A MICRO TESLA TURBINE FOR POWER GENERATION

FROM LOW PRESSURE HEADS AND EVAPORATION DRIVEN FLOWS

Vedavalli G. Krishnan1, Zohora Iqbal1 and Michel M. Maharbiz1
1
University of California Berkeley, CA., US
ABSTRACT
We report on the design, fabrication and testing of a
low pressure head Tesla microturbine. We began
developing this technology as a means of scavenging
energy from fluids flows induced in plant-like
evaporative systems. Unlike traditional inertial turbines,
Tesla turbines have high efficiency when driven with low
pressure flows, are relatively simple to manufacture and
scale down very favorably. The 1 cm3 rotor diameter
turbine presented here is, to our knowledge, the smallest
Tesla turbine reported, with an unloaded peak power of
45 mW (12 cc/sec flow, 17% efficiency) and a peak
efficiency of 40% (< 2 cc/sec flow). Moreover, the entire
turbine is built using a variety of modern commercial
rapid prototyping methods, making its construction
accessible to almost anyone. Beyond applications in
evaporative scavenging, Tesla microturbines may find
use as components in ultrasmall-profile heat engines and
for energy generation from sources of low pressure head
flow.

KEYWORDS
Tesla turbine, viscous turbine, miniaturize turbine,
power MEMS, microturbine.

INTRODUCTION
The cohesive properties of water enable the ascent of
sap to the top of trees against gravity and frictional losses,
driven by evaporation at microscale pores in leaves. For a
100 m tree, this corresponds to a minimum pressure
difference of 10 bars between leaf and root [1], and with a
plant evaporation rate of 5nl/cm2/sec, a power of
15W/cm2 and an energy density of 3 kJ per kg of
evaporated water. Earlier work scavenged energy from
evaporation-induced water flows by charging pumping a
circuit via dielectric-water interface transition between
capacitor plates [2]. In this work, we present a
microturbine which can be driven by evaporative flow
(Figure 1).

Figure 1 (left) power generation concept;(right) turbine

concept. Fluid entering through the inlets spirals inwards
between disks, transferring power to the rotor shaft. Fluid
exists through holes near the center of each disk and
downwards out of the turbine.

Our aim is to design miniaturized turbines (1 25 mm

diameter) that are capable of producing 1 mw 10 W
power outputs. The 10 mm diameter turbines we present
here operate at low Reynolds numbers (NRE ~ 1 15)
corresponding to laminar flow and they transfer energy
using the drag force of viscosity and the adhesive nature
of the flowing fluid. At micro-scale, the surface
area-to-volume ratio increases and surface tension,
adhesion, and cohesion forces play a bigger role
compared to inertial forces. Thus, rotors that use
kinematic viscosity and surface effects (rather than
inertia) become a good choice for micro-scale power
extraction machinery. Previous research work on turbine
scale-down by R.T. Deam et al. [4] has shown that
viscous turbines outperform conventional impact-based
turbines as they are scaled down to millimeter range. In
this paper we present background, theory, fabrication and
test results of our turbines.

THEORY AND SIMULATION

Basic operation of Tesla turbines
In Tesla turbines (Figure 1, right), the adhesion and
viscosity of a moving medium are used to propel closely
spaced disks into rotation. The fluid enters the inner space
between the disks from the periphery and exits through
central holes near the axle (dotted lines). There are no
constraints or obstacles intended to couple inertial forces
(i.e. vanes) as in traditional turbines. The fluid enters
tangentially at the periphery and makes several
revolutions while spiraling towards the central exhaust
(dotted lines). During this process, it transfers momentum
to the disks. Under ideal conditions, there is no slippage
between the tip velocity of the rotating disks and the
tangential velocity of the fluid entering the disks at the
nozzle exit. The efficiency of energy transfer is largely
governed by the smoothness of the medium flow from the
nozzle to the disks (effectively a fluidic impedance
matching problem), the effectiveness of the bearing in
reducing the friction loss, and the size of the active area
for the transfer of the momentum.
Theoretical efficiency
We measure the turbines expansion efficiency, also
known as isentropic or component efficiency. Here the
work output is derived from the moment of inertia of the
rotor and the rotor acceleration and deceleration
characteristic at a given flow rate. The work input is
calculated from the flow rate and the pressure drop across
the turbine.
W. Rice [5] published the first extensive theoretical
work on Tesla turbines, providing results from numerical
simulations of fluid-disk interactions. More recently,

Romanin et al. have provided analytical solutions for

Tesla turbine operation suitable for the regimes tested
here [6].
From Rice et al., the theoretical fluidic-to-mechanical
rotor efficiency can be as high as 80%. The performance
is governed by the rotor, nozzle and fluid characteristics.
Rotor radius, exhaust/rotor radius ratio, and exhaust area.
govern the effective rotor area. The nozzle dimensions
and nozzle positioning affects the nozzle loss and the
nozzle-to-rotor interactions. The kinematic viscosity and
density of the fluid influences the energy transfer. The
bearing and any seals influence the losses. The flow rate
controls the power output and the smoothness of the flow.
A complete description of the analysis is outside the
scope of this paper; see [5, 6]. Applying the Rice et al.
results to our system, theoretical specific power was
calculated for a 1 cm diameter rotor with 20 disks spaced
125 m apart (Figure 2). Table 1 compares the predicted
performance of three different systems: a micro turbine (1
mm disk diameter), a mini turbine (1 cm disk diameter),
and a mini turbine driven with 20 cm3/sec steam at 0.1 bar
pressure.

bearings (1.25 mm OD, Bird Precision, Waltham, MA)

connect the shaft to the housing. These perform well at
<10000 RPM.

Figure 3: (top left) various

photo-etched stainless steel
disks, bronze square axle;
(top right) white light
microscopy (20x) showing
125 m disk and
post-assembly gap
uniformity of rotor stack.
(bottom right) assembled
three 1cm and 2 cm rotors

Power Watts - 80% efficiency

15

10
8

05
0.

Flow rate in cc/sec

Table 2: Rotor Specifications

0.
2

0.

0 .1

0.05

12

0.0

Rotor

Disks

1
2
3

20
20
13

Gap

Exhaust/Rotor radius

0.1

4
2

0. 1

0.
1

h =125 m
h =125 m
h =250 m

ri / ro = 0.47
ri / ro = 0.51
ri / ro = 0.47

Exhaust/Disk
Area

0.105
0.143
0.105

0.05
0.05

0.1
0.15
Head in bar

0.2

0.25

Figure 2: Theoretical maximum specific power (W/cm3)

at a range of flow and pressures for a 1 cm, 20 disk,
125m spacing rotor
Table 1: Specifications and theoretical performance
System

micro
mini-W
mini-S

Flow

Diameter Spacing Isentropic

Efficiency
volume/s
mm
m
%

5l
4cc
20cc

1
10
10

25
125
125

78
74
45

Power
mW

Power
density
mW/cc

0.04
31
1450

8
60
2680

FABRICATION AND TESTING

Turbine fabrication and assembly
Disks of 1 and 2 cm diameters with three different
center exhaust hole patterns were fabricated using
commercial photo etching (Microphoto, Inc., Roseville,
MI) on 125 m thick, 300 series full hard stainless steel
sheets (Figure 3, Table 2). A square axle with rounded
ends was used to enable automatic alignment of the disks.
The spacers were 125 m thick.
We assembled four different rotor stacks with 1 cm
diameter disks: two with 125 m inner disk spacing but
with different exhaust holes designs, one with 250m
spacing, and one with 500m spacing. The number of
disks in the rotor assemblies varied (20, 13 and 8,
respectively) to fit in the same enclosure. The rotors were
held tight by two screws on either side. Ruby Vee

Figure 4: CAD view of enclosure with the 8 nozzles tested.

Center rotor housing diameter:1.013 cm
All noz entry holes diameter :4.04 mm; Nozzle exit info:
Table 3: Nozzle Specifications
Nozzle

Type

Inlet Area
(% of rot area)

Area
(mm2)

Inlet angle
( to tangent)

4,8

Converging
circular
Converging
circular
Converging
circular
Circular
array
funnel

4%

3.28

4%

3.28

15,25,35

2.9%

2.28

0.8%

0.69

15

9%

7.14

15

1,2,6
3
5
7

Testing and characterization

Figure 5 shows the test setup. A gear pump
(EW-74014-40, Cole-Parmer Instrument Company,
Vernon Hills, IL) was used to produce 1 20 cm3/sec
flow rates while the pressure at the nozzle inlet was
measured (DPG8000-100, Omega Engineering, Inc,
Stamford, CT). During operation, the rotation of the
turbine was recorded using a high speed video camera
(FASTCAM-X 1024PCI, Photron, San Diego, CA using
PFC Viewer software). Thermocouples at the top and
bottom of the enclosure (5SC-TT-K-40-36, Omega
Engineering, Inc., Stamford, CT) monitored turbine
temperature.
Figure 5:
Gear pump draws
water from a tank
and drives the rotor.
The nozzle inlet
pressure is
measured using a
gauge and the rotor
movement is
recorded using high
speed camera.

TEST RESULTS AND DISCUSSION

Test data and operation verification
Eight systems with different nozzle and 1cm rotors
are tested with the pump system as shown in Figure 5
Pressure vs. flow rate measurements were carried out
for all the systems (Figure 6). It is observed that the
pressure head for a particular flow is mainly determined
by the nozzle and the three different rotors produced only
small variations in the pressure head.
The Reynolds number (NRE) is calculated using the
rotor spacing and RPM at flow rates from 2 cc/sec to 20
cc/sec; it is < 15 for the 20 disk stacks and < 40 for the 13
disk stacks (Figure 6).
(1)
NRE = 2 f h2 /
Where is the kinematic viscosity, is the density, f
is the rotor revolution/sec, and h is the space between the
disks. For our pressures and flows, Reynolds numbers
varied from 0.5 to 12 for 20 disk rotors R1 and R2; It
varied 2 - 42 for 13 disk stacks with double the disk gap.
In all but one case, only one nozzle was used to inject
fluid during operation (see also Summary). The tested
systems specifications are listed in Table 4, along with
peak performance. Maximum performance for the first

four systems is in Table 5; the maximum power output

with nozzle 3 and rotor 2 (N3-R2) was 55mW at 9%
efficiency whereas nozzle 4 with rotor 1 (N4-R1) had the
best performance of power out of 45mW at 17.3%
efficiency.

0.3

Turbines: Head vs. Flow

5-R2

0
0

20

0.2
0.1

NRE vs. Flow and Head

NRE

0.4
Head in bar

Nozzle impedance mismatch is known to contribute

to large performance degradation in turbines and is
especially important for turbine of this kind [3, 5]. To
explore the nozzle parameter space, we used 3D plastic
rapid prototyping (ProtoTherm 12120 polymer, 0.002
layer thickness, High-Resolution Stereolithography 3,
FineLine Prototyping, Inc., Raleigh, NC) which allowed
us to build designs which would otherwise be
un-machinable. Seven different nozzle types (Figure 4,
Table 3) were tested on rotor performance.

1
4+8
10
Flow in cc/sec

10

0
20

7
20

10
Flow cc/sec

5
0.5
0 0

Head in bar

Figure 6: (left) Pressure head vs. flow for the eight systems
tested. (right) Reynolds number for the tested nozzles at
different flow rate and heads for rotor 1.
Table 4: Six different nozzles and three different rotors were
tested. Data from N3-R3 is used for Figures 8-10; see Table 2
for rotor specifications.
Nozzle# Flow
P Rotation NRE Power
eff
Rotor#
(cc/s) (bar) (rpm)
(mW)
(%)
N1-R1
9
0.11
3499
5.7
10.1
10.5
N3-R1
8
0.15
5590
9.3
20.3
18.4
N3-R2
8
0.13
5264
8.6
19.8
19.7
N3-R3
10
0.19
6522
43
16.9
9.3
N4-R1
12
0.23
7247
12
45.0
17.3
N4+8-R1
14
0.19
6977
11
29.0
10.9
N5-R1
6
0.29
4639
7.6
13.0
8.1
N7-R1
12
0.17
5807
9.5
23.2
11.9
Table 5: Maximum efficiency for four systems. As expected,
highest efficiencies are achieved at lower flow rates and
pressure (see Figure 10).
Nozzle# Flow
P
Rotation NRE Power
eff
Rotor# (cc/s) (bar)
(rpm)
(mW)
(%)
N3-R3
2
0.01 1243
8.1
0.4
36.6
N3-R2
2
0.01 689
1.1
0.32
27.0
N3-R1
5
0.06 3488
5.7
0.87
22.0
N1-R1
6
0.05 2190
3.6
3.6
12.5
N3-R2
15
0.43 9678
16
54.8
9.2

Data analysis and results

Accelerating and decelerating angular velocities
were computed from video data by performing 2nd order
polynomial curve fits on the frequency vs. time data and
extracting the fitted curves slopes at given frequencies
(Figure 7). As the rotor is accelerating while suffering
bearing loss, the sum of the angular acceleration and
angular deceleration magnitudes is used in the calculation
of unloaded torque and work done [3]:
(2)

= J (1 2)
Pout
=2f
(3)
Pin
= Jflow * Pturbine
(4)
(5)
efficiency = Pout / Pin
where is the torque (Nm), J is the moment of inertia of
the rotor (kg m2) which was derived from the geometry of
the rotor components, 1 and 2 are the magnitude of
acceleration and deceleration and f is the rotor rotational

frequency (rev / s).

Figure 8 shows the frequency, acceleration,
deceleration and unloaded torque curves for N3-R3 at 10
cc/s flow rate. Figure 9 shows the torque, power output
and efficiency vs. rpm for the same system at different
flow rates. Figure 10 compares the power output and
efficiency vs. flow rate for four systems; N1-R1, N3-R3,
N3-R2, N3-R1.

nd

Figure 7: Sample raw video data (+), 2 order

polynomial curve fits for the acceleration (solid) and

deceleration (dashed); slopes are 1 and 2. Torque ()
and power output is calculated from 1 and 2.

50

Acceleration

Frequency
100

Deceleration

50

-1000

Efficiency vs. Flow

40

N3-R2
40

20

Effieciency (%)

Deceleration

2000

150

Frequency

100

Power Out vs. Flow

60

Decelerating Rotor

4000

Frequency

Acceleration

Frequency

150

Key findings and next steps

1. Single nozzles exhibited over 20% variation in
efficiency within the tested range. In limited tests
with 2 nozzles placed at 180 degrees from each other
(using nozzles 4 and 8) we did not see a performance
improvement. Further tests are needed at lower flow
rates and with other nozzles.
2. Maximum efficiency was achieved at low flow rates.
The 13 disk rotor stack (rotor3) realized 36%
efficiency for 2 cc/sec flow rate at 0.4 mW power.
3. Higher gap h (rotor 3) and higher inner to outer
radius ratio ri / ro (rotor 2) moved the efficiency
peak to lower flow rates (with respect to rotor 1).
4. Nozzle 4, with the tangential entry angle to the rotor
stack and an exit area 4% of the rotor inlet area (for
rotor 1) achieved the highest power (45 mW) with
17% efficiency for 12 cc/sec flow rate.
5. High exit area (9% of rotor inlet area) and low exit
area (0.8%) nozzles resulted in about 50% lower
efficiency than the peak efficiency area (4%) nozzle.
Power Out (mW)

Accelerating Rotor

SUMMARY

N3-R1
N3-R3
N1-R1

0
0

10

Flow in CC/sec

15

N3-R3
N3-R2
N3-R1

30
20
10

N1-R1
0
0

15

Figure 10 : (left) power out vs. flow; (right) efficiency vs. flow
at maximum rpm for four systems.

Moving forward, we plan to model the loss

mechanisms and perform parametric optimization of the
-5
Torque vs. RPM
Figure 8: (top left) rpm (solid) design to enable 0.1 - 2 cm range Tesla turbine designs
x 10
6
and angular acceleration
Total Torque
for given flow rate, head and power requirements.
(dash). (top right) rpm(solid)
4
REFERENCES:
and angular deceleration
Accelerating Torque
[1] L. Taiz & E. Zeiger, Plant Physiology, Sinauer
(dash) (bottom left)
2
Associates,
Inc., 4th edition, 2002.
acceleration , deceleration and
abs(Decelerating Torque)
[2]
Ruba
T. Borno, Joseph D. Steinmeyer, and
unloaded torque vs. rpm for
0
1000 2000 3000 4000 5000 6000
Michel M. Maharbiz, Charge-pumping in a synthetic
N3-R3 at 10 cc/s flow.
RPM
leaf for harvesting energy from evaporation-driven
flows , Applied Physics Letters, 94,1 (2009)
PowerOut vs. RPM
-5
Torque vs. RPM
x 10
0.03
6
[3] A Guha, B. Smiley, Experiment and analysis for
10 cc/s
an improved design of the inlet and nozzle in Tesla disc
0.02
4
10 cc/s
turbines, Proc. IMechE Vol. 224 Part A: J. Power and
7 cc/s
7 cc/s
Energy, 261-277, Sept 2009
2
0.01
5 cc/s
5 cc/s
[4] R.T.Deam et.al., On Scaling Down Turbines to
2 cc/s
2 cc/s
0
0
0
2000
4000
6000
Millimeter Size, Journal of Engineering for Gas
0
2000
4000
6000
RPM
RPM
Turbines and Power, v 130 / 052301-9, September 2008.
[5] W Rice, Tesla Turbomachinery, Handbook of
Efficiency vs. RPM
Turbomachinery,
CRC Press, 1994.
50
Figure
9:
(top
left)
torque
vs.
[6] Vince Romanin, Van P Carey, Strategies for
40
2 cc/s
rpm, (top right) power out vs.
30
performance enhancement of Tesla turbines for
7 cc/s
5 cc/s
rpm, (bottom left) efficiency vs.
20
combined heat and power applications, Proceedings of
rpm at different flow rates for
10
the ASME 2010, 4th International conference on energy
10 cc/s
N3-R3.
0
sustainability, ES2010-19251, May 2010.
0
2000
4000
6000
0

1.4

1.6

1.8

0
16

-2000
2

16.5

-2000
17

Time Secs

Torque (Nm)

Power Out (W)

Torque (Nm)

Time Secs

Efficiency %

10

Flow in CC/sec

RPM

CONTACT
Vedavalli Krishnan, tel: 510-755-9640;
vedavalli@berkeley.edu