IOP Conference Series: Earth and Environmental Science

Related content
- Some experiences with flow measurement
The comparative analysis of the current-meter in bulb turbines using the differential
pressure method
method and the pressure-time method used for A Adamkowski and M Lewandowski

- Experimental and numerical results of the

discharge measurements in the Kaplan turbine influence of dynamic Poisson effect on
transient pipe flow parameters
penstocks A Adamkowski, S Henclik and M
Lewandowski

- Treatise on water hammer in hydropower

To cite this article: A Adamkowski and Z Krzemianowski 2012 IOP Conf. Ser.: Earth Environ. Sci. 15 standards and guidelines
022021 A Bergant, B Karney, S Pejovi et al.

View the article online for updates and enhancements.

This content was downloaded from IP address 114.4.216.61 on 08/09/2019 at 20:36

26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

The comparative analysis of the current-meter method and

the pressure-time method used for discharge measurements in
the Kaplan turbine penstocks

A Adamkowski and Z Krzemianowski

The Szewalski Institute of Fluid-Flow Machinery, Polish Academy of Sciences,
Fiszera Str. 14, 80-231 Gdansk, Poland

E-mail: aadam@imp.gda.pl

Abstract. The paper presents experiences gathered during many years of utilizing the current-
meter and pressure-time methods for flow rate measurements in many hydropower plants. The
integration techniques used in these both methods are different from the recommendations
contained in the relevant international standards, mainly from the graphical and arithmetical
ones. The results of the comparative analysis of both methods applied at the same time during
the hydraulic performance tests of two Kaplan turbines in one of the Polish hydropower plant
are presented in the final part of the paper. In the case of the pressure-time method application,
the concrete penstocks of the tested turbines required installing a special measuring
instrumentation inside the penstock. The comparison has shown a satisfactory agreement
between the results of discharge measurements executed using the both considered methods.
Maximum differences between the discharge values have not exceeded 1.0 % and the average
differences have not been greater than 0.5 %.

1. Introduction
Water discharge (volumetric water flow rate) belongs to the group of a few basic quantities needed to
determine the hydraulic performance characteristics of hydraulic turbines and pumps. Discharge
always represents the most difficult quantity to measure and accuracy of its measurement is worse and
very difficult to estimate in comparison with the specific hydraulic energy (head) and active power.
Despite immense progress in discharge measurement techniques, this part of the hydraulic machine
performance tests is often a major challenge, even for experienced test teams.
Either the method of the local velocity distribution determined by means of the current-meters or
pressure-time method (the so-called Gibson method) belong to primary methods for discharge
measurement in hydropower systems [1], [2]. Conducting the measurements using these methods
requires not only an appropriate application of measuring devices, but correctly carrying out process
of analyzing the data, including the use of integration techniques.
The paper presents authors’ own experiences gathered during many years of utilizing the
mentioned methods for flow rate measurement in many hydropower plants. The special integration
techniques (concerning both methods) in the form of their own coded programs using the progressive
numerical algorithms have been developed. The techniques differ from the recommendations
contained in the relevant international standards [1], [3], mainly from the graphical and arithmetical
ones.

Published under licence by IOP Publishing Ltd 1

26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

The results of the comparative analysis of both methods used simultaneously during the hydraulic
performance tests of two Kaplan turbines in one of the Polish hydropower plant are presented in the
final part of the paper.

2. Current meter method

Propeller water current meters are very often used in the velocity-area method for absolute discharge
measurements in low-head hydropower systems. The volumetric flow rate is determined by integrating
the distribution of local flow velocities over the entire area of the measuring (hydrometric) cross-
section. The local flow velocity component is measured based on the current meter rotor revolutions
counted in a given time period, and the experimentally determined relationship between the current
meter rotation speed and local flow velocity.
Uncertainty of measurement by the means of the current meters depends on a lot of factors, mainly
on: (1) measurement of local velocity using current meters; (2) measurement of a cross-section
geometry of flow channel; (3) determining of streamlines (current meters orientation); (4) determining
of parameters for velocity function in a boundary layer; (5) the applied method of calculations.
From the authors’ experience it can be concluded that the last two factors may have very important
influence on uncertainty of discharge measurement. The differences can exceed significantly more
than 0.5 % for different approaches of calculations. Particularly it concerns the flows with highly
irregular velocity distribution, especially in case of short intakes.
The velocity field integration methods are based on graphical and arithmetic techniques according
to the very outdated literature [3], [5]. However, due to big progress in computer technology, graphical
and arithmetic techniques have been practically replaced by various numerical schemes. Nowadays,
the most popular ones are based on spline techniques [6], [7]. The expected advantages of the spline
approach are, inter alia, the following: (1) increasing the accuracy of calculation in comparison to
graphical and arithmetical techniques; (2) easy and fast carrying out the flow rate calculations;
(3) easy possibility of applicability to cases with irregular velocity distributions; (4) easy possibility of
an immediate visualization of velocity distributions etc.
The most often applied technique calculates the discharge using Classic Cubic Splines. It should be
highlighted that this approach may lead sometimes to inconsistent with reality (inaccurate) flow
velocity distributions, especially in highly irregular regions of velocity. It is related to the difficulties
to obtain smoothed velocity function that interpolates the areas of strong bending curve as it happens
in a boundary layer region. This problem illustrates figure 1.

Figure 1. Comparison of the curves

based on the classic cubic splines and
the B-spline functions (NURBS).
Example from authors’ own
measurement.
Because of that the authors of this paper decided a few years ago to adapt another more
sophisticated spline technique to interpolate velocities obtained by the means of the current meters.
This technique involves the advanced spline functions, the so-called Non-Uniform Rational B-Splines
(NURBS) [8]. Nowadays, this kind of splines is commonly used in modeling of the complicated
geometrical shapes because of their smoothness. It has been assessed that it represents much better
kind of interpolation than the classic spline functions. Figure 1 presents an example of the comparison
between the interpolation functions based on both mentioned techniques.

2
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

The authors’ own software, called FLOWMAX, is used to calculate volumetric flow rate Q by the
means of current meters in a cross-section A represented in rectangular (x, y) or circular (r, )
coordinate system, which may be written, respectively, as follows:
Q  Vz ( x, y)dxdy Q  rVz (r, )drd (1)
A A

where: x, y – coordinates in Cartesian system; r,  – radial and angular coordinates in cylindrical

system; Vz – normal velocity distribution obtained using current-meters in cross-section A and von
Karman law in a boundary layer zone.
Program gives possibility to calculate the volumetric flow rate from local velocities measured in
the hydrometric sections of different shapes for rectangular closed conduits or open channels (with
optional chamfers and roundings) or circular.
Generally, the essential principal of the NURBS approach can be very shortly characterized as
follows. At each interpolated point, the NURBS function value results from the linear sum of four 3rd
degree polynomials, properly calculated from the given points. The mutual influence of each
polynomial at an interpolated point allows avoiding exaggerated deformation and makes the
interpolated curve be more smoothed than the classic cubic spline in which only one 3rd degree
polynomial is considered at each interpolated point.
According to the program FLOWMAX, the velocity distribution in the boundary layer is
interpolated by the von Karman formula that can be written in the following form:
1m
 x 
Vz  BL x   Vz0  
 (2)
  BL 

where: x – distance from wall, Vz0 – velocity of the nearest current meter to boundary layer, BL –
boundary layer thickness, m – boundary layer exponent dependent on a Reynolds number.
According to the ISO 3354 [3], the boundary layer thickness is calculated as follows:
W  Z
 BL  (3)
Re 0z.2

where: Z – distance from water intake to cross-section of current meters, W – empirical coefficient
recommended by international standards to be equal: W = 0.37, Rez – Reynolds number dependent on
a distance Z calculated as follows:
Vzav  Z
Re z  (4)

where: Vz-av – arithmetical mean velocity from the current meters,  – water kinematic viscosity.
The boundary layer parameter m may be determined according to the procedure described in the
ISO 3354. For this purpose the linear loss coefficient  in a conduit can be calculated, for instance,
using the formula [9]:
2
  

  2  lg 0.27  
2.51
 (5)
 Dh Re  0.4  Re  0.3  0.0053 
  
where:  – wall roughness, Dh – hydraulic conduit diameter of a hydrometric cross-section, Re –
Reynolds number in a cross-section calculated as follows:
Vzav  D h
Re  (6)

The m value is obtained on a basis of a table placed in the ISO 3354 that contains m values for 
ones (it may be introduced by user of the program as well). The software has been successfully used in
the hydraulic turbine performance tests in Poland for several years. It can be especially recommended

3
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

for highly irregular flows with high Reynolds numbers. Such conditions exist often in hydropower
systems.
In authors’ opinion, in cases of machines with a short intake and irregular inflow, using the classic
cubic spline approach in the current meter flow measurement may lead to overestimation of efficiency
of tested hydraulic turbines (underestimation of discharge). The technique of interpolation, presented
and recommended in this paper, may significantly improve the accuracy of discharge measurement
results. An example of the use of this technique, comparing to the pressure-time method use, is
presented in the further part of the paper.

3. Pressure - time method

The method is based on the second law of dynamics as applied to the decelerated mass of liquid
stream flowing through a pipeline. The inertia force of the stopped liquid mass is manifested by the
pressure difference between two measurement sections in the pipeline – see figure 2.

Figure 2. Segment of a
pipeline with marks needed to
explain the theoretical basis of
the pressure-time method.
The discharge is calculated by integrating the recorded pressure difference curve within properly
determined time interval according to the formula [10]:
tf

 p(t)  p (t)  P (t)dt  q

1
Q0  d f l (7)
C
t0

where:  – liquid density, C – geometrical factor of the pipeline segment between cross-sections 2-2
and 1-1 (L length and A cross-section area), p = p2+gz2-p1-gz1 – the pressure difference measured
between sections 2-2 and 1-1 related to a reference level, pd – the dynamic pressure difference
between sections 2-2 and 1-1, Pf is the pressure drop due to friction losses between 1-1 and 2-2 cross
sections, ql – discharge under terminal conditions (usually the leakage rate through the cut-off device
in the closed position that has to be measured or assessed using a separate method), t – time, and (t0, tf)
means time interval in which the flow conditions change from initial to final stage.
The value of C factor has to be determined basing on geometry measurement of L distance between
section 1-1 and 2-2 and A internal pipe cross-section area, from the formula:
L
dx
C
 A(x)
0
(8)

The pressure drop Pf due to hydraulic loss in the pipeline segment between sections 1-1 and 2-2
and the dynamic pressure difference pd in these cross-sections have to be extracted from the
measured static pressure difference p.
In the discussed method the friction pressure drop is determined using the square discharge
function:
Pf  K f Q Q (9)
in which the constant Kf is calculated basing on the measured values of the initial flow conditions:
P f 0
Kf  Kf0  (10)
Q0 Q0

4
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

The value of pd is independent of flow direction and can be calculated from the discharge function:
2 1
pd (t )  Kd Q(t )2 Kd 
2 A22

2 A12 (11)
where:
where: 1, 2 – the Coriolis coefficients for 1-1 and 2-2 sections, respectively [11].
The need for calculating Pf and pd quantities, using their simplified dependence on the square of
the flow rate (equations (9), (10), (11)), unfavorably affects the uncertainty of measurement results of
the method. Therefore, it is of great importance for achieving good accuracy of the flow measurement
performed using the pressure-time method, that the contribution of the pressure difference attributed to
friction loss and difference of dynamic pressures would be possibly the smallest and would not exceed
a certain limit. This requirement can be written in the form of the following inequality:
P f 0  p d 0
 (12)
p m

where: pm – the average value of the static pressure difference measured between the sections during
liquid stream stop, but index ‘0’ refers the initial flow conditions. According to the IEC 60041
standard, the value of  is equal to 0.2 (20 %).
The theoretical basis of the pressure-time method presented above is valid for both turbine and
pump modes of operation. However, the IEC 60041 standard recommends using the method only in
cases of turbine operation mode. The own experiences indicate on the possibility to utilize the method
also in cases of pumping operation mode [12]. One of the necessary requirements in such cases is
correct calculating the pressure drop caused by the friction loss between the hydrometric cross sections
of a pipeline. Typical calculation procedures, including the presented in [1] and [2], assume the
pressure drop to be dependent only on the square of the discharge value, as in the following equation:

Pf  K f Q 2 (13)
The hydraulic losses calculated in accordance with equation (13) do not depend on flow direction
(both are always of the same sign) – as contrary to the equation (9). Therefore, following this way of
calculation may lead to additional error while determining the discharge in the pressure-time method.
Calculation of the initial value of discharge Q0 using the equation (7) requires to specify the time
integration limits t0 and tf. These values should determine the time interval in which the flow is cut-off.
Contrary to t0 time (lower limit of integration), the determination of tf time (upper limit of integration)
presents difficulties. Even precise synchronization of recording of the flow cut-off device run with
measurement of the pressure rise does not ensure the exact determination of tf time value. The reason
for this is often the lack of the strict relation between the time moment at which the closing device
movement is stopped and the time moment of flow cut-off finish – in some cases despite the
termination of flow cut-off run, the closing device is still in motion, e.g. in result of elastic strain.
Therefore, the upper integration limit tf is determined from the character of free pressure oscillations
[1]. These oscillations, as residual ones, remain in the penstock directly after the termination of flow
cut-off. One of the procedures relating to the upper integration limit calculation in the pressure-time
method is given in the IEC 60041 standard. However, the procedure includes mathematical inaccuracy
– it does not ensure to set a zero-value integral of free pressure oscillations that intent to eliminate
their influence on the discharge measurement – what has been proved in paper [12].
On the basis of theoretical consideration above presented, the original program, called GIB-ADAM
has been developed in the Institute of Fluid-Flow Machinery PAS (IFFM PAS) in Gdansk, Poland.
The program is one of the most important tools enabling the use of the pressure-time method in
practice. The first practical application of this method with using the developed GIB-ADAM program
was undertaken in Poland in the second half of 90’s. Since 1998, IFFM PAS has used different version
of the pressure-time method in numerous plants in Poland and in Mexico [13].

5
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

4. Simultaneous use of the pressure-time and current meter methods

During the hydraulic performance tests of two similar Kaplan turbines in one of the Polish
hydropower plant, the water current meter and pressure-time methods were used simultaneously to
measure the discharge. The hydrometric section 0-0 for using the current meter method and the cross
sections 1-1 and 2-2 for the pressure-time method were located in a concrete cylindrical penstock
segment of 4 m inner diameter (D) – figure 3. It is worth highlighting that both methods, in a similar
way, were applied to two turbines.
A cross-section 0-0 was established in distance of about 8 m (2 D) from the inlet of cylindrical
penstock (distance from the intake grating threshold was about 16 m). In this cross-section 25 current
meters were mounted on the stationary supporting frame – figure 4. The trailing edges of current
meters were located in front of supporting frame in distance of 10 diameters of the supporting frame
arm from its axis. Hence, systematic uncertainty of velocity measurement taking into account
supporting frame presence did not exceed 0.2 %.
All applied current meters were calibrated (tarred). The blockage effect was taken into
consideration according the ISO 3354 standard.

Figure 3. Layout of the turbine

penstock with marked
hydrometric sections used for
discharge measurement using the
current meter method (0-0) and
Gibson method (1-1) and (2-2).

Figure 4. The current meters

and the stationary supporting
frame used to mount the current
meters.
Figure 5 presents the samples of the measured local velocity distributions obtained by the means of
the FLOWMAX program. The impulse current meters data acquisition was made by the means of a
digital card connected to computational system of data acquisition.

Figure 5. Example
of profiles of
velocity distributions
to calculate flow
discharge obtained
by the means of
FLOWMAX.
According to the estimation, the total systematic uncertainty of discharge measurement results
using the current meters in the considered cases was not greater then +/–1.5 %.
In the considered tests, the classic version of the pressure-time method was used as the second
method of discharge measurement. This version is based on direct measurement of pressure difference

6
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

between two hydrometric sections of the straight conduit of the same diameter using a pressure
differential transducer. Both hydrometric sections 1-1 and 2-2 for using the pressure-time method were
located in a concrete cylindrical penstock segment of 4 m diameter – figure 6.
Section 1-1 was located about 10 m (2.5 D) downstream the beginning of cylindrical segment
penstock inlet and section 2-2 was located about 10 m upstream the turbine spiral case inlet. The
length of segment between sections 1-1 and 2-2 was about 17 m for both tested turbines. In each of
these sections four pressure taps were uniformly located on the circumference – figure 6.
Due to lack of access to the penstock from outside the special measuring equipment was installed
inside the penstock [13]. In each measuring section (1-1 and 2-2) four flat bars with pressure receiving
holes (taps) were mounted to the internal side of the penstock concrete wall, parallel to the water flow
direction and connected to the manifold by means of the impulse cooper tubes – figure 6.
The manifolds of both hydrometric sections were connected by means of impulse tubes to the sealed
housing with the differential pressure transducer installed inside.
The static pressure difference measured by the precise differential pressure transducer between
sections (1-1) and (2-2) was recorded by the computer data acquisition system with frequency
sampling of 200 Hz. Then the GIB-ADAM software was used to calculate the values of the discharge.
Figure 6. Distribution of
the pressure receiving holes
in the penstock hydrometric
sections 1-1 and 2-2 (left)
and measurement elements
installed inside penstock in
hydrometric section 1-1
(right).
Example of the pressure difference recorded between the measuring cross-sections and flow rate
calculated from this difference is presented in figure 7.

Figure 7. Example of
measurement of discharge
through a turbine using the
pressure-time method.
The water discharge in the final conditions, as the rate of leakage flow through the closed guide
vanes of the tested turbines, was determined basing on the measurement of rate of water level decrease
in the cylindrical segment of the penstock. The total systematic uncertainty of discharge results by
means of the pressure-time method used in the considered cases was estimated on not greater than +/–
1.5 %.
Comparison between discharge measurement results received by means of the pressure-time and
current-meter methods used in the tests of two similar water turbines are presented in figure 8.

7
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

Figure 8. Relative
differences between
discharge measurement
results by means of
Gibson and current-meter
methods used for two
similar turbines.
It shows a satisfactory agreement between the results of discharge measurements using the both
methods. Maximum relative differences between the discharge values have not exceeded 1.0% and the
average differences are not greater than 0.5 %. The greater differences can be observed for lower value
of measured flow rate. Such effect is associated with increased inaccuracy of measurement with the
decrease of the measured discharge values. The results of comparison confirm the reliability of the
techniques and computational software developed and used by the authors of this paper in the
measurement of flow rate through hydraulic machines.
5. Conclusions
Flow rate measurements using the local velocity distribution determined by the current-meters and
pressure-time methods, the most important primary methods in hydropower systems, require in
addition to proper use of measuring devices, correctly conducting analyses of the measuring data,
including the integration techniques. The experiences of the authors of this paper show that the
techniques of integration can affect the measurement results more than 0.5%.
The NURBS functions have been adopted for calculating the discharge from the local velocities
measured using the current meters. It is assessed that the NURBS represent much better kind of
interpolation than the classic cubic spline functions, particularly in area of connections of the very
strong velocity gradients in von Karman law boundary layers with velocity mainstream (core) regions
at measuring cross-sections, and in cases with very irregular highly turbulent flows. On this basis the
NURBS technique may be recommended for practical use of the current meter method, particularly in
cases of highly irregular flows in open channels and closed conduits with high Reynolds numbers.
The pressure-time method has been developed by introducing some modifications to the integration
procedure in comparison to the IEC 60041 and ASME PTC 18 standards. The modifications concern
generally the calculating the hydraulic losses and determining the upper limits of integration of
recorded pressure variations in time.
The results of the comparative analysis of both methods, with introduced modifications, used
simultaneously during the hydraulic performance tests of two Kaplan turbines in one of the Polish
hydropower plant have shown a satisfactory agreement between the results of discharge measurements
executed using the both compared methods. Maximum differences between the discharge values have
not exceeded 1.0 % and the average differences have not been greater than 0.5 %. These results and
the results of many authors’ experiences from the use of the analyzed methods in different types of
hydropower plants, assessed as very satisfactory, can confirm the validity of applicability of
introduced modifications in both the methods.

References
[1] IEC 60041 1991 Field acceptance tests to determine the hydraulic performance of hydraulic
turbines, storage pumps and pump turbines
[2] ASME PTC 18 2002 American National Standard: Hydraulic Turbines and Pump–Turbine,
Performance Test Codes (Consolidation of ASME PTC 18-1992 and ASME PTC 18.1-1978)
[3] ISO 3354 1988 Measurement of clean water flow in closed conduits - Velocity-area method
using current-meters in full conduits and under regular flow conditions
[4] ISO 7194 2008 Measurement of fluid flow in closed conduits - Velocity-area methods of flow
measurement in swirling or asymmetric flow conditions in circular ducts by means of
current-meters or Pitot static tubes

8
26th IAHR Symposium on Hydraulic Machinery and Systems IOP Publishing
IOP Conf. Series: Earth and Environmental Science 15 (2012) 022021 doi:10.1088/1755-1315/15/2/022021

[5] Troskolanski A T 1960 Hydrometry (Pergamon Press Ltd.: Oxford London New York Paris,
PWN: Warsaw)
[6] Berny R and Slota R 2000 Integrated system for hydrometric data processing and discharge
evaluation in open channels and closed conduits Int. Conf. HYDROFORUM’2000, (Gdansk,
Czorsztyn Poland, September 2000) (IMP PAN Publishers) pp 247-274
[7] Leutloff S, Geromiller W, Fischer G 1983 Numerische Auswertung von Flügelmessungen in
Kreisquerschnitten (Elektrizitätswirtschaft, Jg.82, H.5) pp 139-144
[8] Yamaguchi F 1988 Curves and surfaces in computer aided geometric design (Berlin: Springer
Verlag)
[9] Nackab J 1988 Calcul direct, sans iteration, de la parte de charge en conduite par la formule de
Colebrook (La Houille Blanche) No. 1
[10] Adamkowski A 2012 Discharge Measurement Techniques in Hydropower Systems with
Emphasis on the Pressure-Time Method (Chapter no. 5 of the book Hydropower - Practice
and Application) ISBN 978-953-51-0164-2 pp 83-114
[11] Cengel Y A and Cimbala J M 2006 Fluid Mechanics. Fundamentals and Applications (New
York: McGraw-Hill International Edition)
[12] Adamkowski A and Janicki W 2010 Selected problems in calculation procedures for the Gibson
discharge measurement method Proc. of 8th Int. Conf. on Hydraulic Efficiency Measurement
– IGHEM 2010 (Roorkie India) pp 73-80
[13] Adamkowski A, Kubiak J, Sierra E F, Urquiza B G, Janicki W and Fernandez D J M 2008
Hydraulic Review Worldwide 16(6) 40-49