You are on page 1of 22

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

net/publication/326057381

Experimental Evaluation of Advanced Archimedes Hydrodynamic Screw


Geometries

Article  in  Journal of Hydraulic Engineering · August 2018


DOI: 10.1061/(ASCE)HY.1943-7900.0001479

CITATION READS
1 1,572

4 authors, including:

Samuel Harding Dirk Nuernbergk


Pacific Northwest National Laboratory 82 PUBLICATIONS   177 CITATIONS   
28 PUBLICATIONS   131 CITATIONS   
SEE PROFILE
SEE PROFILE

Chris Rorres
Drexel University
65 PUBLICATIONS   1,590 CITATIONS   

SEE PROFILE

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

Archimedes in the 21st Century View project

Turbulence measurements from velocimeters on compliant mid-water moorings, part 2: motion correction View project

All content following this page was uploaded by Chris Rorres on 24 December 2018.

The user has requested enhancement of the downloaded file.


Experimental Evaluation of Advanced Archimedes
Hydrodynamic Screw Geometries
Jerry L. Straalsund 1; Samuel F. Harding 2; Dirk M. Nuernbergk 3; and Chris Rorres 4

Abstract: There is increasing interest in the application of Archimedes hydrodynamic screws (AHS) to generate hydroelectric power.
Although numerous theoretical and experimental investigations have been published on this application of AHS over the last decade, there
have been no experimental evaluations in the literature of blade shapes other than the helicoid. This paper provides the results of bench-scale
testing of three-dimensional (3D)-printed models to evaluate the performance of various shapes and parameters for AHS. Three different
screws are investigated to experimentally compare the new blade shapes: the helicoid, strake, and reverse strake. DOI: 10.1061/(ASCE)
HY.1943-7900.0001479. © 2018 American Society of Civil Engineers.
Author keywords: Archimedes screw; Water power; Hydropower plant; Low head hydropower; Low head turbine; Flight shapes; Helicoid;
Strake; Gap flow; Hydrodynamic screw.

Introduction In 2012, one of the authors [C. Rorres, “Methods and apparatus
for moving fluid using a strake,” Patent No. WO 2014047619 A2
Since Archimedes’ time 23 centuries ago, the blades of an Archi- (2014)] proposed that a blade with the shape of a surface known as
medes screw have had the shape of a helicoid. This shape was used a strake has several possible advantages over a helicoid. Fig. 2 com-
in the past mostly for ease of manufacture. Blades (or flights) of pares the profiles of a strake and reverse strake, together with the
ancient screws were constructed by layers of long, pliable twigs helicoid blade shape. The reverse strake in Fig. 2(c) is simply the
nailed together along a helical path and covered with tar pitch strake in Fig. 2(a) rotated by 180° about a line perpendicular to
(Vitruvius 1914). Later, the ends of wooden staves were inserted the shaft axis, or turned upside down. Fig. 2 shows the experimental
into a helical groove in the central shaft [Fig. 1(a)]. Both techniques orientation of the three screws in which the water moves from the
are difficult to apply if the blades have a shape other than that of a top to the bottom.
helicoid. Today, most manufacturers of small to medium-sized Rorres [“Methods and apparatus for moving fluid using a
screws use metal flights made by welding individual sections strake,” Patent No. WO 2014047619 A2 (2014)] showed that one
formed from flat, slit, annular, metal blanks. These blanks are of the theoretical advantages of a strake is that the resulting screw
formed into a helical shape by hammering, cold pressing, or the has a greater calculated bucket volume per turn, V B , than does a
use of rolling machines [Fig. 1(b)]. helicoid screw. The bucket volume per turn refers to the volume of
The first documented tests using an Archimedes screw for water in one pitch of the screw contained within the buckets of
power generation (rather than the lifting of water) were performed water between the blades when they are filled to maximum capacity
by Brada (1999). Since then, the use of Archimedes hydrodynamic without overflowing. This assumes that the strake screw and heli-
screws (AHS) has spread rapidly worldwide (Lashofer et al. 2013) coid screw have the same outer radius Ra , inner radius Ri , pitch S,
and many more tests and optimizations have been accomplished to inclination β, and number of blades N. The authors hypothesized
improve their performance. However, all reported previous testing that because the strake blade provides a greater bucket volume per
has been performed with screw blades that were helicoidal in shape turn than does a helicoid blade, for a given inlet flow the strake
(e.g., Dellinger 2015; Dellinger et al. 2016; Erinofiardi et al. 2017; screw could operate at reduced speed, thus reducing frictional
Kibel and Coe 2011; Kozyn and Lubitz 2017; Lashofer et al. 2011; losses and increasing efficiency. Although the strake blade presents
Lisicki et al. 2016; Lyons and Lubitz 2013; Lubitz et al. 2014; a greater wetted surface than a helicoid blade, the frictional losses
Nuernbergk 2012; Rohmer et al. 2016; Saroinsong et al. 2015). are very sensitive to the rotational speed; specifically, Muysken
(1932) showed that the power loss because of the wetted friction
of the blades is proportional to the fifth power of the rotational speed.
1
President, Percheron Power LLC, 6855 W. Clearwater, A101-260, Consequently, a decrease in the rotational speed in the case of the
Kennewick, WA 99336. Email: jls@percheronpower.com strake blade should increase its efficiency over that of a helicoid
2 blade even though the strake blade’s wetted surface is greater.
Research Engineer, Pacific Northwest National Laboratory, Richland,
WA 99354. Email: samuel.harding@pnnl.gov
3
Technical Consultant, D-99084 Erfurt, Germany. Email: dirk.nuernbergk@
gmail.com Objectives
4
Professor Emeritus of Mathematics, Dept. of Mathematics, Drexel Univ.,
Philadelphia, PA 19104 (corresponding author). Email: crorres@cs.drexel This paper analyzes the differences in performance for hydropower
.edu
Note. This manuscript was submitted on July 26, 2017; approved on
generation of the three blade shapes in Fig. 2 using scaled test
December 21, 2017; published online on June 13, 2018. Discussion period screws. The objective of the experiment was to measure and com-
open until November 13, 2018; separate discussions must be submitted for pare the performance of the three blade shapes to determine if the
individual papers. This paper is part of the Journal of Hydraulic Engineer- theoretical advantages of the strake blade could be experimentally
ing, © ASCE, ISSN 0733-9429. confirmed. The experimental setup and the measurement results are

© ASCE 04018052-1 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Fig. 1. (a) Manufacture of wooden replacement screw for an old Dutch windmill (image courtesy of Harmannus Noot); and (b) modern
flight-forming machine (image courtesy of Advanced Spiral Technology).

Fig. 3. Side view of Archimedes screw filled just to the top of the
Fig. 2. Three tested blade shapes: (a) strake; (b) helicoid; and central tube (without any overflow over the central tube). In the case
(c) reverse strake. illustrated, the water level entering from the left touches the maximum
filling point F for the first completely formed bucket of water between
the blades.
presented subsequently, followed by the methodology and results
of the comparisons of the efficiencies and other performance
characteristics for the three screw types. Table 1. Geometric parameters of three test screws
Appendix S1 provides details of the experimental setup and cal-
ibration and also provides details of the manufacture of the screws Parameter Symbol Value
using three-dimensional (3D) printing as an easy-to-use prototyp- Number of blades N 4
ing technology. Lee et al. (2015) previously used such 3D printing. Outer radius Ra 95.7 mm
Appendix S2 gives the mathematical descriptions of the screw cross Inner radius Ri 47.8 mm
sections needed to create the computer files for fabrication of the Pitch S 191.4 mm
Inclination angle β 20.6–29.1°
screws using this technology. Appendix S2 also contains math-
Gap G 0.4 mm
ematical descriptions of the zero-thickness surfaces defining these Total length — 766 mm (4 pitches)
three blades and the upper and lower edges of the blade profiles
when they have nonzero thickness.
Fig. 3 also illustrates an idealized concept of the flow behavior
Geometric Parameters of Three Test Screws of an Archimedes hydrodynamic screw. The water is essentially
transported by buckets of water that move with the rotation of
Fig. 3 illustrates the basic geometric parameters of an Archimedes the screw. The volume per turn, V U , is the volume of water con-
screw, including inner radius Ri , outer radius Ra , number of flights tained in one pitch of the screw, which is generally different from
N, pitch S, and inclination β. For the three screws tested, these five the previously defined bucket volume per turn V B . The ratio φ ¼
parameters were fixed throughout the tests at the common values V U =V B of these two volumes is called the filling factor. The screw
listed in Table 1. is said to be underfilled if φ < 1 and overfilled if φ > 1.
Note that Ri =Ra ¼ 0.5 and S=2Ra ¼ 1, which are common Although it may first appear that screws should be sized so that
ratios used in commercial hydropower screws today. Initially, three φ ¼ 1, in commercial practice the design and selection of the tur-
different inclinations of the screw were tested (20.6, 23.9–24.2, and bine represents a balance between the capital cost, desired flow
29.1°). The results presented in this paper are limited to an averaged capacity, and possible physical site constraints, as well as hydraulic
inclination angle β ¼ 24°. losses due to friction and leakage losses. In an idealized scenario

© ASCE 04018052-2 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Table 2. Bucket volume per turn, V B , for three test screws an additional leakage flow will develop when the water overflows
Screw type V B (L) the top of the center tube (Point F). Parameter QL is defined as the
total leakage flow that occurs when the screw is prevented from
Strake 1.911
rotating and a water level, hin , is maintained upstream. For each
Helicoid 1.873
Reverse strake 1.780
test screw, the leakage flow, QL , was measured as a function of
the inlet water height hin .
The equations and methodology utilized to compute and com-
pare the performance of the three screws based upon the measured
for a given flow capacity and turbine dimensions, as φ increases values from the experiment are provided in subsequent sections
toward 1 the rotational speed to pass the same flow decreases, with the experimental results.
which improves efficiency due to reduced friction and turbulence.
However, as φ becomes greater than 1 (the screw is overfilled),
some of the flow spills over the top of the center tube decreasing Test Setup
the efficiency. Typically, the larger the turbine diameter, the greater
are the flow capacity and the capital cost. In some cases, a smaller Figs. 5–7 show the test setup used for the data measurements.
diameter turbine which operates at φ > 1 may deliver the maxi- Appendix S1 provides additional details. The basic test article con-
mum power per unit cost invested for a given flow rate, albeit sisted of the frame, Archimedes screw, bearings, and dynamometer
at a lower efficiency. This maximum-value analysis for commercial components (Fig. 5). The dynamometer components consisted of
plants usually occurs when φ is somewhere between 0.8 and 1.2 an encoder and a magnetic clutch attached to a lever arm with a load
under normal flow conditions. Nuernbergk (2012) and Nuernbergk cell. Appendix S1 provides the details of all the instrumentation
and Lashofer (2016) further discuss this point. used to obtain water levels, torque, speed and flow rate, as well
Table 2 lists the bucket volumes per turn V B of the three test as calibration results of the flow meter and torque measurement
screws computed using MATLAB R2017b programs written by system. Taken collectively, the uncertainty in these instruments
the authors (Rorres 2000). contributed to an uncertainty of 2–3% in calculated power and ef-
Fig. 4 shows cross sections of the three test screws as viewed ficiencies for tests ran at midrange power levels. Bearing friction
for clockwise rotation from above the screw. Each test screw had losses were not measured, but were judged to be negligible com-
5.1-mm-thick blades, and the cross sections of the blades are shown pared with the torque generated by the turbine over the range of
to scale. experimentation. The ceramic hybrid bearings were designed for
high-performance racing bikes and are noted for their exceptionally
low friction. During setup, the turbine rotated freely on the bearings
Measured Parameters of Experiment with little noticeable force.
The test station (Figs. 6 and 7) included two pumps with asso-
The experiment was focused on accurately measuring and record- ciated piping and valves. Water was pumped from the reservoir
ing the following four parameters over a range of operating through a calibrated flow meter, over an upper weir, and into the
conditions for the three test screws: upper bay. The water flowed through the Archimedes screw into the
• rotational speed, n; lower bay, over the outlet weir, and back into the reservoir. Flow
• torque, T; rates were established through the use of one or both pumps and
• total flow rate, Q; and controlled using the bypass and throttle valves. The purpose of
• head difference, H. the upper weir was to minimize the dependence of pump head
The head difference, or geodetic head, H was determined from and associated flow rates on the water level in the upper bay. Two
measurements of the inflow water height hin and the outflow water 7.8-cm-diameter stilling wells (not shown) were fitted with ultra-
level hout (Fig. 3). The inflow and outflow speed of the water were sonic level sensors and connected to the upper and lower water
negligible. measurement ports via flexible polymer tubing.
An additional parameter, the leakage flow, QL , was also mea-
sured. When the screw in Fig. 3 is prevented from turning and
water is introduced upstream, water will leak through the turbine Experimental Procedure
depending upon the upstream water level, hin . At low values of
hin , the water will leak through the gap, G, between the screw The test article was placed in the test stand as illustrated in Fig. 7.
and the trough. If the upstream water level is increased enough, The elevation of the lower end of the test article was adjusted to set

Fig. 4. Cross-sectional views of the three test screws drawn to scale: (a) strake; (b) helicoid; and (c) reverse strake.

© ASCE 04018052-3 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


attained by setting the outlet weir height. Although the outlet water
level, hout , varied slightly (2%) with the flow rate over the flow
rate range of the experiment, the effect was small and was the same
for all three shapes, and calculations of head and efficiency were
based on the actual measured values of hin and hout .
The first series for a given shape and inclination was started at the
lowest flow rate. The first run in the series was initiated by setting
the brake voltage to a low value, starting the pump(s) and allowing
the system to come to a steady state in which the torque, upper and
lower water levels, flow rate, and rotational speed were constant.
After reaching a steady state, the system was operated for an addi-
tional 2 min while data were recorded to a file at a rate of 1 Hz.
Each data set included upper and lower water levels, torque,
speed, and flow rate. After the data record file was saved, the brake
voltage was incremented to a higher level. The system response was
Fig. 5. Assembled test article with Archimedes screw, trough, and a decrease in speed, an increase in water level in the upper reservoir,
dynamometer. and an increase in torque. After attaining steady state, data were
recorded in a new file for an additional 2 min. This process was
repeated with additional increments in brake settings until the
torque became too high for the Archimedes screw to turn. The
pump and valving configuration were then changed to achieve a
higher flow rate, the brake voltage was reset to a low voltage,
and a new series of runs was initiated. For the purpose of analyzing
the steady-state performance of the various test screws, the data
were averaged over 60-s periods for each data set. Appendix S1
describes the quality control procedures used to ensure that these
60-s periods did not include any transient effects. Special care was
taken to adjust and calibrate each measurement device of the test
bench. Appendix S1 further details the test bench, measurement
equipment, and calibration methods, including the precision or
resolution of each device.
Over 600 data sets were obtained for the three test screws. in
Table 3 summarizes the range of parameters and resulting data sets.
Fig. 6. Test setup with test article in test station. Dimensions are in
centimeters.
Experimental Measurement Data for Strake Blade
Figs. 8–11 illustrate the measured inflow height, torque, and cal-
culated mechanical power and efficiency as a function of rotational
speed for the strake blade test screw. Each figure shows seven
different curves in which the flow rate Q was kept fixed at seven
different values.

Interpolation of Measurement Data at Constant


Inflow Height

The curves with the seven filled circles or dots in each of Figs. 8–11
represent constant hin and identify the points on each flow-rate
curve at which hin ¼ Ra . The values of these points were found
by linear interpolation over two measured values of hin (those
immediately above and below Ra ) on each flow-rate curve. From
previous work, it was found that a filling factor φ ¼ 1 occurs at
about an inflow height of hin equal to the outer radius Ra for an
inclination angle of β ¼ 24° (Nuernbergk 2012). This requirement
Fig. 7. Test setup with Archimedes screw installed. (hin ¼ Ra ) was chosen to compare the data of the three screws in
the following section.
Fig. 8 displays the measured inflow height hin versus the rota-
the angle of inclination, which was measured with a calibrated tional speed n. This figure demonstrates that the inflow height de-
digital inclinometer. Data were collected for multiple series of test creased with increasing rotational speed when the flow rate Q was
runs. Each series was performed over a range of torques and speeds kept constant. Very similar curves were found for the helicoid and
while the other variables (Archimedes screw shape, inclination the reverse strake test screws.
angle, lower water level, and flow rate) were held at nominally Fig. 9 shows the measured torque T as a function of the rota-
constant values. Flow rates were adjusted for each series by chang- tional speed n. As expected, the torque decreased with increasing
ing the valving and pump configuration. Lower water levels were rotational speed n along each constant flow-rate curve.

© ASCE 04018052-4 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Table 3. Range of experimental parameters used in evaluation of Archimedes screw performance
Inclination Flow Inlet water Outlet water Head Speed Torque Number of
Shape (degrees) (L=s) level (cm) level (cm) (cm) (rpm) (N · m) data sets
Strake 20.6 1.0–5.8 3.6–14.6 0.0–9.9 24.6–50.0 10–200 0.13–1.41 446
24.2
29.1
Helicoid 20.6 1.7–5.7 3.9–14.4 0.1–8.7 29.1–47.2 24–192 0.19–1.44 198
23.9
29.1
Reverse strake 23.9 1.8–5.6 0.6–14.2 0.0–8.1 28.7–41.1 30–210 0.20–1.17 105

Fig. 8. Normalized inflow head hin =Ra versus rotational speed for Fig. 10. Mechanical power PM of screw versus rotational speed for
seven different flow rates. seven different flow rates.

Fig. 9. Torque T at shaft of screw versus rotational speed for seven


Fig. 11. Power efficiency η versus rotational speed for seven different
different flow rates.
flow rates.

Fig. 10 is the mechanical power PM as determined by Eq. (1) increases. Similar results were found for the helicoid and the
and the torque measurements in Fig. 9 reverse strake test screws.
Eq. (2) gives the theoretical hydraulic power Pth of a hydro-
60PM power site, where ρ is the density of water, g is the gravitational
T¼ ð1Þ
2πn constant, and H is the head difference (Fig. 3)

The mechanical power had a clear maximum for each particular Pth ¼ ρgQH ð2Þ
flow rate. This maximum mechanical power and the rotational where it is assumed that the inflow speed of the water is equal to its
speed at which it was attained both increased as the flow rate outflow speed.
Q increased. Letting PM denote the useful mechanical power at the shaft of
Fig. 10 shows that the mechanical power was approximately the screw, the power efficiency η is defined by
maximized for lower flow rates when the inflow height hin was
equal to Ra . However, for higher flow rates, the inflow height PM
η¼ ð3Þ
must be higher to attain the maximum mechanical power. This cor- Pth
responds well to the typical condition for a hydropower site, in
which the inflow height is constant for low to medium flow rates, The power efficiency curves in Fig. 11 were derived using
but at high flow rates (e.g., flooding conditions) the inflow head Eqs. (2) and (3), and the mechanical power data for PM in Fig. 10.

© ASCE 04018052-5 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


There was a clear maximum for each efficiency curve depending Initially, the leakage flow QL is only the gap flow QG, which
on the flow rate. The highest efficiencies were achieved for low increases approximately linearly with increasing inflow height hin .
flow rates. As the flow rate increased, the maximum efficiency de- At a certain inflow height indicated in the Fig. 12, spilling over the
creased and the rotational speed at which the maximum efficiency central tube begins, after which the leakage flow is the sum of the
was achieved increased. Very similar curves were found for the gap flow QG and the overspilling flow QO.
helicoid and the reverse strake test screws. During testing, it was also found that the leakage flow is depen-
dent on the stationary rotational position of the screw. The leakage
flow changes if the screw is turned by, for example 90°, in the flow
Hydraulic Power Losses and Measurement of direction as shown for the helicoid data in Fig. 12.
Leakage Flow The measured leakage flows were small compared with the
measured total flow, which ranged from 1.893 to 5.678 L=s in these
The mechanical power PM is the theoretical power Pth minus the tests. This corresponds to hydraulic losses QL =Q of 3.5–10.0%.
hydraulic power losses Phydr and the bearing power losses Pbr , so Theoretical models of the leakage flow are found in Muysken
that, for a given screw angle, the power efficiency can also be (1932), Nuernbergk and Rorres (2013), and Kozyn and Lubitz
written as (2015). The leakage flow QL affects the torque and performance
Phydr Pbr of Archimedes screws differently depending on whether the screws
η ¼1− − ð4Þ are conventional screw pumps (used to lift water) or hydropower
Pth Pth
screws, as tested in this experiment. More specifically, for hydro-
Muysken (1932) thoroughly analyzed the hydraulic power power screws the leakage flow, which is defined as the sum of the
losses Phydr , describing them as the sum of losses at the entrance gap and overspilling flows, contributes positively to the torque of
of the screw; leakage losses QL ; friction losses between the water the screw because it contributes to the volume contained within
and the trough, blades, and central tube; loss of kinetic energy of each bucket of the screw.
the water at the outlet of the screw; and losses arising from a non- However, for screw pumps, neither the gap flow nor the over-
optimal water level at the outlet of the screw. Muysken (1932) and spilling flow contributes to the volume being transported to the
Nuernbergk (2012) showed that the most significant losses among upper reservoir because their flow is opposite to the bucket flow
these are the leakage flow loss and the friction loss between the and is not replenished by water from the upper reservoir. This dif-
water and the blades. ference may be important when considering the performance differ-
The leakage flow loss QL is the sum of the gap flow QG between ences between various blade shapes and the importance of dynamic
the trough and the blades and the overspilling flow QO of the water effects of the water entering the screw.
spilling over the central tube. The leakage power loss PL is equal The flow in these scaled test screws was fully turbulent, as it is
to ρgQL H, so that from Eq. (2), the fraction of the power loss due for large-scale commercial screws. Consequently, similar gap-flow
to the leakage flow is behavior can be expected. In this experiment, the gap for the test
screws was less than 1 mm. According to Muysken (1932), as a
PL Q L screw is scaled up, its gap flow increases proportionally to the
¼ ð5Þ
Pth Q square of its outer radius, whereas its total flow increases propor-
tionally to the cube of its outer radius. Therefore, because the gap
flow becomes a smaller proportion of the total flow for larger
Determination of Leakage Flow screws, large-scale commercial screws can be expected to have
greater efficiency than the test screws.
Following testing, leakage flow tests were performed on each of
the three screws. Fig. 12 shows the measured leakage flows QL
of the three screws as a function of the inflow height hin (the height
of the water level of the upper water basin above the lower entrance Comparison of Flow Capacities of Three Test
edge of the trough to the screw as shown in Fig. 3). The test Blades
results in Fig. 12 were performed with the screw in a stationary One of the ways the three test screws were compared was by evalu-
(nonrotating) position. ating their relative flow capacities. Because the experimental data
were generated over a range of nominally constant flow rates, the
flow capacities could be compared by evaluating the inlet heights or
rotational speeds that were required for the three screws to pass the
same flow rate. For the same flow, the screws with the highest
speed or inlet height will have the highest corresponding hydraulic
losses, and therefore will operate less efficiently. There was very
little difference between the three test screws regarding the inlet
water levels and rotational speeds required to achieve the nominal
flow rate of 3.785 L=s (Fig. 13). The data presented subsequently
were normalized to account for minor variations in the flow rates
from run to run by multiplying the measured inlet water height by a
factor 3.785=Q, where Q is the actual flow rate of the run.
Similarly, over the lower range of flow rates (3.785 L=s and be-
low), little difference was found in the flow capacities (the flow rate
for a given inlet height and rotational speed) of the three test screws.
However, at higher flow rates and for the same speed, the strake
Fig. 12. Measurement of the leakage flow QL versus the inflow
required a markedly higher inlet water level to achieve the same
height hin .
flow rate as the other two test screws (Fig. 14). Alternatively,

© ASCE 04018052-6 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Fig. 13. Comparison of inlet water level required for normalized flow Fig. 15. Comparison of flow rate Q for three different screws at
rate of 3.785 L=s for three test screws. different rotational speeds (for hin ¼ Ra ).

Fig. 14. Comparison of inlet water level required for normalized flow Fig. 16. Comparison of available torque T at shaft of screw for the
rate of 5.68 L=s for three test screws. three different screws over different rotational speeds (for hin ¼ Ra ).

for the same inlet water levels, the strake required significantly interpolation of measured data for values of hin immediately above
higher rotational speeds to achieve the same flow rate (Fig. 15). and below Ra.
Keeping hin and H fixed throughout the measurements allowed
comparable measurements of the flows Q and torques T of the three
Comparison of Three Test Blades for Constant Head screws at different rotational speeds. From the measurements of
Difference and Inflow Height Q and T, the theoretical power Pth, mechanical power PM , and
the power efficiency η were determined for different rotational
To effectively compare the performance behavior of the three test speeds. In addition, these measurements were combined to deter-
screws, the measurement data were normalized by interpolation so mine the mechanical power as a function of the flow rate. These
that certain parameters were equal. For example, the number of calculations based on the experimental measurements are described
blades, inner radius, outer radius, and pitch were fixed throughout subsequently.
the experimental testing, and a single inclination angle of 24° was Fig. 15 shows the flow rate as a function of rotational speed. As
selected for data evaluation. In addition, for comparison purposes discussed in the previous section regarding flow capacity, the three
in this section, the data were evaluated and compared by setting the screws behaved similarly at lower rotational speeds. However, at
inflow water height hin and head difference H at fixed values. Spe- higher rotational speeds, the speed of the strake must be higher than
cifically, the inflow water height was fixed at the outer radius Ra of that of a helicoid or reverse strake to pass the same flow. For ex-
the screw (95.7 mm) and the head difference was kept fixed at ample, to maintain a flow rate of 5.047 L=s the strake must rotate
0.325 m (Fig. 3). The head difference H was kept fixed by keeping 25 rpm faster than the reverse strake, even though the strake has the
both the inflow water height hin and outflow water height hout fixed. highest bucket volume per turn (Table 2). This result was counter to
The outlet water height was kept fixed by the use of an overflow the expected behavior of the strake.
weir in the lower reservoir (Fig. 6). The inflow water height was Fig. 16 displays the torque as a function of rotational speed. This
kept mathematically fixed at Ra , as previously mentioned, by linear figure shows remarkable consistency among the three test screws

© ASCE 04018052-7 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Fig. 17. Comparison of available hydraulic power Pth (top curves) and Fig. 19. Mechanical power PN as function of flow rate Q (for
mechanical power PM (bottom curves) of screw at different rotational hin ¼ Ra ).
speeds (for hin ¼ Ra ).

It should be expected that the advantages and disadvantages of the


three shapes will be different when the screws are used for lifting
water, as for irrigation and land-draining purposes and in waste-
water treatment plants. For example, in the case of a power-
generating screw, the gap flow and overflow over the central tube
(the total leakage flow) are both in the direction of the bucket flow,
so they increase the volume in each bucket, and this volume in-
crease contributes to the torque; whereas for a water-lifting screw
these overflows are in the opposite direction to the upward bucket
flow, and so they do not contribute to the water being lifted to the
upper reservoir. The larger theoretical capacity of the strake screw
over a helicoid or reverse-strake screw of the same outer radius
could be an advantage when lifting water.
The differences in bucket volume per turn V B among the three
test screw models were relatively small. In addition, the blades of
the test screws were a greater proportion of the cross-sectional area
Fig. 18. Comparison of power efficiency η for three different screws at than would be expected with full-sized commercial screws. Further
different rotational speeds (for hin ¼ Ra ). work by the authors is expected to include optimized designs
which will provide better separation among the values of the bucket
volume per turn for each screw type.
The differences in the mechanical performance and efficiencies
regarding the decrease of torque as the rotational speed increased. of the three screws at the higher flow rates is postulated by the au-
This decrease was a result of the decreasing filling level as the thors to be associated with the effect of the blade shapes on the
rotational speed increased. dynamics of the water entering the screws. Further work is expected
Fig. 17 displays the theoretical available hydraulic power to include examinations of the effects of the aperture and shape of
Pth and the mechanical power PM. As expected from Eq. (8) the entry blades.
and Fig. 16, all three screws delivered approximately the same
mechanical power at the same rotational speed.
Fig. 18 shows the power efficiency η of the three screws. As
Conclusions
with the torque, the values were fairly close to each other at the
same rotational speeds. This paper presented experimental results for three blade shapes for
Fig. 19 displays the mechanical power as a function of the flow Archimedes hydrodynamic screws used in hydropower generation:
rate rather than the rotational speed. The mechanical powers of the helicoid, strake, and reverse strake. The measurement data for all
three screws were approximately the same for the lower flow rates three test screws were smooth and showed the expected theoretical
(below about 3.7 L=s). However, for the higher flow rates (above
behavior. For example, the torque varied as expected for larger rota-
about 4.5 L=s), the strake’s mechanical power was less than that of
tional speeds proportional to 1=n when the flow rate Q was kept
the helicoid.
constant. There was a clear maximum for each curve of efficiency η
versus rotational speed n. The maximum efficiencies were highest
Discussion and Planned Future Work at the lowest flow rate. With increasing flow rates, the maximum
efficiencies occurred at increasing rotational speeds and increasing
The testing of the three screw shapes in this report was in terms of inlet water levels. These observations were similar for all three test
their use as producers of hydropower through the lowering of water. screws.

© ASCE 04018052-8 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


The comparison of the inflow heights of the three test screws Ra = outer radius of screw (m);
showed that at flow rates of 3.785 L=s and below, the flow Ri = inner radius of screw (m);
capacities of the three shapes were almost identical, despite the dif- S= pitch of screw (m);
ferences in the theoretical bucket volume per turn V B associated T= torque at shaft of screw (N · m);
with the blade shapes. At higher flow rates, the strake design re-
VB = bucket volume per turn (L);
quired significantly higher inflow heights than did either the heli-
VU = volume per turn (L);
coid or the reverse strake, which is counter to the idea that higher
bucket volume per turn will lead to more flow capacity. Conse- β= inclination of screw (rad);
quently, for the strake design, higher rotational speeds were needed η= power efficiency;
than for the helicoid or the reverse strake to pass the required ρ= density of water (kg=m3 ); and
flow rate. φ= filling factor.
A constant inflow water level was then chosen for all three pro-
totypes to facilitate a fair comparison of the different units. Again,
at the lower flow rates the power, torque, flow capacities, and Supplemental Data
efficiencies were remarkably similar for the three shapes. At the
higher flow rates, differences in the mechanical performance and Appendixes S1 and S2, Figs. S1–S9, and Table S1 are available
efficiencies were due to the head differences required to accommo- online in the ASCE Library (www.ascelibrary.org).
date the same flow for a given speed and flow rate. Of the three
screws, the strake design required the highest head difference for
the same flow. References
For power-generating screws, the relative shape of the screw
blades appears to have minimal effect on the torque for the lower Brada, K. 1999. “Wasserkraftschnecke ermöglicht Stromerzeugung über
rotational speeds and flows. All three test screws behaved very sim- Kleinkraftwerke.” Maschinenmarkt 14: 52–56.
ilarly. However, at the higher speeds and flows, the test screws dis- Dellinger, G. 2015. “Etude expérimentale et optimisation des performances
hydrauliques des vis d’Archimède utilisées dans les micro centrales
played differences in the inflow height, volume per turn, and filling
hydroélectriques.” Ph.D. thesis, Université de Strasbourg.
factor actually achieved by the screw. Based on the results of this
Dellinger, G., A. Terfous, P-A. Garambois, and A. Ghenaim. 2016.
experiment, these differences are believed to be due to the different “Experimental investigation and performance analysis of Archimedes
dynamic entrance effects of the three screws, and were in opposite screw generator.” J. Hydraul. Res. 54 (2): 197–209. https://doi.org/10
relation to the theoretical V ma values of the three test screws. .1080/00221686.2015.1136706.
Erinofiardi, E., A. Nuramal, P. Bismantolo, A. Date, A. Akbarzadeh,
A. K. Mainil, and A. F. Suryono. 2017. “Experimental study of screw
Acknowledgments turbine performance based on different angle of inclination.” Energy
Procedia 110 (Mar): 8–13. https://doi.org/10.1016/j.egypro.2017.03
This material is based upon work supported by the US Department .094.
of Energy, Office of Energy Efficiency and Renewable Energy Kibel, P., and T. Coe. 2011. Archimedean screw risk assessment: Strike and
(EERE), under Award No. DE-EE0007247. The authors gratefully delay probabilities. Rep. RA 1108. Totnes, UK: Fishtek Consulting.
acknowledge the support of Pacific Northwest National Laboratory, Kozyn, A., and W. D. Lubitz. 2015. “Experimental validation of gap
leakage flow models in Archimedes screw generators.” Vol. 1 of
which was provided through a Cooperative Research and Develop-
Renewable energy in the service of mankind, edited by A. Sayigh,
ment Agreement and Technology Assistance Program Agreements 365–375. Cham, Switzerland: Springer.
between Percheron Power, LLC and Battelle Memorial Institute, Kozyn, A., and W. D. Lubitz. 2017. “A power loss model for Archimedes
Pacific Northwest Division under US Department of Energy screw generators.” Renewable Energy 108 (Aug): 260–273. https://doi
Contract No. DE-AC05-76RL01830. .org/10.1016/j.renene.2017.02.062.
Lashofer, A., W. Hawle, F. Kaltenberger, and B. Pelikan. 2013. “Die
Wasserkraftschnecke–Praxis, Prüfstand und Potenzial.” Österreichische
Notation Wasser- und Abfallwirtschaft 65 (9–10): 339–347. https://doi.org/10
.1007/s00506-013-0054-x.
The following symbols are used in this paper: Lashofer, A., F. Kaltenberger, and B. Pelikan. 2011 “Wie gut bewährt sich
G = gap between screw and trough (m); die Wasserkraftschnecke in der Praxis.” Wasserwirtschaft, 76–82.
Wiesbaden: Springer.
g = gravitational constant (m=s2 );
Lee, K-T., E-S. Kim, W-S. Chu, and S-H. Ahn. 2015. “Design and 3D
H = head difference (m); printing of controllable-pitch Archimedean screw for pico-hydropower
hin = inflow (upper) water height (m); generation.” J. Mech. Sci. Technol. 29 (11): 4851–4857. https://doi.org
hout = outflow (lower) water height (m); /10.1007/s12206-015-1032-y.
N = number of blades; Lisicki, M., W. Lubitz, and G. W. Taylor. 2016. “Optimal design and
n = rotational speed (m−1 ); operation of Archimedes screw turbines using Bayesian optimization.”
Appl. Energy 183 (Dec): 1404–1417. https://doi.org/10.1016/j.apenergy
Pbr = bearing power losses (kW); .2016.09.084.
Phydr = hydraulic power losses (kW); Lubitz, W., M. Lyons, and S. Simmons. 2014. “Performance model of
PL = leakage power (kW); Archimedes screw hydro turbines with variable fill level.” J. Hydraul.
PM = mechanical power (kW); Eng. 140 (10): 04014050. https://doi.org/10.1061/(ASCE)HY.1943
Pth = theoretical hydraulic power (kW); -7900.0000922.
Lyons, M., and W. D. Lubitz. 2013. “Archimedes screws for microhydro-
Q = total water flow (L=s);
power generation.” In Proc., ASME 7th Energy Sustainability Conf.
QG = gap leakage flow (L=s); New York, NY: ASME.
QL = total leakage flow (L=s); Muysken, J. 1932. “Berekening van het nuttig effect van de vijzel.”
QO = overflow leakage flow (L=s); De Ingenieur 21 (Mei): 77–91.

© ASCE 04018052-9 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


Nuernbergk, D. M. 2012. Wasserkraftschnecken: Berechnung und Renewable Energy 94 (Aug): 136–146. https://doi.org/10.1016/j.renene
optimaler Entwurf von archimedischen Schnecken als Wasserkraftma- .2016.03.044.
schine. Detmold, Germany: Moritz Schaefer. Rorres, C. 2000. “The turn of the screw: Optimal design of an Archimedes
Nuernbergk, D. M., and A. Lashofer. 2016. “Skalierung von screw.” J. Hydraul. Eng. 126 (1): 72–80. https://doi.org/10.1061/(ASCE)
Wasserkraftschnecken—Ein Beitrag zur einer überfälligen Dimensio- 0733-9429(2000)126:1(72).
nierungsrechnung.” WasserWirtschaft 10 (Oct): 25–31. Saroinsong, T., R. Soenoko, S. Wahyudi, and M. N. Sasongko. 2015.
Nuernbergk, D. M., and C. Rorres. 2013. “Analytical model for water “The effect of head inflow and turbine axis angle towards the three
inflow of an Archimedes screw used in hydropower generation.” row bladed screw turbine efficiency.” Int. J. Appl. Eng. Res. 10 (7):
J. Hydraul. Eng. 139 (2): 213–220. https://doi.org/10.1061/(ASCE)HY 16977–16984.
.1943-7900.0000661. Vitruvius. 1914. “The water screw.” Chap. VI in De Architectura.
Rohmer, J., D. Knittel, G. Sturtzer, D. Flieller, and J. Renaud. 2016. Translated by Morris Hicky Morgan. New York: Dover, 1960.
“Modeling and experimental results of an Archimedes screw turbine.” 295–297.

© ASCE 04018052-10 J. Hydraul. Eng.

J. Hydraul. Eng., 2018, 144(8): 04018052


SUPPLEMENTAL DATA

ASCE Journal of Hydraulic Engineering

Experimental Evaluation of Advanced


Archimedes Hydrodynamic Screw Geometries
Jerry L. Straalsund, Samuel F. Harding, Dirk M. Nuernbergk, and
Chris Rorres

DOI: 10.1061/(ASCE)HY.1943-7900.0001479

© ASCE 2018

www.ascelibrary.org
View publication stats

You might also like