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

net/publication/256197237

Air pockets in pipeline systems

Article  in  UPB Scientific Bulletin, Series D: Mechanical Engineering · May 2008

CITATIONS READS
4 2,485

2 authors, including:

Diana Maria Bucur

Polytechnic University of Bucharest
66 PUBLICATIONS   127 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Two phase flow (air-water) View project

Water quality View project

All content following this page was uploaded by Diana Maria Bucur on 31 May 2014.

The user has requested enhancement of the downloaded file.

U.P.B. Sci. Bull., Series D, Vol. 70, Iss. 4, 2008 ISSN 1454-2358

AIR POCKETS IN PIPELINE SYSTEMS

Diana Maria BUCUR1, Eugen Constantin ISBĂŞOIU2

Curgerile mixte apă-aer în sisteme sub presiune pot să apară în conductele şi

galeriile amenajărilor hidrotehnice (centrale hidroelectrice, staţii de pompare,
sisteme de colectare a apelor pluviale etc.) provocând variaţii de presiune care au
uneori urmări catastrofale. La centralele hidroelectrice acumularea unor cantităţi
mari de aer se poate produce prin conductele captărilor secundare, galerii de
aducţiune, castele de echilibru, conducte forţate etc. În staţiile de pompare curgerea
bifazică apare la sifoanele utilizate pentru prevenirea curgerii inverse din bieful
superior. Lucrarea îşi propune să urmărească influenţa prezenţei pungilor de aer
din sistemele de conducte sub presiune, asupra presiunii din întregul sistem.

Mixed flows or two phase flows (air-water) in pressurized systems can arise
in pipe systems of hydroelectric schemes and pumping stations, causing fluctuations
in pressure which sometimes can do significant damage. In hydropower plants, the
air entrainment can take place at the inflow location such as drop chamber, inlet or
intake or because of mass oscillation between the reservoir and conduit system.
Siphon outlets of pump discharge channels are frequently used at pumping station
for the prevention of backflow from upstream. For effective operation it is necessary
to evacuate air downstream from siphon. The purpose of the paper is to study the
influence of air pockets trapped in pipeline systems over the entire system pressure.

Key words: two phase-flow, air pocket, transient flow, pipeline system.

1. Introduction

The presence of air in pipeline systems can result in problems such as loss
of carrying capacity, disruption of the flow, reduced pump and turbine efficiency,
effects on pipe materials and pipeline structure; it can also change the fluid
properties and create environmental concerns at the point of discharge, as fallows:
• air pockets reduce the effective pipe cross section, which results in a
reduction in pipe hydraulic capacity;
• the bulk properties of the fluid (a mixture of air and water) are changed;
this concerns mainly the density and the elasticity.
• the presence of air changes the structure of flow turbulence and possibly
the wall shear as well;

1
Assistant, Dept. of Hydraulics and Hydraulic Machinery University POLITEHNICA of
Bucharest, Romania
2
Professor, Dept. of Hydraulics and Hydraulic Machinery University POLITEHNICA of
Bucharest, Romania

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
36 Diana Maria Bucur, Eugen Constantin Isbăşoiu

• air bubbles introduce vertical momentum into the flow due to their
buoyancy and may thus have significant effects on the flow field;
• in hydraulic transients, the presence of large air pockets results in pressure
waves that are strongly damped and deformed. However, it has also been found
that small accumulations of air may have an adverse effect on pressure transients,
actually enhancing the surge pressures experienced;
• air accumulation in a system may lead to disruption of the flow and such
effects as blow-out or blow-back. For instance, air entrained at a hydraulic jump
may not be able to move downstream with the flow and instead ‘blow back’
through the jump. This can lead to vibration and structural damage and can cause
instabilities of the water surface;
• the presence of air can reduce pump and turbine efficiency. When air-
mixed water is fed into a turbine there is a drop in output and efficiency is
reduced. It can also cause waterhammer pressures. Admission of air to a pump
can cause a loss of priming;
• in ferrous pipelines the presence of air enhances corrosion by making more
oxygen available for the process;
• sealing, a transition from part-full to pipe full flow, can cause vibrations of
the structure and surging of the flow can accompany it;
• air can produce false readings on measuring devices;
• cooling water systems have additives in the water for anticorrosion and this
increases foaming of the water.
In order to measure, to control or to dispose the air that is found in
pipelines, it is important to understand the various ways in which air can enter a
pipe system:
• entrainment at the inflow location such as a drop chamber, inlet or intake;
• entrainment at the outflow location – for instance, sea outfalls may operate
under varying tidal levels and the outlet may be located above the sea level;
• entrainment due to vortices at inlet or intake;
• turbulence in shafts;
• hydraulic jump – the flow within a pipe system may change from gravity
to surcharged flow and under these conditions a hydraulic jump may form;
• direct pumping – direct pumping of air into a system may be carried out to
reduce cavitation pressure;
• pumps – there may be insufficient submergence on the pump or vortices
may form at the inlet causing air to be entrained into the system;
• lines filling or emptying – air transport can occur during filling and
emptying of pipelines. The air movement along the pipeline can be slow during
filling and can become trapped at high points in the system;
• gas formation through biological activity;
• at sections under negative pressure air can leak in at joints and fittings;

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
 Air pockets in pipeline systems 37

• negative pressures at the inlet to the pipe.

2. Background

Two phase flow with large air pockets is present in pipes of hydropower
plants, pumping stations or sewer systems, causing many unpleasant situations
and accidents during operation. For example:
• the headrace from Raul Mare, where, because of secondary intakes (Netiş
and Bodu), air is aspired along with water, which lead to apparition of air pockets
that flow gravitationally upstream or forced downstream. Penetration of these air
pockets could lead to serious failures, as it happened in 1998, when the vane
chamber has been damaged because the violent increase of pressure when a very
large air pocket was released [1];
• siphons on the large discharge pipes from pumping stations, which block
the backward flowing of water and draining of downstream basin, are fitted with
air valves. Faulty operation of this valves can lead to formation of air pockets, or
can allow flows with free surface which decrease pump efficiency [2];
• apparition of air pockets in sewer systems because of rapid water filling
(in case of high intensity precipitations, above the designed values of the pipeline
systems). High pressure variation provoked by appearance and releasing of air
pockets can cause major damages in pipes, like the incidents from Minneapolis,
Minnesota State, USA [3] and from Edmonton, Alberta, Canada [4].
Therefore, to avoid these problems it is important to block the entrapment
of air and the movement of the air pockets through the system with high velocity.
If this is not possible, the air pockets have to be taken out of the system.
Considerable costs incur in providing air release valves and chambers, and in
deepening pipe trenches so as to provide the minimum gradients thought
necessary to enable air bubbles and pockets to move towards the valves. If the
accumulated air can be moved hydraulically out of the pipe system, then
potentially large cost savings are possible. At present there are a number of often
contradictory recommendations for assessing how air may be moved through a
pipe system.

3. Governing equations

The change in the volume of an air pocket inside a control volume, within
the horizontal pipe where U is flow velocity can be written as
dVa
= − AU , (1)
dt
where Va is the air pocket volume, A is the cross-sectional area of the pipe, and t is
time.

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
38 Diana Maria Bucur, Eugen Constantin Isbăşoiu

The momentum equation of the water column is

dU H − H0 UU U2
= −g −λ − , (2)
dt x 2D 2x
where H is the air pressure head, λ is the wall friction factor, x is the length of the
water column and g is the gravitational acceleration (9,81 m/s2).
The governing equation for the air phase is [7]
dH * H * dVa H*
= −k =k Qa , (3)
dt Va dt Va
where Qa is the air discharge out of the orifice, H*, is the absolute air pocket
pressure head and k is the polytropic exponent.
From the investigations over the impact of varying the polytropic
exponent, k, (from 1.0 to 1.4) for adiabatic air pressure oscillation in an air
chamber [8] was found that k = 1.4 gave the best fit to the experimental data (the
asterisk denotes the absolute pressure head).
The air discharge, Qa, can be expressed by
ρ ρ ρa0
Qa = μA0Y 2 g ( H * − H b* ) = μA0Y 2 g ( H * − H b* ) , (4)
ρa ρ a0 ρ a
*
where H b is the absolute initial air pocket pressure head, μ is the discharge
coefficient, which can be taken from conventional hydraulic tables for
compressible flows through orifices and nozzles.
The expansion factor, Y, can be expressed as [7]
2/k
k ⎛⎜ H b* ⎞⎟ 1 − ( H b* / H * ) ( k −1) / k
Y= . (5)
k − 1 ⎜⎝ H * ⎟⎠ 1 − H b* / H *
If H*/H*b, is greater than 1.89, the orifice is choked and the discharge can
be calculated by
k +1
ρ * 2 k −1
Qa = μA0 g H k( ) . (6)
ρa k +1
When the water column reaches the end of the pipe, the impact pressure of
the water column can be calculated by
a⎛ a 2 gH 1 ⎞⎟
2
a ⎛a⎞
H 2 = H 1 + ⎜U 1 + − ⎜ ⎟ + 2U 1 + (7)
g⎜ B ⎝B⎠ B B ⎟
⎝ ⎠
where a is the velocity of the pressure wave when the water column reaches the
orifice.
The coefficient B is defined as

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
 Air pockets in pipeline systems 39

2
⎛ A⎞
B = ⎜⎜ ⎟⎟ + ζ − 1 (8)
⎝ A0 ⎠
where ζ is the minor loss coefficient of the orifice which can be neglected since it
is very small compared with A/A0. U1 and H1, are, respectively, the velocity and
pressure head at downstream section.
When applying equation (7), care must be taken in determining the
appropriate value of pressure wave speed, since it is quite sensitive to the air
content in the flow. For example, the wave speed in a pipeline containing water
with 5% air content is only 20% of that without air [9]. Since the air entrained in
the water column during the filling process is difficult to determine, the wave
speed cannot be theoretically calculated. Therefore, a measured value of wave
speed is needed to apply equation (7). Consequently, this analytical model is not a
predictive tool but rather, provides a means of exploring and explaining the
behavior of this phenomenon [4].

4. The influence of air pockets on transient flow in a rapidly filling

horizontal pipeline

The observations from the physical experiments of a rapidly filling

horizontal pipe [5] which was initially dry confirmed that air trapped in a rapidly
filling pipe can induce high peak pressures, especially when air leakage occurs.
The experiments revealed that there are three types of pressure oscillation
patterns in a rapidly filling horizontal pipe, depending on the relative size of the
leakage orifice.
When no air is released or orifice sizes are small, the cushioning effects of
the air pocket prevent the water column from impacting on the pipe end and, from
generating high waterhammer pressures. However, the maximum pressure
experienced may still be up to four times the upstream driving pressure. In this
case, the pressure oscillation pattern has a long period, and the peak pressure
remains relatively constant for a given initial air volume and upstream driving
head.
When the orifice size is very large, the air cushioning effect vanishes and
the water column can easily impact on the pipe end, inducing a waterhammer
impact pressure. In this case, the maximum pressure decreases with increasing
orifice size, since for the larger orifice sizes, water escapes and mitigates the
waterhammer effect.
For intermediate orifice sizes, the pressure oscillation pattern consists of
both long period pressure oscillations (while the air pocket persists) followed by
short period pressure oscillations (once waterhammer dominates). In this case, the
maximum observed pressures increase rapidly with increasing orifice size, since

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
40 Diana Maria Bucur, Eugen Constantin Isbăşoiu

the cushioning effect of the air pocket decreases as the air release rate increases.
The highest maximum pressures (up to fifteen times that of the upstream head)
were observed at the upper limit of this intermediate region, which occurred at a
fairly consistent orifice size.
If the pipe end is sealed or the orifice size is small, the water column
contains a negligible amount of entrained air and approaches the pipe end along
the bottom. The air trapped on top of the water acts as a shock absorber; therefore,
the overall pressure oscillation pattern has a long period. If the orifice is large, the
water column is highly air entrained and the water front is steep. In this case, the
air release is significant and there is no visible air pocket trapped on top of the
flow by the time that the water column reaches the pipe end. Without the trapped
air pocket, the air cushioning effect vanishes and the water column can easily
slam into the pipe end and induce a sharp and short period waterhammer pressure.
If the orifice is not large enough to allow for a significant air release, the
water front contains a moderate amount of air and is steeper than that observed in
the small air release case. After the water reaches the pipe end, some air is still
trapped within the flow as bubbles or pockets whose random behaviors cause the
pressure oscillation to have a long period, irregular pattern. When the last air
pocket is released, the sharp and short period water impact pressure dominated,
although the peak magnitude was mitigated by the earlier air cushioning effect.
In case of rapidly filling horizontal pipe which is partially full, for small
air release situations, the maximum pressure magnitudes increase with the
tailwater depth because the initial air volume is reduced. If the orifice is large, the
tailwater is disturbed especially when the initial water column length is short;
therefore air pockets are trapped and the orifice may be choked, resulting a lower
water impact pressure on the pipe end.
The L-pipe system does not qualitatively alter the pressure oscillation
pattern and maximum peak pressure magnitude and the gravity effect on reducing
the pressure is obvious only when the initial water column is long and the air
release from the vertical pipe is substantial [5]. The T-pipe configuration was
shown to be able to mitigate the pressure surge during rapid filling, and was found
to be especially effective in reducing the impact pressure when the air release was
large.
The experimental results [4] suggest that the pressure peaks due to trapped
air pocket compression and release are certainly high, enough to blow off
manhole covers and explain sewer ruptures. For most manhole covers, the air
leakage is negligible (d/D ≈ 0.002). So, based on the observed results, a
conservative estimate of the peak pressure (3 times the upstream head) would be
400 to 500 kPa. This is at least one order of magnitude greater than the structural
loads that typical urban sewer systems are designed for.

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
 Air pockets in pipeline systems 41

4. Air pockets movement in sloping pipelines

Laboratory tests and numerical studies in which the air moved downward
sloping pipes in the range of 0 to 22.5 degrees (0 to 1/24), concluded [6]:
• air moves freely upward sloping pipes under its own buoyancy with no
flow. The velocities of air pockets in upwards sloping pipes are similar to the air
pocket velocities observed in downward sloping pipes;
• air spreads widely along horizontal pipes with no flow;
• a critical velocity is required to move air pockets along horizontal and
downward sloping pipes.
The equation for estimation of critical flow velocity for air pocket
movement obtained from experimental shows the dependency of the critical flow
velocity on the slope and air pocket size (and implicitly on the pipe diameter too)
[6]. This equation was developed based on a range of air pocket sizes and the
maximum values of critical velocity associated with each of the air pocket classes
used. For engineering applications is advisable the application of a safety factor, Sf
(for which in principle is suggested a value of 1.1).
v
gD
(
= S f 0.56 sin S + c ) (9)

where c equals: 0.45 for n < 0.06

0.50 for 0.06 ≤ n < 0.12
0.57 for 0.12 ≤ n < 0.30
0.61 for 0.30≤ n < 2
and v is the minimum flow velocity required for movement of an air pocket with
size defined by the parameter n in a downward pipe of slope S and diameter D.
4V
n= a (10)
πD 3
where Va is the volume of the air pocket.
The applicability of this equation is as follows:
• downward slopes from 0 to 22.5 degrees (1/2.4). From other experimental
researches is proven that the relation can be valid for slopes up to 40 degrees
(1/1.19) and that beyond this slope the critical flow velocity may start decreasing
with the slope in a power law relationship;
• air pockets with volume of 0.0005 liters to 5 liters, in a 150 mm diameter
pipe. For larger air pockets the required critical velocity for pocket movement
may not increase significantly (for pocket sizes above 5 liters being reasonable a
value for c = 0.61);
• pipe diameter of 150 mm. But considering previous researches, the scale
effects can be neglected due surface tension, and the results can be extended to
larger pipe diameters.

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
42 Diana Maria Bucur, Eugen Constantin Isbăşoiu

Tests with hydraulic jumps in downward circular pipes suggested that the
rate of air expulsion may be given by
Qa
= 0.0025( Fr − 1)1.8 , (11)
Q
where Qa is the air discharge, Q is the water flow, and Fr is Froude number. Other
factors that strongly influence the rate are the actual flow conditions and cross-
sectional shape of the pipe.
Also, is shown that for the same pipe diameter, the velocity of air pockets
movement down the pipe is highly dependent on the pipe slope: the steeper the
slope, the smaller the ratio air pocket velocity/critical flow velocity becomes. The
time required to remove an air pocket in a steeper pipe can be several times
greater than in a mild slope.
In upward slopes, even in very mild slopes (less than 2 degrees) air
pockets would move for static conditions, thus not requiring a threshold velocity.
The presence of air pockets have been shown, in certain circumstances to
cause both high and low pressure fluctuations which are sufficiently large to
potentially cause pipe fracture and pipe failure. This therefore highlights a need
for consideration of the transient wave interaction with entrapped air pockets
during design stage.
A larger pocket of air has the potential to act as an energy accumulator,
which absorbs the transient pressures in a piped system.
A small pocket of air has the potential to severely exacerbate maximum
peak pressures.
Experimental results have highlighted that a small or large air pocket can
be defined in terms of its effect upon pressure transients, but there are limits upon
size/volume, outside of which, these effects do not occur. This therefore suggests
that a critical spectrum of air pocket volume exists for a particular pipeline
configuration and additionally, a further critical size is relative to the actual
location of the air pocket within the pipeline.
Greater pressure enhancement occurs when small pockets of air are placed
towards the upstream section of the pipeline. Larger pockets of air can also
enhance pressures when located downstream locations, depending on pipeline
configuration.

5. Conclusion

There are many ways to reduce or to control large pressure transients in

pipeline systems. It is known from modeling and experimental studies that the air-
induced pressure is severe only when all of the following conditions are satisfied:
• the filling is rapid or the flow backup is intensive and fast;
• the air is able to be trapped;

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
 Air pockets in pipeline systems 43

• there are exits for air release and the size of the exits are intermediate.
Based on these conditions, the following measures are recommended:
• to reduce the risk of flow backup and rapid filling, the inflow rate should
be reduced by appropriate inlet controls. The dimensions of sewer pipes should be
enlarged if economically possible. In-line and off-line storage devices should be
constructed. Adequate standby power and sump storage for pumped systems
should be provided. The design of interceptors and drop-inlets should consider
energy dissipation. The systems should be well maintained to prevent clogging;
• to reduce the entrapment in pipelines, air vents should be placed in the
crown of conduits to release the air moving along the crown of the conduit. The
variation of pipe section should be smoothly transited, and the changes in section
areas should not be large (i.e. the difference of pipe sizes of the adjacent segments
should not be too big).
To reduce the air release pressure, a surge tank or air chamber should be
upstream of the surge location. The size of the air release valve should be
carefully determined to avoid high impact pressure.
From the planning and design points of view, great attention should be
paid to the locations where the air-related pressure surges are likely to occur since
air entrapment and air release are local events. Usually the downstream end of the
sewer system or dead ends formed by the sewer stubs should be carefully planned
and designed since these are the areas in which surge-events are likely.
Appropriate junction design in these areas is significant in reducing pressure
surges.
With respect to the concern of system response to air pressure transients,
drainage sewer systems should be designed in such a manner that small
disturbances in one reach will not be amplified upon entering a succeeding reach.
The natural periods of adjoining reaches must be sufficiently different from each
other.

REFERENCES
[1] G. Tatu, Miscari nepermanente in aductiunile CHE generate de expulzarea pungilor de aer,
Conferinta Sisteme Hidraulice sub presiune, Bucureşti 1999.
[2] A. Minato, N. Nakajima, T. Nagahara, Simulation of two phase flow in pumping stations,
117eme Session du comité scientifique et technique de la Société Hydrotechnique de France;
2001.
[3] K.H. Wang, Q. Shen, B. Zhang, Hydraulic transients and movement of trapped air in sewers,
Dept. Of Civil Environmental Engineering, Univ. Of Huston/2003.
[4] F. Zhou, F. Hicks, P. Steffler, Analysis of effects of air pocket on hydraulic failure of urban
drainage infrastructure, Canadian Journal of Civil Engineering, 86-94, 31/2004.
[5] F. Zhou, F. Hicks, P. Steffler,. Effects of trapped air during rapid filling of partially full pipes,
Annual Conference of the Canadian Society for Civil Engineering Montréal, Québec, Canada June 5-
8, 2002.

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania
 44 Diana Maria Bucur, Eugen Constantin Isbăşoiu

[6] M. Escarameia, C. Dabrowski, C. Gahan, C. Lauchan, Experimental and numerical studies on

movement of air in water pipelines, Report SR 661, Release 3.0, April 2005.
[7] C.S. Martin, Entrapped air in pipelines, Proceeding of the Second International Conference on Pressure
Surges, London, England,FZ- 1 5, 1976.
[8] H.R. Graze, A rational thermodynamic equations for air chamber design "Third Australasian
Conference on Hydraulics and Fluid Mechanics, Sydney, Australia, pp. 57-6 1.
[9] E.B. Wylie, V.L. Streeter , Fluid Transients, McGraw-Hill, New York, 1979.

5th National Conference of Romanian Hydropower Engineers, Dorin PAVEL

22 – 23 May 2008, Bucharest, Romania

View publication stats