Flow Measurement Using the Dye Dilution Technique

M. Cyrenne P.Eng
Ontario Power Generation

Presented to:
IGHM 2002 International Conference
Toronto, ON
July 17-19, 2002.

Abstract pumps, hydraulic turbines, heat exchangers, service

The Performance and Testing Department of Ontario water pumps and service water systems.
Power Generation, Electricity Production Services The application of our dye dilution technique was
Division, offers a less common but valuable technique developed approximately 30 years ago and has been
for water flow measurement. Our technique utilizes used with success since the early 1970’s. At that time,
fluorescent dye and a very precise procedure which the technique was proven on condenser circulating water
provides highly accurate results. The recommended dye pump performance acceptance tests. Since then, we
for this method is the fluorescent dye, Rhodamine WT. have conducted water flow tests to verify the
The dye is detectable and stable in very low performance of some of the large turbines in Ontario
concentrations, non-toxic, resistant to adsorption, mixes Power Generation’s hydroelectric power stations.
readily in water, and is not usually present in natural We have also conducted performance tests for the
water systems. Its fluorescence level which is Electric Power Research Institute at hydraulic power
proportional to its concentration in water is accurately plant Kootenay Canal in British Columbia and Grand
measured with a precision fluorometer. Coulee Dam in Washington. These tests were conducted
as part of a research project supporting the revision of
The method involves four key functions: the ASME performance test code for hydraulic turbines.
• Injection of dye into the system This method of flow measurement is also included in
• sampling the water downstream the International Standard IEC 41 1991-11 and will be
• preparation of a “standard” of known dilution and included in the next revision of ASME PTC 18
• comparison of the “sample” and the “standard” Hydraulic Turbines. Both of these codes provide
procedures for determining field performance testing of
The technique has proven to be inexpensive, accurate hydraulic turbines and of pump/turbines operating with
and adaptable. The test needs no elaborate facilities and water in either the turbine or pumping mode, by
equipment can be set up quickly and simply. The test measuring flow rate (discharge), head, and power, from
results are extremely accurate, and can reliably achieve which operating efficiency may be determined.
results with an accuracy of ± 1.5%. In addition, the
system is extremely adaptable and has a wide range of Introduction
applications including: condenser circulating water The increase in size of Condenser Circulating Water
Pumps required for the newer generation of nuclear and

Ontario Power Generation 1

fossil fuelled power stations first commissioned during Accurate detection requires the combination of a suitable
the 1960's, posed serious problems where the dye and a sensitive instrument for detection. Early on in
measurement of flow was concerned. Up to this point in the development program, after consultation with
time, these flows were measured with the conventional specialists in the field of dye tracing, it was decided to
propeller type current meters, located on a cross-piece use Rhodamine WT fluorescent dye as the tracer, and a
spanning two diameters of the flow conduit. With the fluorometer as the detecting equipment. The
increase in conduit diameter, and the requirement for development of this dye by a leading chemical
one hundred diameters of upstream straight section to manufacturer, and the fluorometer by Turner Designs,
give uniform flow at the point of measurement, serious made the detection process practical at very low
physical limitations were foreseen. Furthermore, the concentrations.
problems normally associated with the installation of The constant rate injection problem initially proved
these current meters such as system de-watering, more difficult to solve. Early experiments with
isolation and installation of mechanical structures commercially available chemical metering pumps
became significant enough to warrant investigation into proved unsuccessful, mainly due to the fact that these,
other methods of measurement for large water flows. although accurate in dispensing a specific quality over
The use of water soluble tracers in the study of water time, did so in slugs, consistent with their reciprocating
flow is not a new scientific concept. Over a hundred action.
years ago, in France, Schloesing introduced ammonium A volumetric plexiglass cylinder, calibrated to
sulfate into a stream, and chemically analyzed water discharge at a known rate by use of compressed air, was
samples taken downstream to determine the dilution that somewhat more successful, but dependant to a large
had occurred. The possibilities for adaption to flow extent on the pressure downstream of the injection
measurement, which such a scheme provoked, produced nozzle for consistency of results.
a significant interest, and since then a great many tracing In the United States, Mariotte Vessels, floating
techniques have been devised and used, both in Europe syphons and pressurized tanks had been used
and the United States. While the United Kingdom has successfully in streams and open channels, but it was
concentrated on the use of radioactive isotopes for use as considered that these would suffer the same problems as
tracers, specifically Sodium 24, here in North America, our volumetric cylinder when used to inject the tracer
the tendency has been to use soluble chemical tracer into the suction bell of large circulating water pumps.
dyes, particularly those belonging to the fluorescent The United Kingdom Atomic Energy Authority
family. (UKAEA) was finally contacted, and they undertook to
A system has been developed by the Performance manufacture and supply a unidirectional positive
and Testing Department of Ontario Power Generation to displacement injection pump, similar to that used in their
the stage where the method is now accepted by pump radioactive tracer program. This pump has a calibrated
suppliers to Ontario Power Generation as an equally accuracy of ±0.05% on measured volume delivered, and
accurate and practical alternative to the current meter has fulfilled the stringent requirements for constant rate
system previously used. The theory of this method is and volume characteristics necessary for the accurate
included in the International Standard IEC41, Field determination of flow.
Acceptance Tests to determine the hydraulic With some measure of confidence, confirmed by
performance of hydraulic turbines, storage pumps and experiments in connection with these two fundamental
pump-turbines. It will also be included in the next requirements, a developmental program was initiated to
revision of ASME PTC 18 Hydraulic Turbines. identify the techniques, procedures and precautions
which would ensure accuracy and repeatability, at least
equal to existing methods of flow measurement. Several
Development of the System tests were carried out at Ontario Power Generation's
Measurement of the flow by the constant rate injection power stations in conjunction with current meters, and
system is based on a comparison between the the correlation of results have been excellent. During
concentration of a tracer injected at a known rate, with the developmental period, several significant
the concentration of samples extracted from some characteristics of the method, system and procedure
location, sufficiently remote from the injection point to were identified leading to the formation of techniques
ensure good mixing. The two major areas of and procedures, which now form the basis for the flow
development required in the initial stages were: measurement using the dye dilution method.
1) Accurate detection
2) Constant rate injection

Ontario Power Generation 2

Theory • Miscellaneous laboratory equipment
The dye dilution technique for the measurement of
flow is suitable for "once through" pumping systems, Dye Injection Pump
where no recirculation takes place. It involves the It is essential that the dye be injected into the system
injection of Rhodamine WT fluorescent dye at one point at a constant and precisely known rate. The pump used
in a system, and the removal of samples of water from a by Ontario Power Generation is shown in Figure 2. It is
point further downstream, after the dye has been a single stroke positive displacement pump, with
thoroughly mixed with the water. synchronous motor drive. The flow was determined by
Flow rate is established from the determination of the calibration to be 0.12513 litres/ min. ±0.05%, and for
dilution of the injected dye. one charge delivers approximately 1.2 litres over a ten
If this dye is injected at a constant rate, the relationship minute injection period. Flow rates of one half or one
between the concentration and flow is, in its simple quarter of this can be selected by means of variable gear
form: ratios in the drive mechanism.
Q1 C1 = Q2 C2
Where: Q1 = dye injection rate
C1 = concentration of injected dye Fluorometer
Q2 = flow rate to be determined The fluorometer measures the intensity of the
C2 = concentration of dye in water stream fluorescence of a sample of water with dye in solution.
This can be regarded as a relative measurement of the
Thus: Q2 = Q1 x C1 / C2 dye concentration, since fluorescence varies linearly
C1 / C2 is known as the "dilution factor", DF with concentration.
When the injected dye is passing the sampling point,
The dilution factor of a solution can not be measured the fluorometer, which is fitted with a "flow-through"
directly, but can be determined by comparing its door, allows continuous monitoring of a stream of
fluorescence (which is proportional to its concentration) water passing through the fluorometer.
with that of a specially prepared "standard solution" of During the analysis of the test samples, it is used to
precisely known dilution. This standard solution is measure the relative concentrations of test samples and
prepared by diluting a sample of the injected dye by the standard solutions for determination of the flow. For
same amount as it will undergo when injected into the this function, an interchangeable door is installed which
system. The fluorescence level of a test sample and the permits measurement of grab samples in a small test
standard solution are measured in a fluorometer, and the tube or cuvette.
dilution factor of the test sample is determined as
follows:
Rhodamine Wt Dye
DF t = DF s x F s / Ft
This dye, which was developed specifically as a
water tracer dye, is manufactured by Crompton and
Where: DF t = dilution factor of test sample Knowles, and is available in liquid form in 20%
DF t = dilution factor of standard solution solution only. It is detectable by the fluorometer in
Ft = fluorescence level of test sample concentrations lower than 0.01 parts per billion, but in
Fs = fluorescence level of standard solution practice, concentrations of about 10 parts per billion are
aimed for. Of the many fluorescent dyes available,
The flow to be measured is then: Rhodamine WT is amongst the most widely used for
this type of application, because of its low absorptive
Q2 = Q1 x DFs x F s / Ft tendencies, and other advantageous characteristics. In
concentrated form it is a dark red colour, but in the
Figure 1 shows a schematic representation of this concentration used no visible coloration of the water is
technique. apparent. It is also non-toxic and chemically
unreactive.
Equipment and Material
The principal items of equipment are:
Constant Temperature Bath
• Dye injection pump The fluorescence level of Rhodamine WT dye is
• Fluorometer extremely sensitive to temperature change, and varies
• Rhodamine WT fluorescent dye inversely with temperature at a rate of about 2.6% per
• Constant temperature bath degree Celsius. Consequently, the test samples and

Ontario Power Generation 3

 standard solutions must be at the same temperature Since the stock dye is only 20% solution, the injection
during the analysis phase of a test. This is achieved by solution should contain
means of a constant temperature bath, in which the
samples and standards are immersed prior to analysis. Q2 /Q1 x 5 x 10 x 10 -7 % of stock solution
The water in the bath is constantly circulated by
means of a small circulating pump, and it is possible to Example:
maintain the temperature of all samples to within 0.1oC
of each other. The fluorometer door, into which cuvettes If Q2 = 5000 /sec
of samples are placed for analysis, is also circulated with And Q1 = 0.002 litre/sec
water from the bath to maintain a uniform temperature. The injection solution should contain

Miscellaneous 5000 / 0.002 x 5 x 10 x 10-7 = 12.5% of stock dye

The only equipment required is for the preparation of or 125 ml of dye per
the standard solutions. This comprises standard litre of solution.
laboratory glassware, i.e. beakers and 1000 ml. Sufficient solution is prepared in a sealable container
volumetric flasks, and a laboratory weigh scale capable and thoroughly mixed.
of measurement to 0.001 grams.

Test Procedure Injection

The test procedure comprises of the following five A schematic arrangement of the injection system is
steps: shown in Figure 3. Two methods of injection may be
employed:
1. Preparation of the injection solution 1. If direct and close access to the injection point is
2. Injection of the dye possible, the dye is injected into the system through a
3. Sample collection probe inserted part way into the duct. A spring-loaded
4. Preparation of standard solutions non-return valve on the end of the probe prevents back
5. Analysis flow water between injection periods.
2. If the point of injection is unaccessible, such as the
Preparation of Injection Solution suction bell of a submerged suction wet-pit pump, then
In theory, any concentration of dye may be used for during construction, a tube is fitted leading from the
injection. In practice however, a good operating range pump room to the pump suction bell. This tube may be
for the fluorometer is with a concentration of about 10 of considerable length, and its volume may be
ppb, and the injection solution is prepared so that the significant in relation to the low rate of the injection
final concentration after injection into the system will be pump. Problems may then arise because of the dilution
about 10 ppb. The solution is prepared by diluting stock of the dye in this line between injections and the
Rhodamine WT dye with water from the system under inability of the injection pump to effectively purge the
test. Tap water is not used, and it should be noted that line, due to its extremely low flow rate.
chlorine reduces fluorescence. This is overcome by providing a continuous
secondary water supply of about 5 litres per minute
The dilution is calculated as follows: down the injection line, and injecting the dye into this
secondary supply. This has the effect of transporting
C1 = Q2 /Q1 x C2
the dye into the pump suction in the minimum of time.
The quantity of this supply is not critical, but it should
Where: be maintained constant, as any fluctuation would
C1 = concentration of injection solution (ppb) effectively change the rate of injection. Chlorinated tap
C2 = desired concentration after injection (= 10ppb) water in not used, as chlorine may reduce the
Q1 = injection pump flow rate (litre/sec) fluorescence. The supply is usually taken from the
Q2 = anticipated flow being measured (litre/sec) local service water system.
Injection of dye for the measurement of flow in
Thus: C1 = Q2 /Q1 x 10 ppb hydraulic turbines involves extremely large volumes of
flow and sometimes short mixing lengths. This
i.e. the injection solution should contain problem has been resolved the installation of temporary
injection grids. These grids are used to increase the
Q2 /Q1 x 10 x 10-7 % dye number of injection nozzles and therefore facilitate the
mixing process. The dye is also injected with the

Ontario Power Generation 4

 nozzles facing upstream to further improve the mixing
of the dye. Since this factor is of large magnitude, frequently in
order of several millions, the standard solutions are
prepared by "serial dilution", in which each solution is
Sample Collection diluted in turn until the correct factor is attained.
The sampling location is selected to allow sufficient A four serial dilution is usually performed, in which
distance from the injection point to ensure complete the dilution factor for each successive dilution will be:
mixing of the injected dye. This is generally achieved in 1/4
about one hundred diameters of straight pipe, but may be DFp = Q2/ Q1
greatly reduced by the presence of bends or other causes
of flow disturbance in the line. Example:
The sampling equipment is shown in Figure 4. Its If Q1 = 0.002 litre/sec
sole purpose is to determine when the dye is passing the Q2 = 5000 litre/sec
sampling point, and when the sample should be taken. 1/4
No analysis is donicke at this point. DFp = 5000 / 0.002 = 39.764
As dye is being injected at a point upstream, a
continuous sample of water is bled from the system and That is, each 1000 ml. flask of solution will contain
passed through the fluorometer (and then to drain) at a
rate of about 4 litres per minute. As the injected dye 1000 / 39.764
reaches the sampling point, and passes through the
fluorometer, the chart recorder coupled to the signal or 25.148 ml. of the previous solutions.
output of the fluorometer will indicate an increase in For greater accuracy the dye is weighed out and
fluorescence. When this level stabilizes at a plateau, topped up to the 1000 ml. mark. As a safeguard against
indicating uniform dye concentration, a sample is bled errors while preparing the standard solutions, two sets
off slowly over a period of about five minutes. The are prepared independently, and their agreement is
samples are collected in 500 ml. dark glass bottles, and checked during the analysis.
stored away from light until analysis is conducted. As
the last of the slug of dye passes the sampling point
(after about 10 minutes) the chart trace will return to Analysis
zero. A typical chart recording is shown in Figure 5. The arrangement for equipment for the analysis is
Observation of the chart trace will indicate the degree shown in Figure 6. Test samples and standard solutions
of mixing achieved. A straight, level plateau indicates are stored in the temperature equalizing bath until all
good mixing, while erratic trace indicates incomplete temperatures have stabilized and are uniform to within
mixing, in which case sampling should be done further 0.2oC.
down stream. The fluorescence of a test sample and a standard
solution is determined by inserting a cuvette of each into
the fluorometer. This is repeated six times, and the
Preparation of Standard Solutions average value obtained for each.
The standard solutions are prepared in the laboratory It should be noted that the fluorescence level is an
by diluting a sample of the injection solution to the same arbitrary value, and need not be calibrated, since it is
concentration as is anticipated in the samples. If tests only the ratio of the two fluorescence levels which is
have been conducted at several flow rates (as in the case required.
of a pump test), then a standard solution is prepared to The flow is then determined as follows:
match each flow. The water used to prepare the
standards is taken from the system under test, so that Q2 = Q1 x DFs x Fs /Ft
they have the same levels of turbidity, pH, etc. as the test
samples. Chlorinated tap water is not used to prepare Where:
standard solutions.
The dilution factor for preparing the standards will Q1 = flow rate of dye injection pump (litre/sec)
be: Q2 = flow rate to be measured (litre/sec)
DF = Q2 / Q1 DF s = dilution factor of standard solution
Fs = fluorescence level standard solution
Ft = fluorescence level of test sample
Where: Q1 = flow rate of dye injection pump
Q2 = flow rate to be measured

Ontario Power Generation 5

Accuracy of Measurement Advantages of the Technique
Comparative tests were conducted during an • There is no need for down time of equipment before
acceptance test of a condenser circulating water pump, and after tests, with the attendant isolation and de-
where the flow as measured by the dye dilution method watering for the installation and removal of test
was compared with that measured simultaneously by equipment.
current meters, calibrated to within ±1.0%. Eight tests • No modification to the system is necessary for the
were conducted at various flows ranging from 82000 installation of orifices or other types of flow meters.
usgpm to 12400 usgpm, and the average deviation of the • There is no interference with the system operation
flow measured by dye dilution, from that measured by due to pressure drops caused by flow measurement
the current meters was +0.7%. A statistical analysis devices.
which took into account the ±1.0% accuracy of the • Diagnostic tests can be conducted at very short
current meters, indicated a total uncertainty of ±1.6% notice, since no test preparations are necessary, other
(95% confidence level). than to ensure that injection and sampling points are
Error analysis conducted on results from other tests, available.
which took into account the systematic and random
errors for each parameter involved, indicate an
uncertainty of +1.1% (95% confidence level).

Field Experience
With the present methods and equipment, flow in
excess of 63 m3/s (1,000,000 usgpm) can be measured.
This capability could be increased by an increase in the
rate of dye injection.
On-site performance and diagnostic tests using the dye
dilution method have been conducted on the following
pumps and systems:

STATION CAPACITY
Lennox GS - CCW Pumps (8) 5.43 m3 /s (86000 usgpm)
Bruce NGS - CCW Pumps (2) 12.93 m3/s (205000 usgpm
Bruce HWP - Lift Pumps (2) 11.61 m3/s (184000 usgpm)
Bruce HWP - CW Pumps (2) 5.36 m3/s (85000 usgpm)
Douglas Pt. - Process Water Pumps (2) 0.82 m3/s (13000 usgpm)
Pickering NGS - CCW Pumps (2) 11.99 m3/s (190000 usgpm)
Bruce NGS - Moderator Heat
Exchange Study (2) 0.55 m3/s (8700 usgpm)
Galt Energy Systems HGS 4.90 m3/s (173 cfs)
Decew Falls HGS 10.5 m3/s (370 cfs)
Elliot Chute HGS 20 m3/s (20 cfs)
Kakabeka Falls HGS 20 m3/s (706 cfs)
Ontario Power HGS 21.8 m3/s (706 cfs)
South Falls HGS 3.1 m3/s (110 cfs)

Ontario Power Generation 6

 Figure 1
Dye Dilution Water Flow Measurement Technique

Figure 2
Diagram of Constant-Rate-Injection Pump
Ontario Power Generation 7
 Figure 3
Arrangement of Injection Equipment

Figure 4
Arrangement of Sampling Equipment
Ontario Power Generation 8
 Figure 5
Typical Chart Recording During Sampling

Figure 6
Arrangement of Analysis Equipment

Ontario Power Generation 9

 Figure 7
Arrangement of Multi-point Injection

Ontario Power Generation

10