Considerations
for Arced Labyrinth Weirs
The experimental data from the laboratory-scale models tested in this study
should be scalable to predict the performance of geometrically similar and geometrically
comparable prototype structures. The issue of geometric similitude for arced labyrinth
weirs, however, warrants additional comment. Examples of 5-cycle arced labyrinth weirs
at two different size scales (geometrically similar) are presented in Fig. 5-14 (A). An
alternative weir layout to the larger size-scale weir is shown in Fig. 5-14 (B); Lc R, W’,
and Θ remain constant but the cycle scale is reduced by ½, resulting in 2N and an arc of
θ/2. The arced labyrinth weir spillways shown in 5-14 (A) and (B) have geometrically
similar downstream cycles, but the cycle configuration is not geometrically similar (α’
has changed) and the discharge performance is not directly scalable from (A) to (B).
Based on the fact that (A) and (B) are geometrically comparable (reasonably similar or
quasi-similar), (B) should have a similar discharge capacity to (A) and the information
Fig. 5-14. Two geometrically similar arced labyrinth weir spillways, N = 5 (A)
and a geometrically comparable design at ½ scale,
equivalent crest length, and N = 10 (B)
119
presented in this study can be used as a first approximation. However, this design should
still be verified with a physical or numerical model study. Arced labyrinth weir designs
that fall outside the scope of the weirs evaluated in this study should also be verified with
a model study.
Design Example
The following example illustrates the use of the design information presented in
this paper for labyrinth weir spillway design. Design discharge values are typically
obtained from hydrologic and risk assessment flood routing studies. To determine an
initial cycle design with a discharge capacity that meets flow event estimations, it is
recommended that the design method presented in Chapter 4 be used. For this example
the following cycle geometry is used: a quarter-round crest shape, α = 12°, W = 89.6 m, B
= 25.5 m, P = 6.1 m, tw = 45.7 cm. Lc is calculated to be 283.8 m, N = 7, and predicted
Cd are presented in Table 5-7. The head-discharge, tailwater relationships, and spillway
hydrograph can now be estimated from labyrinth cycle discharge (Q/N) and hydraulic
profiling of the downstream channel or chute.
Following the preliminary spillway design, the weir orientation is selected. The
decrease in efficiency for a Projecting orientation for N = 7 should be less than what is
estimated (~5% at HT/P = 0.6) in Fig. 5-7 [e.g., Cd ≥ 0.392*0.95% = 0.369]. The
efficiency of the weir may be increased, according to the data presented in Fig. 5-7, by
placing the cycles in an arced configuration; the results are presented in Table 5-7.
Additional labyrinth weir configurations (e.g., Flush orientation, Rounded Inlet, alternate
crest shapes, etc.) can also evaluated in design development. This study includes
120
Table 5-7. Predicted Cd to confirm calculated results
Crest HT/P
α = 12° Shape 0.20 0.40 0.60
Normal in Channel QR 0.576 0.473 0.392
Linear, Projecting, θ = 0° QR 0.551 0.458 0.369
Arced, Projecting, θ = 10° QR 0.604 0.512 0.411
Arced, Projecting, θ = 20° QR 0.647 0.539 0.415
Arced, Projecting, θ = 20° HR 0.735 0.566 0.425
experimental results for half-round, α = 12° labyrinth weir spillways; therefore, Eq. (5-2)
and Tables 5-3 and 5-4 may be directly applied (e.g., θ = 20° labyrinth weir orientation).
Further adjustments to weir geometry and spillway orientation may follow as the spillway
design is refined.
Although the design tools presented herein will accurately predict the hydraulic
performance of a labyrinth weir spillway, these results have only been confirmed with the
physical models that were tested in this study (Table 5-2). A physical model study is
recommended to verify hydraulic performance and may provide important insights
specific to the spillway location, flow conditions, and geometric designs that may be
outside the scope of this study.
This study provides hydraulic information, specific to labyrinth weir spillways in
a reservoir application, to be used in conjunction with the design and analysis method
presented in Chapter 4. Discharge coefficients as a function of HT/P (graphical and trend
line) are presented for a variety of arced and linear labyrinth weir geometries, specific to
reservoir applications. Discharge rating curves may be modified with Figs. 5-6 and 5-7
for a specific labyrinth weir orientation or cycle configuration. Phenomena were

121
identified (surface turbulence, vortices, local submergence, wakes) that decrease
labyrinth weir discharge capacity in a reservoir application for each tested labyrinth weir
orientation. Also, a standard geometric design layout for an arced labyrinth weir spillway
(cycles configuration follows the arc of a circle) is set forth, including important
geometric parameters.
A comparison (Figs. 5-6 and 5-7) of tested labyrinth weir spillway orientations
(Normal, Inverse, Projecting, Flush, Rounded Inlet, and Arced) showed that that the
projecting arced labyrinth weir had the maximum discharge efficiency, ~5% – 30%
greater than the Normal orientation; no difference in discharge efficiency was observed
between the Normal orientation and the Inverse orientation. The Flush orientation was
~10% less efficient than the Normal orientation. Rounded abutments (Rounded Inlet,
Rabutment ≥ w) were ~2% – 5% less efficient than the Normal orientation; therefore,
rounded abutments decrease flow separation at the abutment walls and improve the
efficiency of the Flush configuration.
This study found that it is possible to over-design a labyrinth weir spillway.
Highly efficient labyrinth weir models (e.g., θ ≥ 20°) may be limited by local
submergence and eventually by the discharge capacity of the outlet labyrinth weir cycles
and exit channel width. As HT increases, local submergence regions also increase,
causing the critical section governing spillway discharge to travel down the outlet
labyrinth cycle and eventually to the downstream channel.
The design tools and information presented herein will accurately design and
analyze labyrinth weirs that are geometrically similar to the models tested. This
information may also be used as a first approximation for geometrically comparable

122
arced labyrinth weir spillways (e.g., Fig. 5-14) that feature geometrically similar outlet
labyrinth cycles (α) but dissimilar inlet labyrinth cycles (α’).
It is recommended that a spillway design be verified with a physical or numerical
model study. A model study would confirm hydraulic performance estimations, and
would include site-specific conditions and any unique flow conditions or geometric
designs outside the scope of this study.
Additional components of this study not presented here include a detailed look at
nappe behavior (including local submergence, nappe interference, and nappe stability),
scale effects, and other labyrinth weir flow phenomena.

123
CHAPTER 6
NAPPE AERATION, NAPPE INSTABILITY, AND NAPPE INTERFERENCE
FOR LABYRINTH WEIRS
Abstract
Nappe aeration conditions for trapezoidal labyrinth weirs on a horizontal apron
with quarter- and half-round crests (6° ≤ sidewall angle ≤ 35°) are presented as a tool for
labyrinth weir design. Specified HT/P ranges, hydraulic behaviors associated with each
aeration condition, and nappe instability phenomena are documented and discussed. The
effects of artificial aeration (a vented nappe) on discharge capacity are presented. Nappe
interference for labyrinth weirs is defined, and the effects of nappe interference on the
discharge capacity of a labyrinth weir cycle are discussed, including the parameterization
of nappe interference regions to be used in labyrinth weir design. Finally, the
applicability of techniques developed for quantifying nappe interference of sharp-crested
corner weirs is examined.
Introduction
A labyrinth weir (Fig. 6-1) is a type of polygonal overflow weir structure that is
characterized by its hydraulic performance and its distinct geometric shape (triangular,
trapezoidal, or rectangular cycles). The geometry of a labyrinth weir cycle produces
complex 3-dimensional flow patterns; the head-discharge relationship of labyrinth weirs
has been determined empirically by the general weir equation [Eq. (6-1)].
2 32
Q= C d Lc 2 g H T (6-1)
3
124
Fig. 6-1. Example of a labyrinth weir
In Eq. (6-1), Q is the weir discharge, Cd is a dimensionless discharge coefficient,
Lc is the centerline length of the weir crest, g is the acceleration constant of gravity, and
HT is the total upstream head defined as HT = V2/2g + h [V is the average cross-sectional
velocity and h is the piezometric head (measured relative to the weir crest elevation) just
upstream of the weir].
The advantages of labyrinth weirs relative to linear weirs can be illustrated by
examining Eq. (6-1). The geometry of the labyrinth weir provides an increase in Lc,
resulting in an increase in Q for a given channel width. If Q is held constant, HT must
decrease as Lc increases; labyrinth weirs require less freeboard for a given design flood
and can facilitate increased reservoir storage under base-flow conditions, relative to linear
weirs. Labyrinth weirs are most efficient at low heads, but as HT increases, the efficiency
of labyrinth weirs declines. Although labyrinth weir Cd values may be less than linear
125
weir Cd values, the increase in Lc typically more than compensates, providing an increase
in discharge capacity relative to linear weirs.
In addition to Lc and HT, labyrinth weir discharge is influenced by the cycle
geometry [e.g., sidewall angle (α), centerline apex length (Ac)], the cycle configuration
(arced or linear), the weir orientation (e.g., Normal, Inverse, Flush, Projecting, Rounded
Inlet), the shape of the weir crest, the approach flow conditions (e.g., the approach angle
of the flow relative to the labyrinth weir cycle), nappe behavior (e.g., nappe aeration
conditions, nappe instability, nappe interference) and the depth of flow downstream of
the weir walls (e.g., tailwater submergence and local submergence). The Cd values
determined from physical modeling indirectly account for these influences on Q.
Discharge coefficients and discharge rating curves for labyrinth weirs have been
determined from physical models of prototype structures [e.g., Avon (Darvas 1971),
Dungo (Magalhães and Lorena 1989), Hyrum (Houston 1983), Keddara (Magalhães and
Lorena 1989), Lake Brazos (Tullis and Young 2005), Lake Townsend (Tullis and
Crookston 2008), Ute (Houston 1982), and Woronora (Darvas 1971)] and from general
labyrinth weir research studies. Preceding prominent design methods have been
presented by Tullis et al. (1995), Magalhães and Lorena (1989), Lux (1989), Hinchliff
and Houston (1984), Darvas (1971), and Hay and Taylor (1970) (Chapters 4 and 5
discuss design methods and information for labyrinth weirs in detail). However,
hydraulic design information is currently inadequate regarding the influence of nappe
interference, nappe aeration conditions, and nappe instability specific to labyrinth weirs.
A nappe is the jet of water that passes over a weir. In this study, four aeration
conditions of the nappe are defined; clinging, aerated, partially aerated, and drowned.
126
Clinging refers to the nappe adhering to the downstream face of the weir wall at lower
values of HT/P. An aerated nappe features an air cavity behind the nappe. As HT/P
increases the air cavity varies spatially and temporally; it becomes non-uniform
(distributed air pockets rather than one continuous air pocket along the weir wall) and
unstable (air pocket size and location changes with time). This condition is referred to as
partially aerated. Finally, the drowned nappe aeration condition features a thick nappe
without an air cavity; this condition occurs at higher values of HT/P. Nappe instability
refers specifically to a nappe with an unsteady or oscillating trajectory. Observations
indicated that nappe instability occurred briefly with the aerated and drowned conditions,
but most frequently with the partially aerated nappe aeration condition.
Nappe interference occurs when two or more nappes collide (Fig. 6-2). For
labyrinth weirs, nappe interference originates at the upstream apex and can produce
wakes downstream of the apex (Fig. 6-2), standing waves [6-3 (A)] and air bulking [6-3
Fig. 6-2. Collision of nappes from adjacent sidewalls and the apex
127
Fig. 6-3. The effects of nappe interference: standing waves (A), wakes and
air bulking (B), and local submergence (C)
128
(B)]. At low HT values, nappe interference is typically comprised of a turbulent nappe
“collision” region. As HT increases, the portion of the labyrinth weir outlet cycle
adjacent to the upstream apex becomes overwhelmed by the discharge from the sidewalls
and apex, thus creating a local submergence condition [see Fig. 6-3(C)]. Local
submergence differs from the traditional tailwater-induced submergence in that local
submergence is independent of the downstream tailwater conditions. The local
submergence region develops downstream of the upstream apex and increases in size as
weir discharge increases. During this study, observations indicated that local
submergence occurred for quarter- and half-round crest shapes.
Nappe interference reduces the local labyrinth weir discharge capacity. The size
of the region influenced by nappe interference is dependent upon α, Ac, crest shape, P,
HT, and the nappe aeration condition; the effects of nappe interference are not explicitly
accounted for in labyrinth weir design methods. For example, a labyrinth weir of four
cycles (N = 4) should have a higher discharge capacity (under common hydraulic
conditions) than an identical labyrinth weir of the same Lc but with 8 cycles (N = 8) (see
Fig. 6-4) because the portion of the weir length affected by nappe interference is larger
for the N = 8 weir.
Indlekofer and Rouvé (1975) explored the concept of nappe interference by
studying sharp-crested corner weirs (α = 23.4°, 31°, 44.8°, 61.7°). A corner weir can be
characterized as a single triangular labyrinth weir cycle with channel boundaries
perpendicular to each sidewall. Indlekofer and Rouvé divided the corner weir into two
flow regions: a disturbed region where the flow from each sidewall converges (colliding
nappes) and a second region where the flow streamlines are perpendicular to the sidewall
129
Fig. 6-4. Example of nappe interference regions for an aerated nappe at low HT/P
(i.e., linear weir flow) (see Fig. 6-5). The length of the crest within the disturbed area
was defined as Ld. By comparing the efficiency of a corner weir to a linear weir, an
average discharge coefficient for the disturbed area, Cd-m; a theoretical disturbance length,
LD; and an empirical discharge relationship were developed [Eq. (6-2)]. Cd-m represents
the efficiency of a corner weir relative to a linear weir (Cd-m = Cd-corner / Cd(90°)).
Applying the linear weir discharge coefficient, Cd(90°), to the corner weir, LD represents
Fig. 6-5. Nappe interference region and parameters as defined by

Indlekofer and Rouvé (1975) for sharp-crested corner weirs
130
 3Q  1 1
Ld =  Lc(α ° ) −  = LD
 2Cd( 90° ) 2 g hm  1 − Cd − m
3 2
1 − Cd − m

(6-2)
the theoretical portion of crest length where Q and Cd = 0 (see Fig. 6-5). In Eq. (6-2), hm
is the head upstream of the weir as defined by Indlekofer and Rouvé (1975); hm
represents a specific upstream depth and includes two velocity components [see
Indlekofer and Rouvé (1975) for details].
Falvey (2003) applied this approach to the experimental results of several
labyrinth weir models. Using corner weir data, Falvey developed an empirical LD
relationship [Eq. (6-3)] as an alternative to polynomial relationships developed by
Indlekofer and Rouvé. Falvey also developed Eq. (6-4) based upon an analysis of
available labyrinth weir experimental data. Falvey does not, however, give a
recommendation with regard to which LD equation is most appropriate or accurate.
Based on an analysis of Tullis et al. (1995) labyrinth weir discharge rating curves, Falvey
proposes a design limit of LD / lc ≤ 0.35 (35% or less of weir length is ineffective), where
lc is the weir sidewall length. For corner weirs and triangular labyrinth weirs, lc = Lc-cycle /
2; for trapezoidal labyrinth weirs, Lc-cycle / 2 = lc + Ac. Falvey also states that additional
research is needed, including ascertaining the validity of Eq. (6-4). In Eq. (6-4), HT/P is
the headwater ratio (total upstream head over the weir height).
LD
= 6.1e− 0.052α ° α ≥ 10° (6-3)
h
 H  
LD = lc  0.224 ln T  + (0.94 − 0.03α °) α ≤ 20° and HT/P ≥ 0.1 (6-4)
  P  
131
The purpose of this study is to provide new information regarding labyrinth weir
nappe aeration conditions, nappe instability, and nappe interference and their influence
on the discharge capacity of labyrinth weirs with quarter-round or half-round crest
shapes. This was accomplished by analyzing trapezoidal labyrinth weir experimental
data sets for 6° ≤ α ≤ 35° with a quarter- and half-round crest shape. Also, the influence
of artificial aeration (vented nappe) is quantified relative to non-vented nappe flow. In
addition, the flow conditions when nappe instability occurs are documented. The
definition for nappe interference is refined, and regions of influence are determined.
Finally, the techniques proposed by Indlekofer and Rouvé (1975) for nappe interference
of corner weirs and the application of these techniques by Falvey (2003) are examined.
Experimental Method
To explore nappe interference, nappe aeration conditions, and nappe instability,
20 labyrinth weirs were fabricated from High Density Polyethylene Plastic (HDPE) and
tested in a rectangular flume (1.2 m x 14.6 m x 1.0 m) at the Utah Water Research
Laboratory (UWRL). Details of the tested model geometries are summarized in Table 6-
1. The flume featured a headbox and baffle to provide uniform approach conditions for a
given discharge rate. The labyrinth weirs were installed on an elevated horizontal apron
with a ramped (2.4 m) upstream floor transition. Willmore (2004) found the effects of
ramped transitions on the discharge capacity of labyrinth weirs to be negligible. Based
upon the findings of Johnson (1996) the influence of flume sidewall effects were also
considered to be negligible.
Calibrated orifice meters in the flume supply piping, differential pressure

132
Model α P Lc-cycle Lc-cycle/w w/P N Crest Type Orientation†

() (°) (mm) (cm) () () () () () ()
1 6 304.8 465.457 7.607 2.008 2 HR Trap Inverse
2-3 6 304.8 465.457 7.607 2.008 2 QR, HR Trap Normal
4-5 8 304.8 354.492 5.793 2.008 2 QR, HR Trap Normal
6-7 10 304.8 287.905 4.705 2.008 2 QR, HR Trap Normal
8-9 12 304.8 243.514 3.980 2.008 2 QR, HR Trap Normal
10-11 15 304.8 199.135 3.254 2.008 2 QR, HR Trap Normal
12-13 20 304.8 154.810 2.530 2.008 2 QR, HR Trap Normal
14 15 152.4 199.135 3.254 4.015 2 QR Trap Normal
15 15 152.4 99.567 3.254 2.008 4 QR Trap Normal
16 15 304.8 99.567 3.254 2.008 4 QR Trap Normal
17-18 35 304.8 98.352 1.607 2.008 2 QR, HR Trap Normal
19-20 90 304.8 122.377 1.000 4.015 - QR, HR - -
†Linear configuration was used for all model orientations
transducers, and a data logger were used to meter the flow rates in the test flume. The
flume was equipped with a stilling well and a rolling instrument carriage that featured
point gauge instrumentation (0.15 mm). The point gauge instrumentation was carefully
referenced to the crest of the labyrinth, which was leveled to ±0.4 mm. The test program
evaluated the influence of artificial aeration (the air cavity behind the nappe was vented
to atmosphere). Each labyrinth weir model with a quarter-round crest shape was tested
with and without a nappe aeration apparatus consisting of an aeration tube for each
labyrinth sidewall [example shown in Fig. 6-6(A)]. Wedge shaped nappe breakers were
tested at three different locations {upstream apex, downstream apex [Fig. 6-6(B)], and
~lc/2 [Fig. 6-6 (C)]} and various location combinations (e.g., upstream and downstream
apex locations) for a labyrinth weir cycle.
Experimental data were collected under steady-state conditions. Q measurement
data were averaged for 5-7 minutes and h was determined using the stilling well equipped
133
(A) (B) (C)
Fig. 6-6. Aeration tube apparatus for N = 2 (A) and nappe breakers located on the
downstream apex (B) and on the sidewall (C)
with a point gauge. A large number of head-discharge data points were collected for all
tested weir geometry, including a system of checks wherein at least 10% of the data were
repeated to determine measurement accuracy and repeatability. Velocity data were
measured inside the weir cycles with a 2-dimensional acoustic doppler velocity probe.
Digital photography and a measurement grid (located on the flume sidewall) were used to
quantify regions of nappe interference. The surface fluctuations of a nappe, and
consequently the fluctuation in the size for the region of nappe interference, increase with
HT and are influenced by cycle geometry. Therefore, nappe interference measurement
accuracy varies from model to model and decreases as HT increases (e.g., ±5 mm for α =
8° and HT/P = 0.1, ±25 mm for α = 10° and HT/P = 0.3, ±15 mm for α = 12° and HT/P =
0.5). In addition to nappe interference, digital photography and high-definition (HD)
digital video recording were used extensively to document the hydraulic behaviors of the
tested labyrinth weirs. Observations noted nappe behavior, nappe aeration conditions,
nappe stability, nappe separation points, areas of local submergence, wakes, harmonic or
recurring hydraulic behaviors for all α tested. Finally a dye wand was used to investigate
the complex 3-dimensional flow characteristics associated with labyrinth weirs.

134
Nappe Aeration Conditions
Labyrinth weirs can experience four different nappe aeration conditions: clinging
(Fig. 6-7), aerated (Fig. 6-8), partially aerated (Fig. 6-9), and drowned (Fig. 6-10). The
shape of the weir crest, P, HT, the depth and turbulence of flow behind the nappe, the
momentum and trajectory of the flow passing over the crest, and the pressure behind the
nappe (sub-atmospheric for non-vented or atmospheric for vented nappes) influence the
aeration condition. As HT increases, a labyrinth weir will transition from clinging to
aerated, to partially aerated, and finally to drowned. However, all four aeration
conditions do not necessarily occur for all labyrinth weir cycle geometries or crest
shapes.
Fig. 6-7. Clinging nappe aeration condition observed for trapezoidal labyrinth weir,
half-round crest shape, α = 12°, HT/P = 0.196
135
Fig. 6-8. Aerated nappe aeration condition observed for trapezoidal labyrinth weir,
quarter-round crest shape, α = 12°, HT/P = 0.202
Fig. 6-9. Partially aerated nappe aeration condition observed for trapezoidal labyrinth
weir, half-round crest shape, α = 12°, HT/P = 0.296
136
Fig. 6-10. Drowned nappe aeration condition observed for trapezoidal labyrinth weir,
quarter-round crest shape, α = 12°, HT/P = 0.604
The discharge efficiency of a labyrinth weir is influenced by the aeration
condition of the nappe. Aeration conditions characterize nappe behavior, which may be
relatively tranquil or may produce pressure fluctuation on the weir wall, noise, and
vibrations. For example, a clinging nappe (Fig. 6-7) is generally more efficient than an
aerated nappe (Fig. 6-8) because sub-atmospheric pressures develop on the downstream
face of the weir. A partially aerated nappe (Fig. 6-9) occurs at larger values of HT/P and
does not have a stable air cavity behind the nappe (varies temporally and spatially). The
air cavity oscillates between labyrinth weir apexes, the amount of the sidewall length that
is aerated fluctuates, and the air cavity may be completely removed and then reappear as
the turbulent levels and unsteady flow behavior behind the nappe fluctuate. Although the
137
air cavity is highly dynamic and causes fluctuating pressures on the downstream face of
the weir, observations noted stable and unstable nappe trajectories (depending upon weir
geometry and flow conditions) for the partially aerated nappe condition. For a stable
nappe, the partially aerated condition had minimal influence on the nappe trajectory.
Further increases in HT cause the nappe to shift from partially aerated to drowned. The
drowned nappe aeration condition features a thick nappe without an air cavity. Ranges of
HT/P that correspond to observed nappe aeration condition for quarter-round and half-
round labyrinth weirs are presented in Figs. 6-11 and 6-12, respectively.
For labyrinth weirs with a smooth quarter round crest shape, clinging conditions
cease at HT/P~0.05. The nappe condition shifts from aerated to partially aerated at 0.25 ≤
Fig. 6-11. Nappe aeration and instability conditions

for labyrinth weirs with a quarter-round crest
138
Fig. 6-12. Nappe aeration and instability conditions

for labyrinth weirs with a half-round crest
HT/P ≤ 0.29, depending on α (α = 8° – 10° have the largest aeration range). The drowned
condition begins at HT/P = 0.31 for α = 6°. As α increases, the inception of the drowned
condition begins at higher values of HT/P. For α ≥ 12, the drowned condition begins at
HT/P = 0.51.
As can be seen in Figs. 6-11 and 6-12, the half-round crest shape produces
different aeration condition ranges than the quarter-round crest. Depending on α, the
clinging condition can be maintained up to HT/P = 0.4 (α = 35°). The Cd values in the
clinging condition range are greater than those in the aerated or partially aerated range, as
exhibited by the abrupt decrease in Cd as the nappe shifts from clinging to aerated or
partially aerated. Labyrinth weirs with 15° ≤ α ≤ 20° were observed to shift directly from
139
a clinging nappe to a partially aerated nappe, and nappe aeration occurred only briefly for
the α = 35° at HT/P~0.15.
Nappe Instability
Figs. 6-11 and 6-12 also present the ranges of HT/P when nappe instability
occurred (α ≥ 12° for quarter-round and half-round crest). Nappe instability refers to a
nappe that has an oscillating trajectory (temporal variations) and may be accompanied by
abrupt shifts in the aeration condition of the nappe. It is a low frequency phenomenon
that occurs under constant upstream flow conditions (i.e., HT, and Q) and is a significant
event for α ≥ 12. Nappe instability affects complete labyrinth weir cycles (two sidewalls
and the downstream apex); nappe oscillations may be synchronized for all labyrinth weir
cycles or temporal variations between cycles may exist. During testing, 3-dimenional
unsteady flow conditions were observed downstream of the sidewalls using dye tracking.
Turbulent mixing in that region created air bulking in the flow around the nappe.
Explorations in the downstream cycle with the dye wand noted turbulent, helical flow
currents traveling relatively parallel and adjacent to the sidewall. The observed
fluctuations of the nappe and turbulent mixing all appeared to contribute to dynamic
pressures behind the nappe. Under these conditions, the nappe was drawn toward the
weir wall, and there appeared to be a critical point when air and/or water were drawn
behind the nappe from the adjacent flow, creating an audible flushing noise. Relatively
large quantities of fluid were introduced in bursts, resulting in the abrupt change of nappe
trajectory.
140
At higher flow rates, air cavity formation and nappe instability diminished
(increased turbulent mixing) and artificial aeration or venting of the nappe was found to
decrease nappe instability and noise. Despite artificial aeration, nappe instability was still
observed to occur (to a lesser degree) for α ≥ 20° in the partially aerated (quarter-round
and half-round crest shapes) and drowned (quarter-round crest shape only) aeration
conditions. Nappe instability was not observed to occur for α < 12° and α < 10° for
quarter- and half-round crest shapes, respectively. The net effect of nappe instability on
prototype structures is unclear; however, avoiding these ranges in labyrinth weir design is
suggested because undesirable and potentially harmful levels of vibration, pressure
fluctuation, and noise may result.
Artificial Nappe Aeration
Artificial aeration, or venting the nappe to atmosphere, had a negligible effect on
the discharge capacity of quarter-round labyrinth weirs (~0.5% to 1.7%). With respect to
discharge efficiency, venting a half-round labyrinth weir crest (using aeration vents or
nappe breakers) reduces the range of HT/P for the clinging nappe aeration condition,
thereby reducing flow efficiency at low heads and partially diminishing the benefit of a
half-round crest shape. Aeration vent or nappe breaker hydraulic effects and placement
for labyrinth weir design are discussed in Chapter 4.
Nappe Interference
Nappe Interference for Labyrinth Weirs. Nappe interference refers to the region
where two or more nappes intersect, and it occurs at the upstream apexes in labyrinth
weirs. Nappe interference locally decreases the weir discharge efficiency in that region.
141
Although the effects of nappe interference and apex influence are inherently accounted
for in the discharge coefficients and rating curves proposed in labyrinth weir design
methods, it is important to characterize and quantify the size of the nappe interference
region to determine if the hydraulic performance of a labyrinth weir design will deviate
from design method predictions (i.e., nappe interference may cause two labyrinth weir
designs with common sidewall angles and weir lengths, but with a different number of
cycles, to exhibit different head-discharge characteristics).
In order to characterize the size of nappe interference regions, Bint was developed
and physically measured; this interference length is illustrated in Fig. 6-13. It describes
the interference region length originating at and perpendicular to the upstream apex wall
to the point where the nappe region intersects the weir crest. Depending upon the
labyrinth weir geometry and the flow conditions, the nappe interference region may
include a turbulent flow region [Fig. 6-13 (D)], a local submergence region [Figs. 6-13
(C), or both [Fig. 6-13 (A)]. Bint may be used to approximate the portion of crest length
within the nappe interference region.
The sizes of the nappe interference regions for the design method proposed in
Chapter 4 are quantified in Figs. 6-14 and 6-15 for quarter-round and half-round crest
shapes. As expected, Bint increases with HT; for the quarter-round crest shape, the aerated
nappe condition increases the size of the turbulent region and, therefore, Bint (HT ≤ 125
mm, which corresponds to 0.1 ≤ HT/P ≤ 0.3). As the nappe shifts from the aerated to the
partially aerated condition, the size of the turbulent region decreases as the influence of
the quarter-round crest shape on the nappe trajectory diminishes (nappe trajectory
becomes less horizontal). The half-round crest shape does not have a flat, horizontal
142
Fig. 6-13. Physical representation of Bint in plan-view (A) and (C) and profile view (B)
and (D) for nappe interference regions
surface and therefore does not feature this anomaly.
Geometric scaling may be used to convert the information presented in Figs. 6-14
and 6-15 to determine the size of the nappe interference region for other labyrinth weir
structures (e.g., Pproto/P * Bint = Bint-proto). Bint is independent of B or Lc; no dimensionless
parameter was found that accurately represents the interference region for all labyrinth
weir geometric configurations, therefore it is not presented as a dimensionless parameter
to determine nappe interference region size for labyrinth weir design. Because families
of curves result from varying B or Lc for a geometrically similar labyrinth weir cycle,
143
700
6 degree QR
8 degree QR
600
10 degree QR
500 12 degree QR
15 degree QR
Bint (mm)
400 20 degree QR
35 degree QR
300
200
100
0
0 50 100 150 200 250 300
HT (mm)
Fig. 6-14. Bint for quarter-round trapezoidal labyrinth weirs, 6°≤ α ≤ 35°
useful dimensionless ratios (e.g., Bint/B) are to be computed after determining Bint.
To quantify the percentage of B that is comprised of Bint for the physical models
tested in this study, Bint/B is presented in Figs. 6-16 and 6-17 for quarter-round and half-
round labyrinth weirs. In general, Bint was approximately 10% to 40% of B during testing
of the quarter-round crest shape labyrinth weirs. Nappe aeration affected the increasing
trend (0.2 ≤ HT/P ≤ 0.35) of Bint/B with α for quarter-round labyrinth weirs. This
anomaly did not occur for half-round labyrinth weirs; in general, the regions of nappe
interference for half-round labyrinth weirs were smaller or equivalent in size to the
regions observed with quarter-round labyrinth weirs. Crest shape appears to have little
144
700
6 degree HR
8 degree HR
600
10 degree HR
500 12 degree HR
15 degree HR
Bint (mm)
400 20 degree HR
35 degree HR
300
200
100
0
0 50 100 150 200 250 300
HT (mm)
Fig. 6-15. Bint for half-round trapezoidal labyrinth weirs, 6°≤ α ≤ 35°
influence on Bint for HT/P ≥ 0.5. As stated previously, a dimensionless approach to nappe
interference was found to produce families of curves. Figs. 6-16 and 6-17 are not
applicable to all labyrinth weirs but do characterize the percentage of the downstream
cycle [(for the labyrinth weir models tested in this study (see Chapter 4)] within the
nappe interference region.
Application of Published Techniques for Nappe Interference. Based on their work
with aerated, sharp-crested corner weirs, Indlekofer and Rouvé (1975) investigated nappe
interference for sharp-crested corner weirs with an aerated nappe. They proposed that the
discharge of any polygonal weir could be determined by assuming the weir is composed
of linear weirs joined with disturbed corner areas. Total weir discharge was determined
145
1.0
6 degree QR
0.9 8 degree QR
10 degree QR
0.8
12 degree QR
0.7
15 degree QR
0.6 20 degree QR
Bint /B
35 degree QR
0.5
0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
HT/P
Fig. 6-16. Bint/B specific to quarter-round trapezoidal labyrinth weirs

tested in this study (6°≤ α ≤ 35°)
from the summation of discharges computed for each portion of the polygonal weir. The
discharge capacity for each weir portion was computed using Eq. (6-2), which requires
selecting appropriate Cd-m and Ld and LD values for each corner (as a function of α) from
figures they developed. The proposed methodology of Indlekofer and Rouvé is based
upon the following assumptions: excluding the disturbed corner areas, the flow passing
over the weir sidewalls is perpendicular to the crest; the weir features a sharp-crest, the
nappes of the weir are stable and fully aerated; the reduction in discharge capacity
(relative to a linear weir) is solely attributable to the colliding nappes; and Ld and LD
increase linearly with HT.

146
1.0
6 degree HR
0.9 8 degree HR
10 degree HR
0.8
12 degree HR
0.7
15 degree HR
0.6 20 degree HR
Bint /B
35 degree HR
0.5
0.4
0.3
0.2
0.1
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
HT/P
Fig. 6-17. Bint/B specific to quarter-round trapezoidal labyrinth weirs

tested in this study (6°≤ α ≤ 35°)
In contrast to corner weirs, the flow passing over a labyrinth weir is not
perpendicular to the crest along the sidewalls except at low upstream head (HT/P < 0.05)
and at the center of the upstream and downstream apexes. Depending on the crest shape
geometry and the upstream flow conditions, regions of nappe interference can be heavily
influenced by crest shape. Also, the stability and aeration condition of the nappe vary
with HT for labyrinth weirs with quarter and half-round crest shapes. In addition to
colliding nappes, labyrinth weir efficiency is influenced by weir geometry, nappe
behavior, upstream flow conditions, and local submergence. For this study, Eq. (6-2) was
147
modified to include HT; the non-linear relationships of LD to HT are provided in Figs. 6-18
and 6-19.
LD and Ld do not vary linearly with HT (a major assumption in the Indlekofer and
Rouvé method); however, this is a reasonable approximation for labyrinth weirs with α ≥
35° (the experiments of Indlekofer and Rouvé were for corner weirs with α > 23°). As
shown in Fig. 6-19, a region of transition exists where the slope of LD decreases with
increasing HT; this region corresponds to the commencement of the partially aerated and
drowned aeration conditions (see Figs. 6-11 and 6-12).
None of the aforementioned assumptions of Indlekofer and Rouvé describe the
general hydraulic behaviors of labyrinth weirs located in a channel or reservoir.
Nevertheless, for lack of a more appropriate alternative, Falvey (2003) applied these
3500
6 degree QR
8 degree QR
3000 10 degree QR
12 degree QR
15 degree QR
2500
20 degree QR
35 degree QR
2000
LD (mm)
1500
1000
500
0
0 50 100 150 200 250 300
HT (mm)
Fig. 6-18. LD as a function of HT for quarter-round labyrinth weirs

148
Fig. 6-19. LD as a function of HT for half-round labyrinth weirs
techniques to quantify the effects of nappe interference for labyrinth weirs. Eq. (6-3) (red
line shown in Fig. 6-20) was developed to determine a disturbed crest length based upon
α. Falvey attributes the nonlinear variation of LD to influences from the downstream
channel.
The non-linear variations of LD/HT (from Fig. 6-18) for quarter-round labyrinth
weirs are plotted for 6° ≤ α ≤ 35°. Theoretically, LD/HT should be 0 at α = 90° and
approach ∞ at α = 0° (no flow); Falvey (2003) limits the empirical equation to α ≥ 10°.
The labyrinth weir experimental data do not match the pattern suggested by these three
relationships. Fig. 6-20 is not recommended as it does not sufficiently describe the nature
of the nappe interference region for labyrinth weirs.
Falvey (2003) also developed Eq. (6-4) (proposed for HT/P ≥ 0.1 and α ≤ 20°)
149
20
6 degree QR 8 degree QR
18
16 35 degree QR Sh Corner Weir
0.2 curve fit Indlekofer a nd Rouvé 0.8 curve fit Indlekofer a nd Rouvé
14 Fa lvey (2003)
12
LD /HT
10
0
0 10 20 30 40 50 60 70 80 90
α°
Fig. 6-20. LD/HT as a function of α
based on plots of LD/Lc-cycle vs. HT/P computed from the experimental results of physical
model studies for eight different labyrinth weir prototype structures. Predictions from
Eq. (6-4), LD-Falvey, and Eq. (6-2), LD, are plotted as LD-Falvey/LD vs. HT/P in Figs. 6-21 and
6-22 for quarter-round and half-round labyrinth weirs.
Based upon the experimental results of this study, Eq. (6-4) appears to under
predict LD for quarter-round labyrinth weirs by ~10% (which may not be sufficiently
accurate) for HT/P ≥ 0.3 and α ≤ 12°. For labyrinth weirs with a half-round crest shape, it
under predicts LD by ~ 5% to 20% for HT/P ≥ 0.2 and α ≤ 12°. The accuracy of Eq. (6-4)
decreases for HT/P < 0.2 and for larger angled labyrinth weirs (α > 12°). Based on these
findings, Eq. (6-4) is not recommended for labyrinth weir design. However, LD does
150
2.0
6 degree QR
1.8 8 degree QR
10 degree QR
1.6 12 degree QR
15 degree QR
1.4
20 degree QR
LD-Falvey / LD
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 6-21. LD-Falvey/LD vs. HT/P for quarter-round labyrinth weirs
2.0
6 degree HR
1.8 8 degree HR
10 degree HR
1.6 12 degree HR
15 degree HR
1.4
20 degree HR
LD-Falvey / LD
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
HT/P
Fig. 6-22. LD-Falvey/LD vs. HT/P for half-round labyrinth weirs

151
accurately describe the difference in required weir length between a labyrinth weir and a
linear weir with common discharges. Also, Cd-m represents the relative efficiency of a
labyrinth weir to a linear weir; therefore, LD and Cd-m are useful parameters when
juxtaposing the hydraulic performance and weir lengths of linear and labyrinth weirs.
As discussed previously, Ld is the crest length within the flow area disturbed by
nappe interference for sharp-crested corner weirs [calculated from Eq. (6-2)]. The
portion of the apron within the disturbed area [(Bd), see Fig. 6-5] is a straightforward
calculation from Ld; Indlekofer and Rouvé defined the boundary of this region to be
perpendicular to the weir wall in the downstream cycle. To determine the accuracy of Bd
for labyrinth weirs, predicted nappe interference length regions were compared to Bint
(measured during physical model testing for this study). The ratio of Bd/Bint vs. HT/P is
presented in Figs. 6-23 and 6-24 for quarter and half-round labyrinth weirs.
Based on the findings presented in Figs. 6-23 and 6-24, Bd is not an accurate
representation of Bint for labyrinth weirs. For example, Bd was ~ 20% to 53% and ~ 10%
to 77% larger than Bint for α = 35° for the quarter-round and half-round weir crest shapes,
respectively. Furthermore, Bd was 14- and 35-times larger for α = 6° at HT/P = 0.1 for the
two different crest shapes, a flow condition where Bd should be minimal. For this
geometry and flow condition, Bint was observed to be ~2% of B during testing (39 mm).
Even including the wake (elevation below the crest) created from nappe collision (152
mm, 7% of B), there is poor agreement; Bd was predicted to be ~545 mm or nearly 25%
of B (at HT/P = 0.8, Bint was measured at 586 mm). Therefore, it is recommended that
Figs. 6-14 and 6-15 be utilized to describe the region of nappe interference for labyrinth
weirs.
152
16
6 degree QR
8 degree QR
14
10 degree QR
12 12 degree QR
15 degree QR
10 20 degree QR
Bd/Bint
35 degree QR
8
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
HT/P
Fig. 6-23. Bd/Bint vs. HT/P for quarter-round labyrinth weirs
This study provides new hydraulic information and insights regarding nappe
aeration conditions, nappe instability, and nappe interference for labyrinth weirs. Twenty
physical models (Table 6-1) were used to determine the influence of these phenomena on
labyrinth weir discharge capacity.
The HT/P ranges for clinging, aerated, partially aerated, and drowned nappe
aeration conditions were identified in Figs. 6-11 and 6-12. These regions are crest-shape
specific, vary nonlinearly with α, and are discussed in detail to clearly characterize nappe
behavior. Nappe aeration conditions also account for changes in Cd; a clinging nappe is
more efficient than an aerated, partially aerated, or drowned nappe. The influence of
153
40
6 degree HR
8 degree HR
35
10 degree HR
30 12 degree HR
15 degree HR
25 20 degree HR
Bd/Bint
35 degree HR
20
15
10
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
HT/P
Fig. 6-24. Bd/Bint vs. HT/P for half-round labyrinth weirs
artificial aeration (vented nappe) on discharge capacity was found to be negligible
(~0.5% – 1.7%), relative to the non-vented nappe conditions, for quarter-round crest
shapes. For half-round labyrinth weirs, aeration vents or nappe breakers limit the
operating range of the clinging nappe and diminish the hydraulic efficiency benefits
provided by the crest shape.
Physical modeling also identified regions of nappe instability for α ≥ 12° for
quarter and half-round crest shapes. Observations noted the presence of sweeping
turbulent flow exiting the downstream cycle, a fluctuating water volume behind the
nappe, dynamic pressures behind the nappe, and turbulent mixing during nappe
instability. For half-round crest shapes, nappe instability is specific to the partially
154
aerated nappe condition. Nappe instability occurred to a lesser degree for α ≥ 20° in the
partially aerated (quarter-round and half-round crest shapes) and drowned (only quarter-
round crest shape) aeration conditions when the nappe was vented. The net effect of
nappe instability on prototype structures is unclear, but it is recommended that these
ranges be avoided in labyrinth weir design, as vibrations, pressure fluctuations, and noise
levels may reach sufficient magnitudes to be undesirable or harmful.
This study refined the definition of nappe interference for the reason that nappe
interaction can produce a turbulent collision region or a region of local submergence,
depending on HT. The effects of nappe interference and consequently apex influence are
inherent but not separately quantified in discharge coefficients and rating curves
proposed in labyrinth weir design methods. However, the size of the nappe interference
region was quantified (Figs. 6-14 and 6-15) to facilitate a comparison between nappe
interference regions between a labyrinth weir design and the design method (e.g.,
maintaining Lc but varying N). Such a comparison will indicate qualitatively if the
hydraulic performance of a labyrinth weir will deviate from design method predictions.
Finally, the techniques proposed by Indlekofer and Rouvé (1975) for nappe
interference of corner weirs and Falvey’s (2003) application of these techniques to
labyrinth weirs were examined. Neither Ld nor LD were found to accurately predict the
extent of the nappe interference regions for labyrinth weirs. Eq. (6-2) does accurately
describe the difference in weir lengths and net discharge efficiencies between a labyrinth
weir and a linear weir of equivalent discharges. Eqs. (6-3) and (6-4) were not validated
for general labyrinth weir application.

155
Additional research is needed to examine the 3-dimensional components of nappe
interference and their local influences on labyrinth weir discharge.

156
CHAPTER 7
SUMMARY AND CONCLUSIONS
Synopsis
The purpose of this research study was to improve the design and analysis of
labyrinth weirs by meeting the objectives presented in Chapter 1. The following
discussion summarizes the contents and contributions of Chapters 2-6.
Chapter 2 – Background and Literature
The background of labyrinth weirs and an extensive review of published literature
are presented in Chapter 2. This includes information regarding labyrinth weir modeling
(analytical approach, similarity relationships), a compilation and refinement of labyrinth
weir nomenclature and terminology, the history and evolution of labyrinth weir design
(including significant design methods and case studies), and a list of labyrinth weirs from
across the globe.
Chapter 3 – Experimental Setup

and Testing Procedure
This study is based upon the experimental results of 32 labyrinth weir physical
models. Chapter 3 documents in detail the experimental setup, fabrication and
installation tolerances, model configurations, and test procedures of this study. Labyrinth
weirs featured a quarter- and half-round crest shape and were installed on a horizontal
platform in an elevated headbox or in a laboratory flume at the Utah Water Research
Laboratory (UWRL). Quarter-round labyrinth weirs were tested with and without an
artificial aeration device. Model configurations included Normal and Inverse orientation
157
in a channel and Flush, Rounded Inlet, and Projecting orientations in a reservoir. The
Projecting orientations included Linear and Arced cycle configurations. The test program
also included four α = 15° models with a quarter-round crest where P, tw, N, and w/P
were varied. Data were collected using calibrated orifice meters and differential pressure
transducers, point gauges, stilling wells, a 2-dimensional acoustic Doppler velocity probe,
a dye injection apparatus, and high definition digital video and still cameras.
Experimental discharge rating curve data sets are comprised of ~60 to 100 individual data
points (total of 2,606 tested flow conditions) and an uncertainty analysis based upon the
method of Kline and McClintock (1953) documented experimental uncertainty. A
system of checks was established wherein at least 10% of the data were repeated to
ensure accuracy and determine measurement repeatability. Experimental data also
includes velocity flow fields, nappe profiling, nappe aeration conditions, nappe instability
conditions, nappe interference regions, regions of local submergence, and other detailed
observations on labyrinth weir hydraulic behaviors and flow phenomena.
Chapter 4 – Hydraulic Design and

Analysis of Labyrinth Weirs
This chapter presents a labyrinth weir design and analysis procedure (Table 4-6)
based upon the results of 20 physical models tested in a laboratory flume. Q is calculated
based on the traditional weir equation [Eq. (4-1)], utilizing HT and selecting the centerline
length of the weir, Lc, as the characteristic length. Tailwater submergence for labyrinth
weirs, as presented by Tullis et al. (2007), is included. The proposed design and analysis
method is validated by juxtaposing the experimental results of this study with other
physical model studies presented in Figs. 4-12 and 4-13, and Table 4-8.
158
Figs. 4-3 and 4-4 present a dimensionless discharge coefficient, Cd, as a function
of HT/P for quarter-round and half-round labyrinth weirs (6° ≤ α ≤ 35°) and linear weirs.
The test results indicate that the increase in efficiency provided by a half-round crest
shape (relative to a quarter-round crest) is more significant for HT/P ≤ 0.4.
Cycle efficiency, ε’, is a tool for examining the discharge capacity of different
labyrinth weir geometries (Figs. 4-8 and 4-9). The results of ε’ indicate how the increase
in crest length compensates for the decline in discharge efficiency associated with
decreasing α.
The experimental results indicate that nappe aeration conditions and nappe
stability should not be overlooked in the hydraulic and structural design of labyrinth
weirs. The results presented in Tables 4-4 and 4-5 indicate flow behaviors that may
include negative or fluctuating pressures at the weir wall, noise, and vibrations. These
tables also aid in the selection of a crest shape. Finally, the effects of nappe ventilation
by means of aeration vents or nappe breakers are put forth, including recommended
placements of vents (one per sidewall) and breakers (one centered on each downstream
apex).
Although the methods and tools presented herein will accurately design and
analyze a labyrinth spillway, a physical model study is recommended to verify hydraulic
performance. A model study would include site-specific conditions that may be outside
the scope of this study and may provide valuable insights into the performance and
operation of the labyrinth weir.

159
Chapter 5 – Arced and Linear labyrinth
Weirs in a Reservoir Application
This chapter provides hydraulic information specific to labyrinth weir spillways in
a reservoir application. It is to be used in conjunction with the design and analysis
method presented in Chapter 4. Discharge coefficients as a function of HT/P (graphical
and trend line) are presented for a variety of arced and linear labyrinth weir geometries,
specific to reservoir applications. Discharge rating curves may be modified with Figs. 5-
6 and 5-7 for a specific labyrinth weir orientation or cycle configuration. Phenomena
were identified (surface turbulence, vortices, local submergence, wakes) that decrease
labyrinth weir discharge capacity in a reservoir application for each tested labyrinth weir
orientation. Also, a standard geometric design layout for an arced labyrinth weir spillway
(cycles configuration follows the arc of a circle) is set forth, including important
geometric parameters.
A comparison (Figs. 5-6 and 5-7) of tested labyrinth weir spillway orientations
(Normal, Inverse, Projecting, Flush, Rounded Inlet, and Arced) showed that that the
projecting arced labyrinth weir had the maximum discharge efficiency, ~5%-30% greater
than the Normal orientation; no difference in discharge efficiency was observed between
the Normal orientation and the Inverse orientation. The Flush orientation was ~10% less
efficient than Normal orientation. Rounded abutments (Rounded Inlet, Rabutment ≥ w)
were ~2% – 5% less efficient than the Normal orientation; rounded abutments decrease
flow separation at the abutment walls and improve the efficiency of the Flush
configuration.
160
This study found that it is possible to over-design a labyrinth weir spillway.
Highly efficient labyrinth weir models (e.g., θ ≥ 20°) may be limited by local
submergence and eventually by the discharge capacity of the labyrinth weir cycle outlets.
As HT increases, local submergence regions also increase resulting in the critical section
that governed spillway discharge to travel down the outlet labyrinth cycle and eventually
to the downstream channel.
The design tools and information presented herein will accurately design and
analyze labyrinth weirs that are geometrically similar to the models tested. This
information may also be used as a first approximation for geometrically comparable
arced labyrinth weir spillways (e.g., Fig. 5-14) that feature geometrically similar outlet
labyrinth cycles (α) but dissimilar inlet labyrinth cycles (α’).
It is recommended that a spillway design be verified with a physical or numerical
model study. A model study would confirm hydraulic performance estimations, and
include site-specific conditions and any unique flow conditions or geometric designs
outside the scope of this study.
Chapter 6 – Nappe Aeration, Nappe

Instability, and Nappe Interference
for Labyrinth Weirs
This chapter provides new hydraulic information and insights regarding nappe
aeration conditions, nappe instability, and nappe interference for labyrinth weirs. 20
physical models (Table 6-1) were used to determine the influence of these phenomena on
labyrinth weir discharge capacity.

161
The HT/P ranges for clinging, aerated, partially aerated, and drowned nappe
aeration conditions were identified in Figs. 6-11 and 6-12. These regions are crest-shape
specific, vary nonlinearly with α and are discussed in detail to clearly characterize nappe
behavior. Nappe aeration conditions also account for changes in Cd; a clinging nappe is
more efficient than an aerated, partially aerated, or drowned nappe. The influence of
artificial aeration (vented nappe) on discharge capacity was found to be negligible
(~0.5% to 1.7%), relative to the non-vented nappe conditions, for quarter-round crest
shapes. For half-round labyrinth weirs, aeration vents or nappe breakers limit the
operating range of the clinging nappe and diminish the hydraulic efficiency benefits
provided by the crest shape.
Physical modeling also identified regions of nappe instability for α > 15° and α >
12° for quarter and half-round crest shapes, respectively. Observations noted the
presence of sweeping turbulent flow exiting the downstream cycle, a fluctuating water
volume behind the nappe, dynamic pressures behind the nappe, and turbulent mixing
during nappe instability. For half-round crest shapes, nappe instability is specific to the
partially aerated nappe condition. Nappe instability occurred to a lesser degree for α ≥
20° in the partially aerated (quarter-round and half-round crest shapes) and drowned
(only quarter-round crest shape) aeration conditions when the nappe was vented. The net
effect of nappe instability on prototype structures is unclear, but it is recommended that
these ranges be avoided in labyrinth weir design, as vibrations, pressure fluctuations, and
noise levels may reach sufficient magnitudes to be undesirable or harmful.
This study refined the definition of nappe interference for the reason that nappe
interaction can produce a turbulent collision region or a region of local submergence,

162
depending on HT. The effects of nappe interference and consequently apex influence are
inherent but not separately quantified in discharge coefficients and rating curves
proposed in labyrinth weir design methods. However, the size of the nappe interference
region was quantified (Figs. 6-14 and 6-15) to facilitate a comparison between nappe
interference regions between a labyrinth weir design and the design method (e.g.,
maintaining Lc but varying N). Such a comparison will indicate qualitatively if the
hydraulic performance of a labyrinth weir will deviate from design method predictions.
Finally, the techniques proposed by Indlekofer and Rouvé (1975) for nappe
interference of corner weirs and the application of these techniques by Falvey (2003) to
labyrinth weirs were examined. Neither Ld nor LD were found to accurately predict the
extent of the nappe interference regions for labyrinth weirs. Eq. (6-2) does accurately
describe the difference in weir lengths and net discharge efficiencies between a labyrinth
weir and a linear weir of equivalent discharges. Eqs. (6-3) and (6-4) were not validated
for general labyrinth weir application.

163
REFERENCES
Afshar, A. (1988). “The development of labyrinth weir design.” Water Power and Dam
Construction, 40(5), 36-39.
Amanian, N. (1987). “Performance and design of labyrinth spillways.” M.S. thesis, Utah
State University, Logan, Utah.
ASCE. (2000). Hydraulic modeling: Concepts and practice, Manual 97, ASCE, Reston,
Va.
Babb, A. (1976). “Hydraulic model study of the Boardman Reservoir Spillway.” R.L
Albrook Hydraulic Laboratory, Washington State University, Pullman, Wash.
Brinker, D. (2005). “Boyd lake spillway – An innovative approach to using a labyrinth

weir.” Proc. of the 2005 Dam Safety Conference, ASDSO.
CH2M-Hill, (1973). “Mercer Dam Spillway model study.” Dallas Oregon, March
International Report.
Chanson, H. (1995). “Air bubble diffusion in supercritical open channel flow.” Proc.
12th Australasian Fluid Mechanics Conference AFMC. Sidney, Australia. (2), 707-
710.
Chanson, H. 1997. Air bubble entrainment in free-surface turbulent shear flows.

Academic Press, London, U.K.
Copeland, R. and Fletcher, B. (2000). “Model study of Prado Spillway, California,

hydraulic model investigation.” Report ERDC/CHL TR-00-17, U.S. Army Corps of
Engineers, Research and Development Center.
Crookston, B.M. and Tullis, B. (2010). “Hydraulic performance of labyrinth weirs.”

Proc. of the Int. Junior Researcher and Engineer Workshop on Hydraulic Structures
(IJREWHS ’10). Edinburgh, U.K.
Darvas, L. (1971). “Discussion of performance and design of labyrinth weirs, by Hay

and Taylor.” J. of Hydr. Engrg., ASCE, 97(80), 1246-1251.
Emiroglu, M., Kaya, N. and Agaccioglu, H. (2010). “Discharge capacity of labyrinth

side weir located on a straight channel.” J. of Irr. and Drain. Engrg., 136(1), 37-46.
Falvey, H. (1980). Chapter 2. Practical experiences with flow-induced vibrations, E.

Naudascher, and D. Rockwell, eds. Springer-Verlag, Berlin, Heidelberg, New York.
386-398.
164
Falvey, H. (2003). Hydraulic design of labyrinth weirs. ASCE, Reston, Va.
Falvey, H., and Treille. P., (1995). “Hydraulics and design of fusegates.” J. of Hydr.
Engrg., ASCE, 121(7), 512-518.
Gentilini, B. (1940). “Stramazzi con cresta a planta obliqua e a zig-zag.” Memorie e

Studi dell Instituto di Idraulica e Construzioni Idrauliche del Regil Politecnico di
Milano, No. 48 (in Italian).
Hay, N., and Taylor, G. (1970). “Performance and design of labyrinth weirs.” J. of
Hydr. Engrg., ASCE, 96(11), 2337-2357.
Hinchliff, D., and Houston, K. (1984). “Hydraulic design and application of labyrinth
spillways.” Proc. of 4th Annual USCOLD Lecture.
Houston, K. (1982). “Hydraulic model study of Ute Dam labyrinth spillway.” Report
No. GR-82-7, U.S. Bureau of Reclamation, Denver, Colo.
Houston, K. (1983). “Hydraulic model study of Hyrum Dam auxiliary labyrinth

spillway.” Report No. GR-82-13, U.S. Bureau of Reclamation, Denver, Colo.
Indlekofer and Rouvé, G. (1975). “Discharge over polygonal weirs.” J. of Hydr. Engrg.,
ASCE, 110(HY3), 385-401.
Johnson, M. (1996). “Discharge coefficient scale effects analysis for weirs.” Ph.D.
dissertation, Utah State University, Logan, Utah.
Kabiri-Samani, A. (2010). “Analytical approach for flow over an oblique weir.”

Transaction A: Civil Engineering. 17(2), 107-117.
Khatsuria, R., Deolalikar, P., and Bhosekar, V. (1988). “Design of duckbill spillway and
reversed sloping curved stilling basin.” Salauli Project, 54th R&D Session, CBI&P,
Ranchi, India.
Kline, S. and McClintock, F. (1953). “Describing uncertainties in single-sample

experiments.” American Society of Mechanical Engineers, 75(1), 3-8.
Kozák, M. and Sváb, J. (1961). “Tort alaprojzú bukók laboratóriumi vizsgálata.”

Hidrológiai Közlöny, No. 5. (in Hungarian)
Laugier, F. (2007). “Design and construction of the first piano key weir (PKW) spillway
at the goulours dam.” Hydropower & Dams, Issue 5.
Lempérière, F., Ouamane, A. (2003). “The piano keys weir: a new cost-effective
solution for spillways.” The Int. J. on Hydropower and Dams, 10(5).
165
Lopes, R., Matos, J., and Melo, J. (2006). “Discharge capacity and residual energy of
labyrinth weirs.” Proc. of the Int. Junior Researcher and Engineer Workshop on
Hydraulic Structures (IJREWHS ‘06), Montemor-o-Novo, Hydraulic Model Report
No. CH61/06, Div. of Civil Engineering, the University of Queensland, Brisbane,
Australia, 47-55.
Lopes, R., Matos, J., and Melo, J. (2008). “Characteristic depths and energy dissipation
downstream of a labyrinth weir.” Proc. of the Int. Junior Researcher and Engineer
Workshop on Hydraulic Structures (IJREWHS ‘08), Pisa, Italy.
Lopes, R., Matos, J., and Melo, J. (2009). “Discharge capacity for free-flow and
submerged labyrinth weirs.” Proc. 33rd IAHR Congress, Water Engineering for a
Sustainable Environment, Vancouver BC, Canada, 1054-1061.
Lux, F. (1984). “Discharge characteristics of labyrinth weirs.” Proc. of Conference on

Water for Resources Development, ASCE, Cour d’Alene, ID.
Lux, F. (1985). “Discussion on ‘Boardman labyrinth crest spillway.’” J. of Hydr.

Engrg., ASCE, 111(6), 808-819.
Lux, F. (1989). “Design and application of labyrinth weirs.” Design of Hydraulic

Structures 89, M. Alberson, and R. Kia, eds., Balkema/Rotterdam/Brookfield.
Lux, F. and Hinchliff, D. (1985). “Design and construction of labyrinth spillways.” 15th
Congress ICOLD, Vol. IV, Q59-R15, Lausanne, Switzerland, 249-274.
Magalhães, A., and Lorena, M. (1989). “Hydraulic design of labyrinth weirs.” Report
No. 736, National Laboratory of Civil Engineering, Lisbon, Portugal.
Magalhães, A., and Lorena, M. (1994). “Perdas de energia do escoamento sobre soleiras
em labirinto.” Proc. 6º SILUSB/1º SILUSBA, Lisboa, Portugal, 203-211. (in
Portuguese).
Matos, J. and Frizell, K. (1997). “Air concentration measurements in highly turbulent

aerated flow.” Proc. 27th IAHR Congress, Theme B, San Francisco, Calif., USA,
1:149-154.
Matos, J. and Frizell, K. (2000). “Air concentration and velocity measurements on self-
aerated flow down stepped chutes.” Proc. 2000 Joint Conference on Water Resources
Engineering and Water Resources Planning and Management, ASCE, Minneapolis,
USA (CD-ROM).
Matthews, G. (1963). “On the influence of curvature, surface tension, and viscosity on
flow over round-crested weirs.” Paper No. 6683, University of Aberdeen, Aberdeen,
Scotland, 511-524.
166
Mayer, P. (1980). “Bartletts Ferry project, labyrinth weir model studies.” Project No. E-
20-610, Georgia Institute of Technology, Atlanta, Ga.
Melo, J., Ramos, C., and Magalhães, A. (2002). “Descarregadores com soleira em
labirinto de um ciclo em canais convergentes. Determinação da capacidad de vazão.”
Proc. 6° Congresso da Água, Porto, Portugal. (in Portuguese).
Naudascher, E., and Rockwell, D. (1994). Flow induced vibrations - an engineering

guide. Balkema Press, Rotterdam, Brookfield.
Noori, B. and Chilmeran, T. (2005). “Characteristics of flow over normal and oblique
weirs with semicircular crest.” Al Ra_dain Engrg. J., 13(1), 49-61.
Page, D., García, V., and Ninot, C. (2007). “Aliviaderos en laberinto. presa de María
Cristina.” Ingeniería Civil, 146(2007), 5-20 (in Spanish).
Phelps, H. (1974). “Model study of labyrinth weir – Navet pumped storage project.”
University of the West Indies, St. Augustine, Trinidad, West Indies.
Quintela, A., Pinheiro, A., Afonso, J., and Cordeiro, M. (2000). “Gated spillways and
free flow spillways with long crests. Portuguese dams experience.” 20th ICOLD Q79-
R12, Beijing, China, 171-189.
Ribeiro, M., Boillat, J., Schleiss, A., Laugier, F., and Albalat, C. (2007). “Rehabilitation
of St-Marc dam. Experimental optimization of a piano key weir.” Proc. of the 32nd
Congress of IAHR, Venice, Italy.
Rouve, G., and Indlekofer, H. (1974). "Abfluss über geradlinige wehre mit
halbkreissförmigem überfallprofil." Der Bauingenieur, 49(7), 250-256 (in German).
Savage, B., Frizell, K., and Crowder, J. (2004). Brian versus brawn: The changing world
of hydraulic model studies. Proc. of the ASDSO Annual Conference, Phoenix, Ariz.,
CD-ROM.
Taylor, G. (1968). “The performance of labyrinth weirs.” Ph.D. thesis, University of

Nottingham, Nottingham, England.
Tullis, B. and Crookston, B.M. (2008). “Lake Townsend Dam spillway hydraulic model
study report.” Utah Water Research Laboratory, Logan, Utah.
Tullis, B. and Young, J. (2005). “Lake Brazos Dam model study of the existing spillway
structure and a new labyrinth weir spillway structure.” Hydraulics. Report No. 1575.
Utah Water Research Laboratory. Logan, Utah.
167
Tullis, B., Young, J., & Chandler, M. (2007). “Head-discharge relationships for
submerged labyrinth weirs.” J. of Hydr. Engrg., ASCE, 133(3), 248-254.
Tullis, P. (1992). “Weatherford Spillway model study.” Hydraulic Report No. 311, Utah
Water Research Laboratory, Logan, Utah.
Tullis, P. (1993). “Standley Lake service spillway model study.” Hydraulic Report No.
341, Utah Water Research Laboratory, Logan, Utah.
Tullis, P., Amanian, N., and Waldron, D. (1995). “Design of labyrinth weir spillways.”
J.of Hydr. Engrg., ASCE, 121(3), 247-255.
USACE (1991). “Sam Rayburn Dam – spillway.” Value Engineering Team Study. U.S.
Army Engineer District, Kansas City, Mo.
Vasquez, V., Boyd, M., Wolfhope, J., and Garret, R. (2007). “A labyrinth rises in the
heart of Texas.” Proc. of the 28th Annual USSD Conference. Portland, Ore.
Vermeyen, T. (1991). “Hydraulic model study of Ritschard Dam spillways.” Report No.
R-91-08, U.S. Bureau of Reclamation, Denver, Colo.
Villemonte, D. (1947). “Submerged weir discharge studies.” Engineering News Record,

866.
Waldron, D. (1994). “Design of labyrinth spillways.” M.S. thesis, Utah State University,
Logan, Utah.
Willmore, C. (2004). “Hydraulic characteristics of labyrinth weirs.” M.S. report, Utah

State University, Logan, Utah.
Wormleaton, P., and Soufiani, E. (1998) “Aeration performance of triangular planform

labyrinth weirs.” J. of Environ. Engrg., ASCE, 124(8), 709-719.
Wormleaton, P., and Tsang, C. (2000). “Aeration performance of rectangular planform

labyrinth weirs.” J. of Environ. Engrg., ASCE, 127(5), 456-465.
Yildiz, D., and Uzecek, E. (1996). “Modeling the performance of labyrinth spillways.”
Hydropower, 3:71-76.
Young, J. (2005). “Submergence effects on head-discharge relationships for labyrinth

and sharp-crested linear weirs.” M.S. thesis, Utah State University, Logan, Utah.
168
APPENDICES
169
APPENDIX A
SCHEMATICS OF TESTED LABYRINTH WEIR PHYSICAL MODELS IN THE

RECTANGULAR FLUME FACILITY
170
Fig. A-1. Schematic of 2-cycle, trapezoidal 6° quarter- and half-round

labyrinth weirs, normal orientation
Fig. A-2. Schematic of 2-cycle, trapezoidal 6° half-round

labyrinth weir, inverse orientation
171


172


173
Fig. A-7. Schematic of 4-cycle, trapezoidal 15° quarter-round


174

175
APPENDIX B
SCHEMATICS OF TESTED LABYIRNTH WEIR PHYSICAL MODELS IN THE

RESERVOIR FACILITY
176

projecting orientation, linear cycle configuration (θ = 0°)
B-2. Schematic of 5-cycle, trapezoidal 6° half-round labyrinth weir,

projecting orientation, arced cycle configuration (θ = 10°)
177


178

flush orientation, linear cycle configuration

rounded inlet orientation, linear cycle configuration
179

projecting orientation, linear cycle configuration (θ = 0°)

180


181

flush orientation, linear cycle configuration

rounded inlet orientation, linear cycle configuration
182
APPENDIX C
VISUAL BASIC CODE, SPECIFIC TO RECTANGULAR

FLUME FACITY, USED IN MICROSOFT EXCEL
183
Option Explicit
'for use with 4-ft rectangular flume with transmitters in UWRL (9-15-2007)
'calibration of o-plates 2/2/2010
Function flowt4(Size, dH)
Dim beta, a, Dorifice, Dpipe, pi, C, g As Double
pi = 3.14159265359
g = 32.174
If (Size = 8) Then
C = 0.6205 ' previously was 0.616 '0.6033
Dorifice = 5.5839 '5.719
Dpipe = 7.932 '7.625
beta = Dorifice / Dpipe
a = Dorifice ^ 2 * pi * 0.25 / 144
Else
If (Size = 20) Then
C = 0.6282 'previously was 0.611
Dorifice = 14.625
Dpipe = 19.5
a = Dorifice ^ 2 * pi * 0.25 / 144
Else
End If
End If
flowt4 = C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
End Function
'for use with reservoir headbox on lower floor level with transmitters in UWRL
Function flowtRes(Size, dH, g, leak)
Dim beta, a, Dorifice, Dpipe, pi, C, Calib As Double
pi = 3.14159265359
'Calibrated coefficient and precise geometry for each nominal orifice size in inches to
feet
If (Size = 4) Then
C = 0.6197
a = 1.5 ^ 2 * 3.14159 * 0.25 / 144
beta = 1.5 / 4.026
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) - leak
ElseIf (Size = 8) Then
C = 0.6106
a = 5 ^ 2 * 3.14159 * 0.25 / 144
beta = 5 / 7.981
Calib = 1 '+ 0.0357131
184
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) * Calib - leak
C = 0.6029
a = 14.016 ^ 2 * 3.14159 * 0.25 / 144
beta = 14.016 / 19.25
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5)
Calib = 1 - (0.000071079566 * flowtRes ^ 2 - 0.002182705515 * flowtRes +
0.024449497333)
flowtRes = flowtRes * Calib - leak
Else: flowtRes = "Check Meter!"
End If
End Function
'4-ft Flume Calculation
'To determine uncertainty in single sample measurement, from Kline and McClintock
1953
Function SSUCd4ft(Size, mA, deltaH, Q, Ptgage, Ht, P, Lc, W, Yplatform, Yramp, Yref,
g)
Dim beta, Aorifice, Dorifice, Dpipe, pi, C As Double
Dim wQ, wLc, wHt, wC, wW, wPtgage, wH, wP, wYplatform, wYramp, wYref, wmA,
H
Dim dQ, dH, dP, dYplatform, dYramp
pi = 3.14159265359
Lc = Lc / 12 'convert from inches to feet
W = W / 12 'convert from inches to feet
'Calculate Q in 4-ft flume

If (Size = 8) Then
C = 0.6205 ' previously was 0.616 '0.6033
Dorifice = 5.5839 '5.719
Dpipe = 7.932 '7.625
Aorifice = Dorifice ^ 2 * pi * 0.25 / 144
Else
If (Size = 20) Then
Dorifice = 14.625
Dpipe = 19.5
Else
End If
End If
185
Q = C * Aorifice * (2 * g * deltaH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
H = Ptgage - Yref
Ht = H + Q ^ 2 / (2 * g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 2)
'Assign values from instrumentation

'wQtank = 0.0015
wQ = 0.0025 * Q
wLc = (1 / 32) / (2 * 12) '+- 1/64 of inch
wW = (1 / 16) / 2 '+- error, can read smaller but have to average diff flume widths
wPtgage = 0.0005 / 2 '+-error in feet
wYref = 0.0005 / 2 '+-error in feet
wmA = 0.01 / 2 '+-error in mA
wYramp = (1 / 32) / (2 * 12) '+- 1/64 of inch
wYplatform = (1 / 32) / (2 * 12) '+- 1/64 of inch
'Calculate uncertainties
wH = (((wPtgage / H) ^ 2 + (wYref * (-1) / H) ^ 2) ^ (1 / 2)) * H
'Calc wHt by taking derivatives
dH = 1 - (Q ^ 2) / (g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 3)
dQ = Q / (g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 2)
dP = -Q ^ 2 / (g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 3)
dYplatform = Q ^ 2 / (g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 3)
dYramp = Q ^ 2 / (g * W ^ 2 * (H + P + Yplatform - Yramp) ^ 3)
wHt = (((wH * dH / Ht) ^ 2 + (wQ * dQ / Ht) ^ 2 + (wP * dP / Ht) ^ 2 + (wYplatform *
dYplatform / Ht) ^ 2 + (wYramp * dYramp / Ht) ^ 2) ^ (1 / 2)) * Ht
'%Uncertainty of single Cd value from labyrinth in 4-ft flume

SSUCd4ft = ((wQ / Q) ^ 2 + (-wLc / Lc) ^ 2 + (-27 / 8 * wHt / Ht) ^ 2) ^ (1 / 2)
End Function
Function flowt3(Size, dH, g)
Dim beta, a, C As Double
If (Size = 2) Then
C = 0.6345
a = 1.035 ^ 2 * 3.14159 * 0.25 / 144
beta = 0.507
Else
If (Size = 4) Then
C = 0.6277
a = 3 ^ 2 * 3.14159 * 0.25 / 144
186
beta = 0.7452
Else
If (Size = 10) Then
C = 0.707
a = 10.508 ^ 2 * 3.14159 * 0.25 / 144
beta = 10.508 / 12
Else
If (Size = 12) Then
C = 0.6151
a = 8.005 ^ 2 * 3.14159 * 0.25 / 144
beta = 0.6671
Else
End If
End If
End If
End If
flowt3 = C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
End Function
Function Calc_CdT(Q, Ht, Tlength)
Function_Cd = 3 / 2 * Q / (Tlength / 12 * (32.2 * 2) ^ 0.5 * Ht ^ (3 / 2))
End Function
Function Calc_CdE(Q, Ht, Elength)
Function_CdE = 3 / 2 * Q / (Elength / 12 * (32.2 * 2) ^ 0.5 * Ht ^ (3 / 2))
End Function
Option Explicit
'Calculate Specific Weight of Water as a function of Temperature (Fahrenheit)

Function GAMMAH2O(wdTemp)
GAMMAH2O = 59.364982 + 3.0750805 * Cos(0.0078331697 * (wdTemp) -
0.24302151)
'Slight adjustment of gamma to match values given in Engineering Fluid Mechanics 7th
edition by Crowe, Elger, Roberson
If wdTemp = 40 Then
GAMMAH2O = 62.43
ElseIf wdTemp = 50 Then
GAMMAH2O = 62.4
187
End If
End Function
'Calculate surface tension as a function of temperature (sigma in lbf/ft)

Function SigmaH2o(wdTemp)
SigmaH2o = -5.6230368808E-06 * wdTemp + 0.005376033645
End Function
188
APPENDIX D
VISUAL BASIC CODE, SPECIFIC TO RESERVOIR FACITY,

USED IN MICROSOFT EXCEL
189
Option Explicit
'calibration of o-plates 2/2/2010
Function flowt4(Size, dH)
Dim beta, a, Dorifice, Dpipe, pi, C, g As Double
pi = 3.14159265359
g = 32.174
If (Size = 8) Then
C = 0.6205 ' previously was 0.616 '0.6033
Dorifice = 5.5839 '5.719
Dpipe = 7.932 '7.625
a = Dorifice ^ 2 * pi * 0.25 / 144
Else
If (Size = 20) Then
Dorifice = 14.625
Dpipe = 19.5
a = Dorifice ^ 2 * pi * 0.25 / 144
Else
End If
End If
flowt4 = C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
End Function
'for use with reservoir headbox on lower floor level with transmitters in UWRL
Function flowtRes(Size, dH, g, leak)
Dim beta, a, Dorifice, Dpipe, pi, C, Calib As Double
pi = 3.14159265359
'Calibrated coefficient and precise geometry for each nominal orifice size in inches to
feet
If (Size = 4) Then
C = 0.6197
a = 1.5 ^ 2 * 3.14159 * 0.25 / 144
beta = 1.5 / 4.026
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) - leak
C = 0.6106
a = 5 ^ 2 * 3.14159 * 0.25 / 144
beta = 5 / 7.981
Calib = 1 '- 0.0357131 '+
190
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) * Calib - leak
C = 0.6029
a = 14.016 ^ 2 * 3.14159 * 0.25 / 144
beta = 14.016 / 19.25
flowtRes = (C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5)
Calib = 1 - (0.000071079566 * flowtRes ^ 2 - 0.002182705515 * flowtRes +
0.024449497333)
flowtRes = flowtRes * Calib - leak
Else: flowtRes = "Check Meter!"
End If
End Function
'4-ft Flume Calculation
1953
Function SSUCd4ft(Size, mA, deltaH, Q, Ptgage, Ht, P, Lc, W, Yplatform, Yramp, Yref,
g)
Dim beta, Aorifice, Dorifice, Dpipe, pi, C As Double
H
pi = 3.14159265359
'Calculate Q in 4-ft flume

If (Size = 8) Then
C = 0.6205 ' previously was 0.616 '0.6033
Dorifice = 5.5839 '5.719
Dpipe = 7.932 '7.625
Else
If (Size = 20) Then
Dorifice = 14.625
Dpipe = 19.5
Else
End If
End If
191
Q = C * Aorifice * (2 * g * deltaH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
H = Ptgage - Yref

'wQtank = 0.0015
wQ = 0.0025 * Q
wLc = (1 / 32) / (2 * 12) '+- 1/64 of inch
wYramp = (1 / 32) / (2 * 12) '+- 1/64 of inch

SSUCd4ft = ((wQ / Q) ^ 2 + (-wLc / Lc) ^ 2 + (-27 / 8 * wHt / Ht) ^ 2) ^ (1 / 2)
End Function
'Reservoir Calculation
1953
Function SSUCdRes(Size, mA, deltaH, Q, Ptgage, Ht, P, Lc, W, Yplatform, Yramp,
Yref, g, leak)
Dim beta, Aorifice, Dorifice, Dpipe, pi, C, Calib As Double
H
pi = 3.14159265359
'Calculate Q in Reservoir
192
If (Size = 4) Then
C = 0.6197
Aorifice = 1.5 ^ 2 * 3.14159 * 0.25 / 144
beta = 1.5 / 4.026
Q = (C * Aorifice * (2 * g * deltaH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) - leak
C = 0.6106
Aorifice = 5 ^ 2 * 3.14159 * 0.25 / 144
beta = 5 / 7.981
Calib = 1 '+ 0.0357131
Q = (C * Aorifice * (2 * g * deltaH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5) * Calib - leak
C = 0.6029
Aorifice = 14.016 ^ 2 * 3.14159 * 0.25 / 144
beta = 14.016 / 19.25
Q = (C * Aorifice * (2 * g * deltaH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5)
Calib = 1 - (0.000071079566 * Q ^ 2 - 0.002182705515 * Q + 0.024449497333)
Q = Q * Calib - leak
Else: Q = "Check Meter!"
End If
H = Ptgage - Yref

'wQtank = 0.0015
wQ = 0.0025 * Q
wLc = (1 / 32) / (2 * 12) '+- 1/64 of inch
wYramp = (1 / 32) / (2 * 12) '+- 1/64 of inch
193

SSUCdRes = ((wQ / Q) ^ 2 + (-wLc / Lc) ^ 2 + (-27 / 8 * wHt / Ht) ^ 2) ^ (1 / 2)
End Function
Function flowt3(Size, dH, g)
Dim beta, a, C As Double
If (Size = 2) Then
C = 0.6345
a = 1.035 ^ 2 * 3.14159 * 0.25 / 144
beta = 0.507
Else
If (Size = 4) Then
C = 0.6277
a = 3 ^ 2 * 3.14159 * 0.25 / 144
beta = 0.7452
Else
If (Size = 10) Then
C = 0.707
a = 10.508 ^ 2 * 3.14159 * 0.25 / 144
beta = 10.508 / 12
Else
If (Size = 12) Then
C = 0.6151
a = 8.005 ^ 2 * 3.14159 * 0.25 / 144
beta = 0.6671
Else
End If
End If
End If
End If
flowt3 = C * a * (2 * g * dH) ^ 0.5 / (1 - beta ^ 4) ^ 0.5
End Function
Function Calc_CdT(Q, Ht, Tlength)
Function_Cd = 3 / 2 * Q / (Tlength / 12 * (32.2 * 2) ^ 0.5 * Ht ^ (3 / 2))
End Function
194
Function Calc_CdE(Q, Ht, Elength)
Function_CdE = 3 / 2 * Q / (Elength / 12 * (32.2 * 2) ^ 0.5 * Ht ^ (3 / 2))
End Function
Option Explicit
'Calculate Specific Weight of Water as a function of Temperature (Fahrenheit)

Function GAMMAH2O(wdTemp)
GAMMAH2O = 59.364982 + 3.0750805 * Cos(0.0078331697 * (wdTemp) -
0.24302151)
'Slight adjustment of gamma to match values given in Engineering Fluid Mechanics 7th
edition by Crowe, Elger, Roberson
If wdTemp = 40 Then
GAMMAH2O = 62.43
ElseIf wdTemp = 50 Then
GAMMAH2O = 62.4
End If
End Function
'Calculate surface tension as a function of temperature (sigma in lbf/ft)

Function SigmaH2o(wdTemp)
SigmaH2o = -5.6230368808E-06 * wdTemp + 0.005376033645
End Function
195
CURRICULUM VITAE

1725 South 1240 West Logan, Utah 84321
Email: bcrookston@gmail.com
Phone: (435) 760-2938
Education
Ph.D. Water Resources Engineering, Department of Civil & Environmental Engineering

Utah State University. Expected Fall 2010. Advisor Blake P. Tullis.
Dissertation: Labyrinth Weirs
M.S. Hydraulic Engineering, Department of Civil & Environmental Engineering

Utah State University. Spring 2008.
Thesis: A Laboratory Study of Streambed Stability in Bottomless Culverts
B.S. Civil Engineering, Department of Civil & Environmental Engineering

Utah State University. Spring 2006. Cum Laude
Minor Spanish, Department of Languages, Philosophy and Speech Communication
Areas of Research Interest
Hydraulic Structures, Ecohydraulics, Free-surface Flows, Water Resources Engineering,

Sedimentation and Erosion, Physical Modeling, Applied Computational Fluid Dynamics
Honors and Awards
2007-2010 Ph.D. Remission Award

2007-2010 Golden Key International Honor Society
2005-2010 Research Assistantships
2008 USSD Scholarship 2008
2006 ASCE/AISC Student Steel Bridge Rocky Mountain Regional Comp. Champions
2005 National Deans List
2005 Dee Hansen Scholarship
2003 Learning Assistance Scholarship
1998 Eagle Scout Award
Affiliations
IAHR/ International Junior Researcher and Engineering Workshop on Hydraulic

IJREWHS Structures, 2010 Organizing Committee Member & Participant, 2008 Participant
ASCE American Society of Civil Engineers, Student Member
USU Chapter Officer 2005-2006
USSD United States Society on Dams, Hydraulics of Dams Committee Member
EWRI Environmental and Water Resources Institute of ASCE, Member
EWB Engineers Without Borders, Student Member
196
Research and Engineering Experience
Research Assistant—Utah Water Research Laboratory, USU, Logan, Utah

• May 2005 – Present
• Experimental Design, Data Collection Procedures, Quality Control
• Research Group Management, Supervision of Physical Model Fabrications
• Reports, Peer-reviewed Papers, Conference Presentations, Funding Solicitations
• Spillways, Fish Passage Culverts, Free-surface Flows, Sedimentation, Erosion, Pumps,
Valves
• Physical Modeling, Computational Fluid Dynamics, Instrument Calibration
Steel Bull Engineering—Senior Design Team, USU, Logan, Utah

• January 2005 – May 2006
• Team Captain – Member of Design Team, Manufacture Team, & Competition Team
• ASCE/AISC National Student Steel Bridge Participants
• First Place ASCE/AISC Rocky Mountain Regional Conference
• First Place USU In-house Competition
Extra-curricular Activities
Engineers Without Borders—USU Peru Team, Logan, Utah

• September 2004 – September 2006
• Potable wells, Pumps, Storage, Water Quality, Recruitment, Donations, Public
Relations
• Collaboration between USU, Chiclayo City Engineers, & Eagle Condor
Humanitarian
Teaching Experience
Teaching Assistant—Depart. of Civil and Environmental Engineering, USU, Logan, Utah

• Fall 2007 - Present
• Substitute Instructor for Faculty
• Lectures in Elementary Fluid Mechanics (CEE 3500), Hydraulic Structures Design (CEE
6540), Hydraulics of Closed Conduits (CEE 6550)
Dynamics (ENGR 2030) Recitation Instructor—Depart. of Civil and Environmental

Engineering, USU, Logan, Utah
• Spring Semester 2008
• Supplementary Lectures, Example Problems, Homework Tutor
ESL Supervisor & Tutor—BYU-Idaho Reading Center & ESL, Rexburg, Idaho
• December 2001 – April 2003
• Certified as Level 2-Advanced Tutor of CRLA, April 17, 2003
• English as a Second Language (ESL), Reading Comprehension, Vocabulary,
Speed Reading, Advanced Study Skills
Languages
English (Native), Spanish (Fluent)

197
Publications
Crookston, B.M. and Tullis, B.P. (Under review, submitted Nov. 2010). Arced and
Linear Projecting Labyrinth Weirs in a Reservoir Application. 34th International
Association for Hydro-Environment Engineering and Research World Congress,
Brisbane, Australia, June 2011.
Crookston, B.M. and Tullis, B.P. 2011. The Design and Analysis of Labyrinth Weirs.
Proceedings of the 31st Annual USSD Conference, San Diego, California, April, 2011.
Crookston, B.M. and Tullis, B.P. (2011). Hydraulic Characteristics of Labyrinth Weirs.
International Workshop on Labyrinth and Piano Key Weirs. Liege, Belgium. Feb, 2011.
Crookston, B.M. and Tullis, B.P. (Under review - submitted March 2010) Incipient
Motion of Gravels in a Bottomless Arch Culvert. International Journal of Sediment
Research.
Crookston, B.M. and Tullis, B.P. 2010. Hydraulic Performance of Labyrinth Weirs. 3rd
International Junior Researcher and Engineer Workshop on Hydraulic Structures,
Edinburgh, Scotland, May 2010
Paxson, G., Crookston, B., Savage, B., Tullis, B., and Lux III, F. 2008. The Hydraulic
Design Toolbox: Theory and Modeling for the Lake Townsend Spillway Replacement
Project. 2008 Association of State Dam Safety Officials, Indian Well, California,
September 2008
Crookston, B.M. and Tullis, B.P. 2008. Labyrinth Weirs. 2nd International Junior
Researcher and Engineer Workshop on Hydraulic Structures, Pisa, Italy, August 2008
Crookston, B.M. and Tullis, B.P. 2008. Scour and Riprap Protection in Bottomless Arch
Culverts. 2008 World Environmental and Water Resources Congress, Honolulu, Hawaii,
May 2008
Crookston, B.M. 2008. A Laboratory Study of Streambed Stability in Bottomless

Culverts. M.S. Thesis, Utah State University, Logan, Utah.
Tullis, B.P. and Crookston, B.M. 2008. Lake Townsend Dam Spillway Hydraulic Model
Study Report. Schnabel Engineering, February, 2008
Crookston, B.M. and Tullis, B.P. 2007. A Laboratory Study of Streambed Stability in
Bottomless Culverts. 2007 World Environmental and Water Resources Congress, Tampa,
Florida, May 2007
Tullis, B.P. and Crookston, B. M. 2007. Dam Outlet Works, More than Just Equations.
Dam Safety 2007, Austin, Texas, September 2007
Crookston, B.M. and Tullis, B.P. 2006. Preliminary Study of Scour in Bottomless
Culverts (FHWA-AK-RD-06-05)
View publication stats

Labyrinth Weirs: December 2010

Uploaded by

Copyright:

Available Formats

Labyrinth Weirs: December 2010

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Labyrinth Weirs: December 2010

Uploaded by

Copyright:

Available Formats

See discussions, stats, and author proﬁles for this publication at: https://www.researchgate.

Thesis · December 2010

Dam Safety View project

Hydraulic Structures View project

The user has requested enhancement of the downloaded ﬁle.

Brian Mark Crookston

A dissertation submitted in partial fulfillment

Civil and Environmental Engineering

UTAH STATE UNIVERSITY

Copyright © Brian Mark Crookston

All Rights Reserved

Brian Mark Crookston, Doctor of Philosophy

Utah State University, 2010

Major Professor: Dr. Blake P. Tullis

elevations and increase flow capacity; nevertheless, it can be difficult to engineer an

discussed. The validity of this method is presented by comparing predicted results to

data from previously published labyrinth weir studies.

A standard geometric design layout for arced labyrinth weirs is presented.

Insights and comparisons in hydraulic performance of half-round, trapezoidal, 6° and 12°

design and analysis method.

Nappe aeration conditions for trapezoidal labyrinth weirs on a horizontal apron

of nappe interference regions to be used in labyrinth weir design. Finally, the

applicability of techniques developed for quantifying nappe interference of sharp-crested

corner weirs is examined.

United States Society on Dams (USSD).

Tullis, for his instruction, support, encouragement, and guidance.

support, expertise, direction, and advice.

Brian Mark Crookston

LIST OF TABLES ...............................................................................................................x

LIST OF FIGURES .......................................................................................................... xii

NOMENCLATURE ........................................................................................................ xix

Background and Motivation ............................................................1

2. BACKGROUND AND LITERATURE ......................................................5

Labyrinth Weirs ...............................................................................5

Analytical Approach ............................................................6

Headwater Ratio (HT/P) .........................................13

Labyrinth Weir Design Methods ...................................................27

Early Investigations ...........................................................27

Labyrinth Weir Case Studies .........................................................35

Amanian M.S. Thesis (1987) .............................................37

3. EXPERIMENTAL SETUP AND TESTING PROCEDURE....................44

Test Facilities .................................................................................44

Rectangular Flume Facility ................................................46

Physical Models .............................................................................52

Materials, Fabrication, and Installation .............................52

Test Procedure ...............................................................................57

4. HYDRAULIC DESIGN AND ANALYSIS

Discharge Rating Curves ...................................................75

Summary and Conclusions ............................................................92

5. ARCED AND LINEAR LABYRINTH WEIRS

Previous Design Methods ..................................................99

Geometric Layout of Arced Labyrinth Weirs ..................104

Summary and Conclusions ..........................................................120

6. NAPPE INTERFERENCE AND NAPPE INSTABILITY .....................123

Nappe Interference for Labyrinth Weirs ..............140

Summary and Conclusions ..........................................................152

7. SUMMARY AND CONCLUSIONS ......................................................156

Appendix A: Schematics of Tested Labyrinth Weir Physical

CURRICULUM VITAE ..................................................................................................195