You are on page 1of 16

EXPERIMENTAL STUDY OF CROSS-FLOW TURBINE

By Shahram Khosrowpanah,1 A. A. 3Fluzat,2 Member, ASCE, and


Maurice L. Albertson, Fellow, ASCE

ABSTRACT: The performance of the cross-flow turbine, CFT, also known


as the Banki turbine, was studied experimentally by varying the number
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

of blades, the runner diameter, and the nozzle entry arc under flow/head
variations. The results, as presented in this paper, show that the
maximum efficiency of the CFT at anyflow/headcombination increases
as the nozzle entry arc increases or the aspect ratio of the runner
decreases. Equations have been suggested to obtain the optimum
number of blades for maximum efficiency of the turbine and tofindthe
specific speed of the CFT as a function of unit discharge and nozzle
entry arc.
INTRODUCTION
At the present time, hydroelectric energy provides only about 23% of the
total electric generation, less than half its potential, on a worldwide basis
(Armstrong 1978). Many third world countries have a great need for
electric power especially those without a grid system. The cross-flow
turbine (CFT) has a good potential for hydroelectric power development of
both advanced and developing countries (Inversin 1981). Various inter-
national organizations have been able to adapt this turbine to the rural
areas in third world countries (Inversin 1981). In this study certain gaps
that existed in the knowledge about the CFT have been filled by conduc-
ting a laboratory study of the turbine.

PRINCIPLE OF OPERATION OF THE CFT


The CFT, shown in Fig. 1, operates by a rectangular jet flowing through
a ring of blades on a barrel-shaped rotor. The jet imparts energy to the
rotor first as it enters the rotor and again as it leaves the rotor on the
opposite side. This turbine, which is a radial-impulse type, has been
classified as a small-scale or ultra-low head turbine based on the U.S.
Army Corps of Engineers' definition (1979). However, Ossberger Tur-
binenfabrik (1981) has manufactured CFTs for a head of 200 m (659 ft) and
flow of 11.32 m3/s (400 cu ft/sec). The range of specific speed of the CFT
varies from 9 to 46 in English units (Warnick 1984). The reader is referred
to Khosrowpanah (1984) for other characteristics and advantages of the
CFT.

REVIEW OF LITERATURE
The Australian engineer A. G. M. Michell obtained a patent for this
turbine in 1903. The turbine was further developed by Donat Banki in West
•Asst. Prof., Univ. of Guam, Water & Energy Res. Inst., Mangilao, Guam 96923.
2
3
Asst. Prof., Dept. of Civ. Engrg., Clemson Univ., Clemson, SC 29634.
Prof., Dept. of Civ. Engrg., Colorado State Univ., Fort Collins, CO 80523.
Note. Discussion open until August 1, 1988. To extend the closing date one
month, a written request must be filed with the ASCE Manager of Journals. The
manuscript for this paper was submitted for review and possible publication on
August 18, 1986. This paper is part of the Journal of Hydraulic Engineering, Vol. 114,
No. 3, March, 1988. ©ASCE, ISSN 0733-9420/88/0003-0299/$1.00 + $.15 per page.
Paper No. 22272.
299

J. Hydraul. Eng. 1988.114:299-314.


Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

FIG. 1. Components of Cross-Flow Turbine (Ossberger 1981): (1) Nozzle; (2)


Runner; (3) Blade; (4) Shaft; (5) Bearing; (6) Casing; (7) Draft Tube; (8) Guide Vane;
(9) Air Valve

Germany, and became known as the Banki or Michell-Banki Turbine


(Sonnek 1923). Banki developed blade design methods, an expression for
maximum efficiency, the ratio of the inner to outer diameter of the runner,
and energy transfer at each stage. Sonnek (1923) modified the Banki theory
and recommended that the nozzle width be narrower than the runner width
to reduce the separation on the back side of the blades. Mockmore and
Merryfield (1949) tested a model of CFT and concluded that 1) The CFT
can operate efficiently under a wider range of flow variations than most
turbines; and 2) the maximum efficiency occurs at the same speed for all
openings of the guide vane at a constant head. Shepherd (1956) showed
that theoretically, 72% of the shaft energy is transferred to the rotor during
the first stage, because the fluid and blade angles can be calculated and
controlled. Varga (1959) investigated the variation of static pressure along
the nozzle of the CFT. He verified Banki's relationship between the
peripheral velocity and runner velocity resulting in maximum efficiency. In
his study of the CFT, Haimerl (1960) stated that the CFT is not a perfect
impulse turbine because at the nozzle exit the pressure is slightly higher
than atmospheric. Johnson et al. (1983) studied effectiveness of new
materials in the construction of the CFT. Nakase et al. (1982) studied the
effect of nozzle shape and size on performance of the CFT. They defined
a variable, called nozzle throat-width ratio TR , and suggested that among
the values of this variable studied, the most suitable nozzle is one with

So
TR = -~- = 0.26 (1)
/lA

in which S0 = nozzle throat width; ry = outer radius of the runner; and X


= nozzle entry arc (see Fig. 2). The upper limit of the value of TR in their
study was 0.44. Comparison of their work with the results of the present
study is given later.
300

J. Hydraul. Eng. 1988.114:299-314.


8.2* |

1. 14.5
7,2 D
I
10.1
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

I I'8-?'

47.0

1 t

luT

FIG. 2. Model Dimensions for 30.48-cm (12-in.) Runner Diameter with Nozzle Entry
Arc of 90°

The most intensive development work on the CFT has perhaps been
done by Ossberger (1981), by building 7,500 units in the past 60 years. In
1984 Van Dixhorn conducted an experimental study of the blade forces in
the CFT. The blade shapes and curvatures of his study are not used in this
work since simplicity has been considered the primary factor in designing
the turbine.

THEORETICAL BACKGROUND

The cross-flow turbine has experienced relatively little investigation to


this date. Balje (1981) defines this turbine as one that intakes water through
half of the rotor and exhausts the water through the other half. His
relatively exhaustive analysis of the turbine is based entirely on this
assumption. He reports maximum efficiency of the turbine to occur at a
speed ratio of 0.4-0.5 for the entering stage and at higher speed ratios for
the exiting stage. Speed ratio is the ratio of inflow jet speed to the
peripheral velocity of the runner. His treatment of the turbine confirms the
blade arrangement used in this study. He reports that 70% of the power is
generated in the first stage and 30% in the second stage of the turbine.
However, all of Balje's analysis is based on his original assumption that
flow enters through half of the rotor and exists through the other half. It
will be explained later that this was not the case in the present study. The
maximum theoretical efficiency he reports for the CFT is 73% without a
draft tube and 82% with a draft tube.
The efficiency of the CFT, similar to other turbines, is defined as
301

J. Hydraul. Eng. 1988.114:299-314.


Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

FIG. 3. Velocity Diagram of Cross-Flow Turbine

Power
= OUtput = Pout
Power input P in
in which Pln = yQH; Pout = Tu>; 7 = specific weight of water; Q =
discharge; H = total head at the nozzle tip; T = shaft torque; and co =
angular velocity of the runner. Factors affecting efficiency of the turbine
are given below.
Speed Ratio
The velocity diagram of the CFT in the radial-tangential plane is shown
in Fig. 3. The condition for occurrence of maximum efficiency for the CFT
was analyzed by Sonnek (1923) who showed that based on certain
assumptions, efficiency is maximized when

Vi 2 cos a!
in which a1 = angle of attack; U1 = peripheral velocity of the runner; and
Vx = absolute velocity of the flow at the first stage. The ratio U1/V1 is
called the speed ratio, which can be expressed in terms of total head and
rotational speed of the runner as
U, _ TTD^N
(4)
vrm^jim •
302

J. Hydraul. Eng. 1988.114:299-314.


in which Dx = Outer diameter of the runner (Fig. 3); N = runner rpm; H
= total head at the nozzle tip; and the number 60 is conversion factor from
minutes to seconds. The effect of the speed ratio on efficiency is discussed
later.

Number of Blades
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

One of the important considerations of the runner design for the CFT is
selecting the optimum number of blades. Too many blades will increase
losses, the weight of the runner and the cost of the turbine. Too few blades,
on the other hand, cause higher losses by flow separation on the back side
of the blades. The turbine parameter reflecting the role of the number of
blades is the solidity ratio a, defined as:
a
<x = y • ••• (5)

in which a = the radial distance from outer to inner edge of the blades; and
t = distance along the runner periphery between two consecutive blades.
Using Fig. 3, and neglecting blade thickness
TTDI
t= (6)
n
and
Di-D2_Dl( D2\

The solidity ratio, which is a geometric variable, depends on the number of


blades and also the diameter of the runner. The diameter of the runner and
also its width are usually designed based on the available head and
discharge. Thus, the solidity ratio (or the number of blades) will also
depend on head and discharge. Khosrowpanah (1984) derived the equation
Z>! cos P] v w .
= (8)
T 6Q'
where the left-hand side is called the aspect ratio, for which Balje (1981)
suggests a value of 1.0. Substituting into Eq. 5 from Eqs. 6, 7, and 8:

a (9)
~ 2^ ~B y Di) cos pjV^n x " "
in which B = the runner width along the shaft; and Q' = the unit discharge,
which is a dimensionless representation of discharge, as will be defined in
the next section. The angle pj is defined in Fig. 3. In Eq. 9, the ratio D2/D1
varies within a narrow range of 0.66-0.69 according to the theoretical work
done on the CFT (Sonnek 1923). The ratio DJB, called the aspect ratio, is
determined by the designer. Ossberger (1981) has built turbines with aspect
ratios of up to 3.5. Balje (1981) assumed an aspect ratio of 1.0 for his
theoretical calculations. In this study the aspect ratio of 2.0 gave better
results than 1.0, as will be discussed later. The ultimate choice will
303

J. Hydraul. Eng. 1988.114:299-314.


probably be in the range 1.0-2.0. Choosing the aspect ratio, Eq. 9 will give
the optimum number of blades if an optimum value for the solidity ratio a
can be found.
TURBINE CONSTANTS
The variables which can be measured and which affect the turbine
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

performance can be expressed in a functional equation as


<KG, H, Pmu N,DuB) =0 (10)
in which all the variables are previously defined. This list is not exclusive,
but adequate for the purpose of this paper. The variables of Eq. 10 are
customarily combined into turbine constants for the study of homologous
turbine units (Balje 1981). The turbine constants needed to explain and/or
discuss the results of this study are listed below (Khosrowpanah 1984;
Warnick 1984):

Q
Q' = DiHm = u n i t discharge (11)

-Pout
P' = ^2^3/2 = u n i t Power (12)

NDX
N' = -jrvT - unit rotational speed (13)

= NP™t
Ns irS/4 - specific speed (14)

These turbine constants are not all dimensionless, but they are the
conventional parameters used to describe homologous turbines (Balje
1981). Other parameters such as DXIB and efficiency iq are important design
parameters and will be utilized in the presentation of the results.
TESTED MODEL
A model of the CFT was designed, constructed and, tested at Colorado
State University. In the design three factors were considered to be most
important:
1. Simplicity of design and construction (as a result, use of expensive
shapes of blades and nozzle was avoided).
2. Low cost.
3. Obtaining maximum efficiency.
This model was designed for a total head of 1.79 m (5.9 ft) and a discharge
of 0.03 m3/s (1.19 cfs) with the assumed efficiency of 65%. The ratio of the
inner to outer diameter of the runner, D2/D1, was selected to be 0.68 which
is the value used by Nakase et al. (1982), and Varga (1959). Other nozzle
characteristics are given in Table 1. The dimensions of the model were as
follows.
Runner number 1 had a diameter of 15.24 cm (6 in.), and runners 2, 3,
and 4 had a diameter of 30.48 cm (12 in.). The blade inlet and outlet angles
304

J. Hydraul. Eng. 1988.114:299-314.


TABU 1. Characteristics of Tested Runners
Runner Nozzle

Runner Outer Blade Number Entry arc


number diameter (in.) length (in.) of blades (degrees) Number Discharge (cfs)
(D (2) (3) (4) J (5) (6) (7)
1 6 6 20 58 A6 0.72-0.89
1 6 6 20 78 B6 0.72-0.89-1.04
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

1 6 6 20 90 C6 0.72-0.89-1.04
2 12 6 20 58 A12 1.04-1.19-1.44
2 12 6 20 78 B12 1.04-1.19-1.44
2 12 6 20 90 C12 1.04-1.19-1.44
3 12 6 15 58 A12 1.19-1.44
3 12 6 15 78 B12 1.19-1.44-1.56
3 12 6 15 90 C12 1.19-1.44-1.56
4 12 6 10 58 A12 1.44-1.56
4 12 6 10 78 B12 1.44-1.56
4 12 6 10 90 C12 1.44-1.56
Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m 3 /s.

(Pi and (32 in Fig. 3) were 30° and 90° with the length B (along the shaft) of
15.24 cm (6 in.). Based on previous constructions of the turbine (Ossberger
1981), 20 blades were chosen for runner number 1 and 20, 15, and 10 blades
for runners 2, 3, and 4, respectively. The casing was open to the air at the
top. No draft tube was used.
Nozzles for all the runners had the same length as the blades (B) but
varying entry arcs of 58°, 78°, and 90°. Larger nozzle entry arcs are not
recommended without the use of a guide vane (Ossberger 1981) and hence
were not used. This point disputes Balje's assumption of flow entering
through half of the rotor. The vertically admitting nozzle was arranged as
shown in Fig. 2. The nozzle throat width ratio SJr^ for these three nozzles
was 0.41 which is within the range tested by Nakase et al. (1982). In this
test apparatus, there were no guide vanes upstream of the nozzle.

EXPERIMENTAL EQUIPMENT AND TEST PROCEDURE

The experimental setup is shown in Figs. 2 and 4. The flow rate supplied
to the nozzle was measured with a calibrated Venturi meter. The velocity,
pressure, and total heads were measured at the tip of the nozzle by two
pitot tubes traversing across the nozzle in two orthogonal directions. The
datum for all head measurements was a line passing through the center of
turbine shaft. The rotational speed of the runner was measured by a
directly attached electronic tachometer. The output torque of the runner
was measured with a Prony brake.
The model was tested using runners 1, 2, 3, and 4 with a series of nozzle
openings of 58°, 78°, and 90° under different discharges as shown in Table 1.
The testing procedure for the experiment was as follows. The pump was
started and the desired discharge accurately adjusted by a gate valve.
Velocity, pressure, and total heads were measured across and perpendicu-
lar to the nozzle tip normal to the flow. As an additional check, the wall
pressure at the nozzle tip was measured by using a mercury manometer.
The rotational speed of the runner was first recorded before applying any
load. The shaft was then loaded by means of the Prony brake. The force on

305

J. Hydraul. Eng. 1988.114:299-314.


Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

FIG. 4. Tested Model at Colorado State University (CSU)

the Prony brake arm and the corresponding rotational speed of the runner
were recorded at each load. The pressure on the wall and discharge
through the model were checked several times during the test in order to
ensure steady operation of the system.

RESULTS
Performance characteristics of the CFT under tested conditions are
presented in terms of turbine constants (Eqs. 11-14) for runners 1,2,3, and
4, as summarized in Tables 2 through 6. The results are presented as
follows:
Effect of Nozzle Entry Arc
The variation of the efficiency of the turbine with unit rotational speed of
the runner (Eq. 13) at a discharge of 0.04 m3/sec (1.44 cfs) for runner 2, is
shown in Fig. 5. By increasing the nozzle entry arc from 58° to 90° the
maximum efficiency increased. The highest efficiency in this experiment
was obtained for a nozzle entry arc of 90°. At the smaller nozzle opening
(58°) the static pressure at the nozzle tip is higher than it is at larger
openings (see Tables 2-6). This higher pressure at the nozzle reduces
efficiency because 1) The angle of attack, a, probably deviates from its
design value of 16°, as also observed by Nakase et al. (1982); and 2) a
bigger portion of the flow travels on the outside by the blades. In Fig. 6 a
comparison is made between the variation of efficiency and unit speed of
the runner with nozzle entry arc of 90° from runner number 2 and the
experimental results of Nakase et al. (1982). This curve was not actually
plotted by Nakase et al., but inferred from their data. The results obtained
at Colorado State University (CSU) are about 3% more efficient, possibly
because of slightly smaller throat width ratio used at CSU. Effect of the
nozzle entry arc on efficiency has not been theoretically treated by
previous researchers.

306

J. Hydraul. Eng. 1988.114:299-314.


TABLE 2. Performance Characteristics of Runner Number 1, with Varying Dis-
charge (Note: runner diameter = 6 in.; runner width = 6 in.; nozzle width = 6 in.;
and number of blades = 20)
Unit
Power power Unit Specific
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

Dis- Nozzle Total Power output RPM Run- Unit output RPM Speed speed
charge entry arc head input (max) at away discharge (AT') at ratio at
^.vmax
in. Nsit
(ofs) (degrees) (ft) (hp) (hp) (%) ^Umax (RPM) (60 (max) At/max A(/max AT/max
(D (2) (3) (4) (5) (6) (7) (8) (9). (10) (n) (12) (13)
0.72 58 3.08 0.2516 0.1559 62 290 681 1.640 0.1153 83 0.53 28.06
0.72 78 1.84 0.1503 0.1097 73 225 590 2.1231 0.1758 85 0.54 34.77
0.72 90 1.30 0.1061 0.0843 79.5 183 520 2.5251 0.2274 80 0.53 38.00
0.89 58 4.77 0.4816 0.2600 54 356 770 1.6300 0.0998 82 0.53 25.75
0.89 78 2.98 0.3009 0.1860 62 290 675 2.0622 0.1446 84 0.54 31.94
0.89 90 2.25 0.2272 0.1522 67 240 620 2.3733 0.1803 81 0.53 34.00
a a a a a a a a a a
1.04 58 "
1.04 78 4.32 0.5097 0.2752 54 350 760 2.001 0.1225 84 0.54 29.48
1.04 90 3.25 0.3834 0.2300 60 300 714 2.3071 0.1570 83 0.54 32.97
a
High rotational speed.
Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m 3 /s.

TABLE 3. Performance Characteristics of Runner Number 2, with Varying Dis-


charges (Note: runner diameter = 12 in.; runner width = 6 in.; nozzle width = 6
in.; and number of blades = 20)
Unit
Power power Unit Specific
Dis- Nozzle Total Power output Run- Unit output RPM Speed speed
RPM (N') at ratio at W s at
charge entry arc head input (max) ^-i/max
(%) at£#m,x
away discharge
in
(cfs) (degrees) (ft) (hp) (hp) (RPM) (2') (max)
Aw max At/max Humax
(D (2) (3) (4) (5) (6) (7) (8) (9)
(10)
(11) (12) (13)
1.04 58 4.00 0.4720 0.3304 70 165 398 0.5200 0.0413 83 0.53 16.76
1.04 78 1.93 0.2277 0.1684 74 115 328 0.7486 0.0628 83 0.54 20.74
1.04 90 1.10 0.1297 0.1037 80 • 87 262 Q.9916 0.0898 83 0.54 24.86
1.19 58 5.38 0.7264 0.4576 63 190 445 0.5130 0.0366 82 0.53 15.68
1.19 78 2.80 0.3780 0.2494 66 138 365 0.7111 0.0532 83 0.54 19.02
1.19 90 1.47 0.1985 0.1449 73 100 277 0.9814 0.0813 83 0.54 23.51
1.44 58 8.54 1.3952 0.7673 55 240 539 0.4927 0.0307 82 0.53 14.40
1.44 78 4.20 0.6861 0.4116 60 170 432 0.7026 0.0478 83 0.54 18.14
1.44 90 2.44 0.3986 0.2630 66 130 349 0.9218 0.0690 83 0.54 21.86

Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m 3 /s; 1 ft = 30.48 cm.

Table 3 shows that the rotational speed of the runner at the maximum
efficiency decreases as the nozzle entry arc increases from 58° to 90° at any
discharge. Since the blade geometry is fixed in the CFT, the smaller entry
arc causes a less efficient runner operation and a higher head at the nozzle
tip. As expected, the drop in the rotational speed of the runner is
proportional to the square root of the drop in total head at the nozzle tip.
This result is confirmed by Eq. 4 for a constant speed ratio Ul/V1, and by
Eq. 13 for a constant unit rpm.

Effect of Speed Ratio


Variation of the efficiency with the speed ratio UjVi is shown in Fig. 7.
307

J. Hydraul. Eng. 1988.114:299-314.


TABLE 4. Performance Characteristics of Runner Number 3, with Varying Dis-
charge (Note: runner diameter = 12 in.; runner width = 6 in.; nozzle width = 6
in.; and number of blades = 15)
Unit
Power power Unit Specific
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

Dis- Nozzle Total Power output RPM Unit output RPM Speed speed
(iV') at ratio at N,<£
charge entry arc head input
(cfs) (degrees) (ft) (hp)
(max) at
(hp) £ # „ , « (%> Effm„
Runaway discharge
(RPM) <6')
(n At/max Effm&x Atfmax
(4) (max) (13)
(1) (2) (3) (5) (6) (7) (8) O) (11) (12)
(10)
1.19 58 4.36 0.5886 0.4002 68 175 414 0.5699 0.0439 84 0.54 17.57
1.19 78 2.00 0.2700 0.1998 74 120 312 0.8414 0.0706 85 0.55 22.55
1.19 90 1.12 0.1512 0.1194 79 90 263 1.1244 0.1007 85 0.55 26.99
1.44 58 7.20 1.1762 0.7175 61 220 470 0.5366 0.0371 •82 0.53 15.66
1.44 78 3.19 0.5211 0.3439 66 149 368 0.8062 0.0603 83 0.54 20.49
1.44 90 1.98 0.3251 0.2275 70 120 303 1.0233 0.0816 85 0.55 24.36
a a a a a a a
1.56 58
1.56 78 4.09 0.7240 0.4416 61 170 385 0.7713 0.0533 84 0.54 19.42
1.56 90 2.40 0.4247 0.280 66 130 325 1.01 .0753 84 0.54 23.03
a
Blade breaks.
Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m3/s; 1 ft = 30.48 cm.

TABLE 5. Performance Characteristics of Runner Number 4, with Varying Dis-


charges (Note: runner diameter = 12 in.; runner width = 6 in.; nozzle width = 6
in.; and number of blades = 10)
Unit
Power power Unit Specific
Nozzle Total Power output RPM Run- output RPM Speed speed
Discharge entry arc head Input (max) At/max at away Unit (/•') (N') at ratio at W s at
(cfs) (degrees) (RPM) discharge (max)
(ft) (hp) (hp) (%)
Effm* £5max ^Tmax Aw max
d) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13)
1.44 58 6.21 1.0145 0.5376 53 210 450 0.5778 0.0347 84 0.55 15.70
1.44 78 2.48 0.4052 0.2471 61 130 332 0.9144 0.0632 83 0.54 20.76
1.44 90 1.43 0.2336 0.1471 63 104 262 1.2041 0.0860 86 0.55 25.50
1.56 58 7.59 1.3433 0.6850 51 232 487 0.5662 0.0327 84 0.55 15.24
1.56 78 3.17 0.5611 0.3142 56 150 365 0.8761 0.0556 84 0.55 19.87
1.56 90 1.78 0.3150 0.1890 60 114 285 1.1692 0.0795 85 0.55 24.10
Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m3/s; 1 ft = 30.48 cm.

Using an angle of attack a, of 16° as used in this study, Eq. 3 indicates that
the speed ratio of 0.52 will result in maximum efficiency. The line for UJVi
= 0.52 has been shown in Fig. 7. However, Table 3 shows maximum
efficiency to occur at slightly higher values of the speed ratio (0.53-0.54).
Noting that Eq. 3 is not derived as an exact equation, this agreement is
quite acceptable, (Balje, 1981).

Effect of Runner Size


By reducing the diameter of the runner from 30.48-15.24 cm (12-6 in.)
the maximum efficiency of the turbine decreased as observed by compar-
ing Tables 2 and 3. Since the speed ratio at the maximum efficiency for
these two runners is the same, the losses through the turbine should be
higher for runner number 1. Visual observation of flow path through runner
number 1, not presented here due to lack of space, verified this fact by
308

J. Hydraul. Eng. 1988.114:299-314.


TABLE 6. Performance Characteristics of Runner Numbers 2, 3, and 4, under
Varying Number of Blades (Note: runner diameter = 12 in.; runner width = 6 In.;
nozzle width = 6 in.; and discharge = 1.44 cps)
Unit
Power power Unit Specific
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

Number Nozzle Total Power output RPM Run- Unit output RPM Speed speed
of entry arc head input (max) QtFmax. at away discharge (/>') (AO at ratio at W* at
blades (degrees) CM (RPM) <2') (max)
(ft) (hp) <%)Mffmax At/max ^i/max At/max
(D (2) (3) (4) (S) (6) (7) (8) (9) (10) (11) (12) (13)
58 8.54 1.3952 0.7673 55 240 539 0.4927 0.0307 82 0.53 14.40
20 78 4.20 0.6861 0.4116 60 170 432 0.7026 0.0478 83 0.54 18.14
90 2.44 0.3986 0.2630 66 130 349 0.9218 0.0690 83 0.54 21.86
58 7.20 1.1762 0.7175 61 220 470 0.5366 0.0371 82 0.53 15.66
15 78 3.19 0.5211 0.3439 66 149 368 0.8062 0.0603 82 0.54 20.49
90 1.98 0.3251 0.2275 70 120 303 1.0233 0.0816 83 0.55 24.36
58 6.21 1.0145 0.5376 53 210 450 0.5778 0.0347 84 0.54 15.70
10 78 2.48 0.4052 0.2471 61 130 332 0.9144 0.0632 83 0.54 20.70
90 1.43 0.2336 0.1471 63 101 262 1.2041 0.0860 84 0.55 24.77
Note: 1 in. = 2.54 cm; 1 cfs = 0.028 m 3 /s; 1 ft = 30.48 cm.

90.0
NOZZLE ENTRY AHC 58° D
60.0 NOZZLE ENTRY ARC 7B° O
70,0
NOZZLE ENTRY ARC 90° A

60.0

50.0-

•10.0-

30.0

2Q.0

10.0

n n
. 1 1 1 1 1 1 i 1 1 1
0.0 30.0 50.0 90.0 120.0 ISO.O IBO.O 210.0 240.0 270.0 300.0
UNIT R.P.M.

FIG. 5. Performance Curves for Runner Number 2, for Various Nozzle Entry Arcs
with Discharge of 0.04 m3/s (1.44 cfs)

indicating a high tendency of water to rotate around the shaft in the smaller
nozzle (Khosrowpanah 1984).

Effect of Number of Blades


Experimental results for runners number 2, 3, and 4, with 20, 15, and 10
blades respectively, at a discharge of 0.04 m3/s (1.44 cfs), are summarized
in Table 6. Fig. 8 shows the maximum efficiency for these runners with
nozzle entry arcs of 58°, 78°, and 90°, indicating that the highest efficiency
occurs at about 15 or 16 blades. Flow pattern observations during testing
did confirm the smallest separation pocket for the 15-blade runner.
According to Eq. 9, for a selected value of D2IDl = 0.68 and DXIB = 2.0
(runners 2, 3, and 4) and A. = 90°
309

J. Hydraul. Eng. 1988.114:299-314.


Source Symbol X r,X
CSU A 90° 0.41
BO.O-
Nakase V 90° 0.44
BO.O-

TO,0-
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

00.0- /^
SCO-

«0.0

30.0

,0,,

,0.0

0.) 10.0 1 60-0 BO.O 100.0 UO.O MOO IS 0.0


40-0
Unit R.P.H.

FIG. 6. Comparison of Results from Nakase et al. (1982) and from CSU for Runner
Number 2

100.0

90.0
NOZZLE ENTRY ARC 58° d
NOZZLE ENTRY ARC 78° O
80.0 NOZZLE ENTRY ARC 90° A
70.0

^ 60.0
Lj
H 50.0

I.J
10.0

30.0

20.0

10.0

0.0
0.00.1 0.2 0.3 0.4 0.5 0.6 0.7 O.S 0.9 1.0 I.I 1.2 J . l 1.1 I.S I.G 1.7 1.0 1.9 2.0
SPEXO NUT 10 U I / V I

FIG. 7. Variation of Efficiency with Speed Ratio 1/,/V1 for Runner Number 2, for
Various Nozzle Entry Arcs with Discharge of 0.04 m3/s (1.44 cfs)

cr oc Q' (15)
•n

This proportionality is plotted in Fig. 9, which indicates a value of cr = 1.03


for maximum efficiency. Since the terms of Eq. 15 are dimensionless, this
value of the solidity ratio can be used as an index for runner design, for the
case of D2/D1 = 0.68 and A. = 90°. The use of the optimum value of the
solidity ratio in Eq. 9 will also confirm the result of Fig. 8.

310

J. Hydraul. Eng. 1988.114:299-314.


eo.a
NOZZLE F.NTRY ARC 58° a
NOZZLE EHTRY ARC 78° o
M.e NOZZLt W H Y ARC 90° A

to a

^ ^
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

«s.o-

W.O-

0.0
la t.e (.0 s.n IO.S •).«

FIG. 8. Variation of Maximum Efficiency with Number of Blades for Various Nozzle
Entry Arcs with Discharge of 0.04 m3/s (1.44 cfs)

NOZZLE EHTRY ARC 90° a


BO.O
D2/Dx = 0 . 6 8

eo.o

70.0

J ^ ~ ^ <
eo.o
*
5O.0

10.0-

».o-

ZO.O

10.0

0.0
0.0 O.t 0.2 0.1 0.1 0.5 0.8 0.9 1.0 I.I 1.2 1.3 t.l I.S 1.6 1.7 I.B 1.9 2.0 2.1
SOLIDITY RATIO a/T = 0

FIG. 9. Variation of Maximum Efficiency with Solidarity Ratio a

Specific Speed
The specific speed of the CFT was calculated by using Eq. 14 and the
measured values of AT, P, and H for all the runners as shown in Tables 2-6.
Analyses of these data have indicated dependence of the specific speed on
both nozzle entry arc and unit discharge. Therefore, a multiple linear
regression analysis was conducted between the specific speed as the
dependent variable and the nozzle entry arc and unit discharge as the
independent variables. The results were:
Ns = 6.67 - 18.85 X + 24.34 Q' (16)
for runner number 1 (Dr/B = 1.0), and
311

J. Hydraul. Eng. 1988.114:299-314.


Ns = 9.44 - 6.34 X + 25.07 Q' (17)
for runner number 2 (DXIB = 2.0). If Eq. 17 is applied to runners 3 and 4,
the predicted specific speeds will be very close to the measured values. It
should be noticed that the specific speeds obtained in this study fall within
the range given for the CFT by Warnick (1984).
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

CONCLUSIONS

The results of this study would be most useful in designing CFTs to


operate at maximum efficiency when conditions vary from the assumptions
made for the previous theoretical treatments of the turbine. The following
summary conclusions provide helpful information for the design of the
turbine. Given this information, the optimum design dimensions of the
CFT components will come from simultaneous solution of Eqs. 9, 11, 12,
13 or 14, and 16 or 17.
1. The unit discharge passing through the runner increases by increas-
ing the nozzle entry arc from 58° to 90°, increasing the runner aspect ratio,
or reducing the number of blades.
2. The maximum efficiency of the CFT at any flow/head combination
increases by increasing the nozzle entry arc from 58° to 90° as shown
consistently in Tables 2-6.
3. The maximum efficiency r\ at a constant nozzle throat-width ratio
SQ/J-JX occurs at about the same runner unit rotational speed N', regardless
of the nozzle entry arc X, flow/head variations, runner diameter, or the
number of blades.
4. The runaway speed of the runner decreases as the nozzle entry arc
increases, which corresponds to a decrease in the velocity of the water at
the nozzle tip, as shown in Tables 2-6.
5. By reducing the aspect ratio DJB of the runner from 2.0 to 1.0,
efficiency decreases about 20%.
6. Efficiency of the CFT shows strong dependence on the number of
blades.
7. The optimum number of blades in this experiment was approxi-
mately 15 for the 30.48-cm (12-in.) runner diameter with the aspect ratio of
2.0. Eq. 9 has been proposed for the optimum number of blades for any
size turbine.
8. The maximum efficiency for the runners tested occurs at about the
speed ratio U1/Vl of 0.53-0.54, which is very close to the theoretical value
of 0.52.
9. For each nozzle entry arc, the total head decreases as the number of
blades decreases for a constant discharge, as summarized in Tables 2-6.
10. The optimum value of the solidity ratio o- is 1.03 for the nozzle entry
arc of 90° and the ratio D2IDi = 0.68. This value is used for proper design
of the CFT.
11. The specific speed Ns of the CFT varies with nozzle entry arc,
aspect ratio, and speed ratio as given by Eqs. 16 and 17.

APPENDIX I. REFERENCES
Armstrong, E. L. (1978). The impact of the world's energy problems on lowhead

312

J. Hydraul. Eng. 1988.114:299-314.


hydroelectric power. University of Idaho, Moscow, Idaho, 1978.
Balje, O. E. (1981). Turbomachines. John Wiley & Sons Inc., New York, N.Y.
Feasibility studies for small-scale hydropower additions (1979). U.S. Army
Corps of Engineers, Hydrologic Engineering Center, Davis, Calif., Jul.
Haimerl, L. A. (1960). "The cross-flow turbine." Water Power, 22(1), 5-13.
Inversin, A. R. (1981). Micro-hydropower schemes in pakistan, a case study.
National Rural Electrification Cooperative Association, International Programs
Division, Washington, D.C.
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

Johnson, W., Ely, R., Ill, and White, F. (1983). "A new approach to ULH
development using the cross-flow turbine." Proc, Water Power 1983:
Internal. Conf. on Small-Scale Hydropower, Tennessee Valley Authority/U.S.
Army Corps of Engineers, September 18-21, 1983, pp. 1403-1407.
Khosrowpanah, Sh. (1984). "Experimental study of the C-F-T," dissertation
presented to Colorado State University, Fort Collins, Col., in partial
fulfillment of the requirements for the degree of Doctor of Philosophy.
Mockmore, C. A., and Merryfield, F. (1949). "The Banki water turbine."
Bulletin Series No. 25, Engineering Experiment Station, Oregon State System
of Higher Education, Oregon State College, Corvallis, Oreg.
Nakase, Y., Fukutomi, J., and Watanabe, T. (1982). "A study of cross-flow
turbine, effects of nozzle shape on its performance." Small hydro-power fluid
machinery (Annual Winter Meeting) American Society of Mechanical
Engineers, New York, N.Y., 129-133.
Ossberger Turbinenfabrik. (1981). Ossberger, Weissenburg, Bavaria, Germany.
Shepherd, D. G. (1956). Principles 'of turbomachinery. The MacMillan Company,
New York, N.Y.
Sonnek, E. (1923). The theory of the cross-flow turbine. Berlin Verlag Von Julius
Springer.
Van Dixhorn, L. R. (1984). "Experimental determination of blade forces in a
cross-flow turbine," thesis presented to the Virginia Polytechnic Institute and
State University, Blacksburg, Va., in partial fulfillment of the requirements for
the degree of Master of Science.
Varga, J. (1959). "Tests with the Banki water turbine." Acta Technica
Academicae Scientiarum Hungaricae, 27(1/2), 79-102.
Warnick, C. C , Mayo, H. E., Jr., Carson, J. L., and Sheldon, L. H. (1984).
Hydropower engineering, Prentice Hall, Inc., Englewood Cliffs, N.J.

APPENDIX II. NOTATION

The following symbols are used in this paper:

AR = aspect ratio of runner;


a = radial rim width;
B = width of runner;
Z), = outer diameter of runner;
D2 = inner diameter of runner;
Q = gravitational acceleration;
H = available head of water;
N = rotational speed of runner;
N' = unit rotational speed of runner;
Ns = specific speed;
n = number of blades;
P' = unit power;
Pin = input power;
•P
'out = output power;
Q = discharge;
Q' = unit discharge;
313

J. Hydraul. Eng. 1988.114:299-314.


outer radius of runner;
inner radius of runner;
nozzle throat width;
nozzle throat-width ratio;
blade spacing;
peripheral velocity of runner;
absolute velocity of flow;
Downloaded from ascelibrary.org by University of Exeter on 07/15/15. Copyright ASCE. For personal use only; all rights reserved.

relative velocity;
angle between absolute and peripheral velocity;
blade angle;
specific weight of water;
hydraulic efficiency of turbine;
maximum efficiency of turbine;
nozzle entry arc; and
solidity ratio.

first-stage inlet;
first-stage outlet;
second-stage inlet; and
second-stage outlet.

314

J. Hydraul. Eng. 1988.114:299-314.

You might also like