- Email: [email protected]
Design and development of a novel water flow measurement system
Accepted Manuscript Design and Development of a Novel Water Flow Measurement System Shiv Kumar Jaiswal, Sanjay Yadav, Ravinder Agarwal PII: DOI: Reference:
S0263-2241(17)30226-9 http://dx.doi.org/10.1016/j.measurement.2017.04.018 MEASUR 4699
To appear in:
Measurement
Received Date: Revised Date: Accepted Date:
31 December 2015 1 April 2017 10 April 2017
Please cite this article as: S.K. Jaiswal, S. Yadav, R. Agarwal, Design and Development of a Novel Water Flow Measurement System, Measurement (2017), doi: http://dx.doi.org/10.1016/j.measurement.2017.04.018
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and Development of a Novel Water Flow Measurement System Shiv Kumar Jaiswal1, Sanjay Yadav1# and Ravinder Agarwal2 1
CSIR-National Physical Laboratory, New Delhi -110012 India 2
#
Thapar University, Patiala -147004, India
Author for correspondence; Email: [email protected]
Abstract
The present paper reports the design and developmental aspects of a prototype of an automatic primary water flow standard. The system is based on a unique technique of carrying out multiple weighing using double weighing tanks. The multiple weighing technique has advantages of increasing the collection time by collecting large water mass thereby reducing the diverter error, the main source of measurement uncertainty, consequently improving measurement uncertainty of the system. The longer collection time minimizes the effect of diverter error. This makes cumbersome process of diverter error evaluation insignificant and may be avoided. Another advantage is to use either of the two tanks in case of any one of the tanks becomes out of order and using another tank as stand alone in a single weighing compromising bit measurement uncertainty. The system thus developed is capable of measuring water flow rate and / calibrating different water flow meters having size (15 to 100) mm, in the flow range from 0.3 m3/h to 100 m3/h. The estimated relative expanded uncertainties of the system for collected water mass of 2000 kg, 1500 kg, 1000 kg and 500 kg are found as 0.055 %, 0.056 %, 0.061 % and 0.081 % (at k=2), respectively which is sufficient for industrial dissemination. The present designed and developed prototype system is a low-cost solution to users and industries which is easier to operate and maintain with comparable existing systems.
Keywords
Uncertainty, Volume flow rate, Water flow measurements, Diverter error, Multiple weighing, Double weighing tanks.
1. Introduction
Water is one of the most important fluids essentially needed for survival of living things. Although, at present, the availability of water may be considered plenty but industrialization and alarming climatic changes,
have urged the researchers to use water effectively,
judicially, adequately to save future generations. In the context, the role of metrologists, manufacturers and associated agencies is very vital to device instrumentation and techniques for the accurate and precise measurement of quantity of water for its optimum use and also for exact quantification, estimating appropriate tariff and curbing misuse. Normally, water flow measurement system (WFMS) based on static weighing method is used worldwide for measurement and calibration of different types of flowmeters [1-2]. In a static gravimetric method, various sources of errors and uncertainties are collected mass, collection time and liquid density. Most of these errors and uncertainties contributions are improved by the use of advanced instrumentation but improvement in the diverter error (part of collection time) is a slightly complex process. It is not possible to improve diverter error only using better instrumentation but special techniques are required to this effect. Several researchers have carried out investigations on improvement of diverter error at various NMIs i.e. National Institute of Standards and Technology (NIST), USA; Physikalisch-Technische Bundesanstalt (PTB), Germany and National Metrology Institute of Japan (NMIJ), Japan [36]. During 2007, PTB, Germany established high accuracy WFMS with an expanded uncertainty of total flow as 0.02 % which is a result of several studies carried out over the years starting from 1998 [7-14]. In this system, a new diverter design having an angular encoding transducer and electric diverter actuator with suitable electronics and software has been used to achieve targeted diverter timing error. The NIST, USA has also worked during the same period on reducing the diverter error and developed a new WFMS based on self canceling diverter (uni-diverter) achieving diverter error less than 0.01 %. The uni-diverter moves in the same direction against liquid jet both at the beginning and at the end of the water collection. Uni-directional travel of the divider reduces the error due to asymmetry in the divider actuated motion, the liquid jet velocity profile and the position of the diverter trigger [15-18]. The WFMS developed by NMIJ, Japan having diverter error as 0.003 % is based on double wing method [19] wherein both the wings of the diverter move in the same direction against the liquid jet at the beginning and end of the measurement. This design has limitation of
adjusting the start position for the next diversion. To address the issue, further, a novel rotating double wing (RDW) type diverter has been developed achieving diverter error <0.002 % which is the best in the world [20]. In India, the responsibility for establishment, maintenance, upgradation and dissemination of national flow standards lies with National Physical Laboratory, New Delhi (NPLI) as the National Metrology Institute (NMI) for India. Consequently, NPLI established water measurement laboratory in joint collaboration with PTB, Germany during 1992-1998 in the flow range of 0.1 m3/h to 600 m3/h using static-gravimetric method as per ISO 4185 [1-2, 2123]. There are two test rigs, namely, DN50 and DN200 in this facility for calibration of water flowmeters of different sizes ranging from 10 mm to 200 mm. This highly complex automated system comprised of lot of electronics and controls, required highly skilled manpower and resources to maintain. Although several researchers have carried out a lot of works as reported above but reducing diverter error is still a challenge. This challenge can be addressed by faster actuation, symmetric triggering and longer collection time. This motivated the authors to develop a simpler, easy to maintain, cost effective system having improved diverter error. In this paper, a novel method based on multiple weighing utilizing two weighing tanks has been designed and developed for achieving longer collection time to improve diverter error for which theoretical aspects have already been reported elsewhere [24]. . 2. Design Features of Water Flow Measurement System
The design of the developed WFMS is based on gravimetric method as per ISO 4185 [1]. This method is also considered as primary method in flow metrology wherein flow is realized in fundamental units of mass, length and time. A block diagram and photograph of the presently designed and developed system are shown in Fig. 1(a) and (b), respectively. The schematic diagram showing operation of the system is depicted in Fig. 2. The system based on a unique multiple weighing technique, is mainly comprises of a flow generation system, test section and a timed weighing system [9]. The detailed description of each section is as follows;
(a)
(b) Fig. 1. Double weighing tank based water flow measurement system, (a) block diagram (b) photograph showing weighing tanks with diverter and fishtail.
Fig. 2. Operating principle of double weighing tank based WFMS.
2.1. Flow generation system The WFMS is a closed loop flow system. A underground storage tank made of a cemented concrete structure [size (25×3×3) m3] is having capacity of storing 225 m3 water. The temperature stability of the storage tank is found to be better than ±1 °C/h for calibrating medium i.e. water. Three constant velocity pumps, each of 37 kW (50 hp), are used. These can be operated individually or in combination as per requirement, to pump water from the storage tank to constant level over head tank (CLOHT) through a customized manifold (DN300). The head and capacity of the each pump are 30 m and 240 m3/h, respectively. Therefore, the total capacity of 3 pumps is 720 m3/h. The CLOHT at 25 meter height, with effective capacity of 21 m3 is in a position to supply water to the test rig up to a maximum flow so that any type of flow meter can be calibrated. The effective capacity and height are adequate to take care of all losses due to various fittings and providing the desired flow rate of 600 m3/h. There exists a weir for maintaining the water level constant. The main features of the CLOHT are maintaining the water level constant, water surface free from ripples, restricting the formation of swirl, highly stable flow and constant head to the test rigs. The diameter of pumping line is 300 mm. There is an overflow line of 200 mm diameter. From the CLOHT, a separate manifold (DN400) is derived which splits the flow into two separate pipelines of 50 mm and 200 mm [3-4]. The prototype system has been developed by deriving/ taping a separate line of 100 mm from the 400 mm manifold. Admittedly, the flow generation system is of the previous water flow measurement system, however, the 100 mm test rig is the new development. A 15 kW air compressor with refrigerated type dryer is used for
pneumatically operated valves. The control panel of the system is now upgraded with latest electronics, instrumentation and controls.
2.2 Test section Test section is comprised of a 100 mm piping system (110D upstream and 30D downstream lengths, where D is diameter of the pipeline) for obtaining optimum and fully developed pipe flow and better flow stability. The Meter under Test (MUT) is installed in the test line. Since CLOHT has been used for better flow stability, the use of flow conditioner is avoided in the line. The adopters of different sizes have been fabricated to install MUT. The diaphragm type 100 mm flow control valve has been used in downstream side for control of flow of the line. The current to pressure converter has been integrated with flow control valve which operates on (4 to 20) mA.
2.3 Timed weighing system The weighing system containing 4 load cells (1000 kg capacity each) are used for the weighing purpose. The maximum limit set on load cell indicator is 2500 kg to accommodate the weight of weighing tank. Although, the capacity of weighing tank is 2500 kg but actual capacity used is up to 2000 kg in the present investigation to avoid any overflow. Two such identical weighing systems have been developed and used. The conical (V) shape design of the weighing tanks at the bottom end results into faster draining which is required in case of multiple weighing method. A plate has been suspended beneath the weighing tank as a platform for keeping the dead weights for calibration of the weighing system. The each weighing tank is connected with 200 mm drain valve for draining the water to sump tank. A simpler fishtail type diverter has been designed and adopted for easier maintenance. The fishtail changes the flow from circular to the rectangular cross-section of the nozzle. The flow nozzle has aspect ratio of 12.5. During calibration, the diverter valve switches the flow to either bypass line or collection tank. The diverter error is the main uncertainty contribution in collection time. For the timing error measurement, the use of a single proximity switch is advantageous for easy triggering of frequency counter. However, in the previous system, 3 proximity switches were used. These switches require automation for triggering. The schematic diagram of the diverter is depicted in Fig. 3. The electronic circuitry for automation of diverter and drain valves have been designed and developed. The functions, operations and data acquisition for load cell indicators, frequency counter/timer, digital multimeter, diverter and drain valves have been made automatic in LabVIEW software. Fig.
4 shows the screen shot of software developed for single weighing. The flow diagram of the system showing data acquisition and control is shown in Fig. 5.
Fig. 3. Schematic diagram of diverter with design parameters.
(a)
(b) Fig. 4. Screen shots of the Automation Software developed for single weighing.
Fig. 5. Data acquisition and control chart of prototype WFMS.
3. The Measurement Principle
In a conventional water flow measurement system, diverter valve is used to direct flow either into the weighing system or into a bypass, which is ultimately returned to sump tank due to being a closed loop system. The measurement of flow is determined by collecting a prescribed mass (m) of steadily flowing water over a measured time interval (t). Since the flow is dynamic, therefore, the output of the flowmeter (i.e. frequency/ current/ other output form) is averaged during collection interval. During the calibration, other quantities (e.g., water temperature; air temperature, humidity and barometric pressure) are measured to determine air density and water density (ρ) for determining air buoyancy correction factor as per NIST formula [16]. The buoyancy correction is applied to the mass indicated by the weighing system. The mass flow rate is calculated by dividing corrected collected mass by the collection time. The volume flow rate is determined by dividing mass flow rate by density of the water. The volume flow rate indicated by MUT ( qvMUT ) is compared against calculated by the standard ( q vSTD ) and error (e) of the MUT is determined from the following equation:
e=
(qvMUT − qvSTD) ×100 qvSTD
The flowmeter error is also determined in terms of standard volume as follows:
(1)
e=
[(qvMUT × t ) − V STD] ×100 VSTD
(2)
Some flowmeters are calibrated in terms of the meter K-factor, which is defined as follows:
meter K - factor =
N
=
V STD
f
(3)
qvSTD
where, N is the no. of pulses output by MUT, f is the frequency output of the MUT, VSTD and q vSTD are the volume and volume flow rate measured by WMFS.
The relative standard uncertainty of flowmeter error is determined by the following equation [3-4, 13, 16]: 2
2
2
2
ue u u u u uρ = Ae + I + m + t + e e I m t ρ
2
(4)
where, ue is the combined standard uncertainty in flowmeter error determination, uAe is the repeatability (Type A uncertainty) in the flowmeter error, uI is the standard uncertainty of digital multimeter in current measurement, um is the standard uncertainty in collected water mass (inclusive of all the components of weighing system), ut is the standard uncertainty in collection time (inclusive of all the components of collection time measurement) and uρ is the standard uncertainty in water density measurement.
The relative standard combined uncertainty of meter K-factor is determined as follows: 2
u K −factor u u = AK −factor + f K − factor K − factor f
2
2 2 u m u t uρ + + + m t ρ
2
(5)
where, uK-factor is the standard combined uncertainty in flowmeter meter K-factor determination, uAK-factor is the repeatability in the meter K-factor measurement and uf is the standard uncertainty of frequency counter in frequency measurement.
4. Results and Discussions
The evaluation of the individual uncertainty contributions as described by earlier researchers [13, 16], is carried to establish the traceability of the system and assign the associated measurement uncertainty. All the instruments used in the present investigations i.e. load cells with weighing tank, universal counter, digital multimeter, barometric pressure indicator, temperature and humidity indicator, PT-100 sensor & indicator were calibrated against appropriate national standards of respective parameters. The evaluation of uncertainty contributions is described herein;
4.1.1 Collected Water Mass
The uncertainty of the collected water mass depends on the balance indication, balance drift, balance calibration, buoyancy correction, leaks and splashes, storage effects and the evaporation. The weighing system was calibrated using 20 weights of 50 kg each up to 1000 kg. Beyond 1000 kg and up to 2000 kg, a calibration technique of equivalent mass of water was used (Fig. 6). The measurement uncertainty estimated with weighing system is found to be 0.2 kg (at k=2) from 500 kg to 2000 kg. The buoyancy correction is applied to the collected mass values and its uncertainty contribution is almost negligible. The drift of the balance/ weighing system is not included in the uncertainty budget but may be considered later on after acquiring sufficient data over the time. The effect of leaks and splashes are assumed negligible as there were no leaks and splashes observed in the system during measurement. The storage and evaporation effect is not studied in the present investigation. However, combined contribution of buoyancy correction uncertainty, storage and evaporation effect is taken as 0.005 % (at k=1) as reported by other researchers [13, 16]
Fig. 6. Calibration of weighing tank using mass standards method.
4.1.2 Collection Time
The main uncertainty contributions for the collection time are timer calibration and diverter error. For the time measurements, a frequency counter/ timer of nano second resolution was used and its relative uncertainty is estimated to be better than 0.0001% (at k=1). The diverter error was evaluated as per ISO 4185 standard [1] using eq.(6);
n n ∆ ∑ ∑ m t q i i i =1 ∆tm = t m 0 . i =1 − 1 n −1 q m0 t m n
(6)
where, ∆tm is the diverter timing error, q0 is average flow rate during the single-burst filling of the weighing tank with tm being the metered time of flow, qn is the average flow rate during the n bursts, m0 is the mass of collected water during single-burst filling of the weighing tank, ∆mi is the incremental change in balance read out in ti time interval during collection in the weighing tank. The time ti shows individual time measurement of n increment. The diverter error thus evaluated is shown in Fig. 7 without proximity switch adjustment. The PC clock was also used as timer along with frequency counter for improvement in the error. The PC clock is also used as timer in multiple weighing technique for improvement of diverter error. The combined use of PC clock and frequency counter has the reduced the diverter error considerably and it is found to be within 0.025% (at k=1) for the flow range 30
% to 100 %. This method is suitable where there is no provision for adjustment of proximity switch. The proximity switch was adjusted for optimizing the diverter error. The diverter error is found to be within 0.02 % (at k=1) in the flow range 10 m3/h to 100 m3/h (Fig. 8).
Fig. 7. Plot of diverter error verses flow rate without adjustment of proximity switch.
Fig. 8. Plot of diverter error verses flow rate after adjustment of proximity switch.
4.1.3 Water Density
The routine municipal water supply water is used in the experiments. The water density is determined using weighing method and its relative uncertainty is estimated as 0.01% (k=2). The uncertainty due to repeatability of MUT and digital multimeter / frequency counter (depending upon current or frequency output) is not included. Accordingly, the overall
system uncertainties at 2000 kg, 1500 kg, 1000 kg and 500 kg collected water mass are estimated as 0.055 %, 0.056 %, 0.061 % and 0.081 % (at k=2), respectively. The uncertainty budget of the new WFMS is summarized in Table 1.
Table 1 Uncertainty budget of new water flow measurement system of size DN100
Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 1500 kg 1000 kg 500 kg 0.005 0.0067 0.01 0.02 0.0058 0.0077 0.0116 0.0231 0.005 0.005 0.005 0.005
0.0001 0.025 0.005 0.0271 0.055
0.0001 0.025 0.005 0.0279 0.056
0.0001 0.025 0.005 0.0302 0.061
0.0001 0.025 0.005 0.0401 0.081
The calibration of an electromagnetic flow meter requires data acquisition and evaluation of current output and pulse/ frequency output using digital multimeter and frequency counter respectively. The estimated relative combined measurement uncertainty including all the uncertainty components was found to be 0.083 % at k=2. The detailed uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h using prototype WFMS is given in Table 2. Table 2 Uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h in current
output mode
Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Digital multimeter calibration Repeatability (Type A uncertainty) of flowmeter Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 0.005 0.0058 0.005
0.0001 0.025 0.005 0.005 0.031 0.0415 0.083
In the WFMS, the installation effect is one of significant uncertainty contributions apart from the instrumentations used to achieve traceability [25]. Therefore, an interlaboratory comparison (ILC) is required to prove the performance of system [26-30]. An interlaboratory comparison (ILC) was carried out with Fluid Control Research Institute (FCRI), India in the flow range (10 to 90) m3/h using an electromagnetic flowmeter of DN80 size. The flowmeter was provided several times more than the minimum upstream and downstream length requirement of the flowmeter. The flowmeter was calibrated in current output mode by both the laboratories. Table 3 and Fig. 9 show the results of interlaboratory comparison. From the Table 3 and Fig. 9, it is observed that the normalized error value, En, is found to be < ±1, which proves the satisfactory performance of the system. Table 3 Interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Sr. No.
Nominal Flow Rate (m3/h)
1. 2. 3. 4. 5.
10 30 50 70 90
NPLI assigned value Average Error (%) -0.792 -0.865 -1.012 -0.783 -0.891
Type A Uncertainty (%) 0.029 0.033 0.025 0.027 0.031
Expanded Uncertainty (%) 0.088 0.098 0.079 0.081 0.092
FCRI assigned value Average Error (%) -0.880 -0.745 -0.952 -0.873 -0.966
Type A Uncertainty (%) 0.010 0.007 0.010 0.014 0.009
Expanded Uncertainty (%) 0.080 0.080 0.080 0.080 0.080
En value
0.74 -0.95 -0.53 0.79 0.62
Fig. 9. Plot of interlaboratory comparison results of electromagnetic flowmeter of DN80 size.
Both the weighing systems are used in case of multiple weighing. The PC clock is used for time measurement. This PC clock was calibrated using universal frequency counter / timer
using comparison method. Fig. 10 shows the screen shot of the software developed for multiple weighing.
(a)
(b) Fig. 10. Screen shots of the Automation Software developed for multiple weighing.
For prototype WFMS, the concreted drain line of old DN50 Test Rig has been used which can handle the maximum flow up to 70 m3/h. Although, higher drain capacity is achieved using 2500 kg weighing tank with 200 mm drain valve but due to limitation of concrete drain line, the diameter of the connecting metallic drain pipe was reduced from 200 mm to about 100 mm. This process increases the draining time from 18 s to 55 s for 2500 kg water. Due to this reason, the multiple weighing was carried out at flow rate of 40 m3/h using a weighing capacity of 1000 kg. A comparative results of error and measurement uncertainty in single
and multiple weighing is given in Table 4 and Fig. 11. The flowmeter was calibrated in a single weighing mode using a universal frequency counter and its error was determined as -0.94 % with expanded uncertainty of ±0.11 %. Here the corrected diverter error was used to show the difference of flowmeter error and uncertainty due to diverter error. It is clearly evident from the data that measurement uncertainty improves considerably in case of multiple weighing (Fig. 11). The effect of diverter error decreases as the collected mass (& collection time) increases. This shows if collected mass and collection time are sufficiently high, then the effect of diverter error is insignificant and it is not required to take into consideration while estimating measurement uncertainty. This enables users to avoid the cumbersome process of evaluation of diverter error.
Table 4. Comparison of single weighing and multiple weighing data where diverter error is
uncorrected and proximity switch is not adjusted.
S. No.
Type (single/ multiple weighing)
1.
Single weighing
Collected Mass
Flowmeter *Expanded Error uncertainty Standard Estimate Standard % at k=2 uncertainty (ρ) uncertainty U (flow) (ut) (uρ) % % kg/m3 %
Collection Time
Water Density
Estimate (M) kg
Standard uncertainty (uM) %
Estimate (t) s
1000
0.02
90
0.0951
994
0.005
-0.845
0.205
2.
2000
0.02
180
0.0476
996
0.005
-0.856
0.121
3.
3000
0.02
270
0.0317
996
0.005
-0.864
0.099
4.
4000
0.02
360
0.0238
996
0.005
-0.858
0.089
5000
0.02
450
0.0190
996
0.005
-0.867
0.085
6000
0.02
540
0.0159
996
0.005
-0.859
0.082
7.
7000
0.02
630
0.0136
996
0.005
-0.888
0.080
8.
8000
0.02
720
0.0119
996
0.005
-0.933
0.079
9.
9000
0.02
810
0.0106
996
0.005
-0.961
0.078
10.
10000
0.02
900
0.0095
996
0.005
-0.986
0.078
5. 6.
Multiple weighing
*Water density uncertainty of 0.01% (at k=2) and Type A uncertainty of 0.03 % have been considered for the expanded uncertainty estimation.
-0.94
(a)
-0.94
(b) Fig. 11. Comparison of single weighing and multiple weighing measurement results.
The developed WFMS is simpler as compared to the WFMSs of NIST, PTB and NMIJ. The automation has been kept to the optimum level. The cheaper instrumentation has been used
in this system. Hence, in overall, our present developed system is simpler, easy to use and low-cost solution to industries.
5. Conclusion
The following conclusions are drawn from the studies; i)
A new Water Flow Primary Standard (size DN100) is designed and developed indigenously.
ii)
The standard thus developed is based on gravimetric method which is suitable to use in flow range 0.3 m3/h to 100 m3/h.
iii)
The main features of the developed system are that it can be used either in single or multiple weighing modes, easy to maintain and low-cost industrial solution. The minimum and optimum automation is used to make it ease in operations.
iv)
The single weighing is a commonly used method but multiple weighing is advantageous in case of high flow where collected mass is limited by weighing tank size.
v)
The relative expanded measurement uncertainties associated with a use of single weighing tank are found to be 0.055 %, 0.056 %, 0.061 % and 0.081 % at 2000 kg, 1500 kg, 1000 kg and 500 kg, respectively. The measurement uncertainties of this order are considered to be excellent for such a primary standard.
vi)
An interlaboratory comparison with FCRI was carried out to check the performance of the developed system and it was found satisfactory.
vii)
The novelty of the system is that it does not require the cumbersome process of diverter error evaluation using multiple weighing. The overall measurement uncertainty in multiple weighing is found to be 0.08 % in comparison to 0.21 % in single weighing which is a significant improvement.
viii)
There is a scope of further improvement if better flow stability and flow velocity profile are achieved using flow conditioner with CLOHT. Using high accuracy weighing system, densitometer and timer will further supplement to the improvement of uncertainty of WFMS.
Acknowledgement
The authors are thankful to Director, CSIR-National Physical Laboratory, New Delhi and Director, Thapar University, Patiala for their constant encouragement. Two of the authors
express their gratitude towards Dr. Ranjana Mehrotra, Head, Physico-Mechanical Metrology Division, CSIR-National Physical Laboratory and Dr. A K Bandyopadhyay for their support and encouragement. Authors are thankful to Council of Scientific and Industrial Research (CSIR) for granting fund to CSIR–National Physical Laboratory under project PSC-0111 for progress of the research work.
References
[1]
ISO 4185:1980, Measurement of liquid flow in closed conduits – Weighing methods, 1st edition, 1980.
[2]
ISO 9368-1:1990, Measurement of liquid flow in closed conduits by weighing methods – Procedure for checking the installations, Part 1: Static weighing system, 1st edition, 1990.
[3]
ISO 5168:2005, Measurement of liquid flow in closed conduits – Procedure for the evaluation of uncertainties, 2nd edition, pp 1-65, 2005.
[4]
JCGM 100:2008, Evaluation of measurement data – Guide to the expression of uncertainty in measurement, 1st edition, 2008.
[5]
NIST Special Publication 250, NIST calibration services for water flowmeters, water flow calibration facility, NIST, Gaithersburg, August 2006.
[6]
S.K. Jaiswal, S. Yadav, A.K. Bandyopadhyay and R. Agarwal, Global water flow measurement and calibration facilities: Review of methods and instrumentations, MAPAN-Journal of Metrology Society of India, 27(2) (2012) 63-76.
[7]
W. Poeschel and R. Engel, The concept of a new primary standard for liquid flow measurement at PTB, Braunschweig, in: Proc. of the 9th International Conference on Flow Measurement (FLOMEKO ‘98), Lund, Sweden, June 15-17, 1998, pp 7-12.
[8]
R. Engel and U. Klages, A novel approach to improve diverter performance in liquid flow calibration facilities, in: Proc. of the 10th International Conference on Flow Measurement (FLOMEKO 2000), Salvador, Brazil, June 5-8, 2000.
[9]
W. Poeschel, R. Engel, D. Dopheide, H. J. Baade, H. J. Kecke, R. Praetor, N. Weist and E. Kurras, A unique fluid diverter design for water flow calibration facilities, in: Proc. of the 10th International Conference on Flow Measurement (FLOMEKO 2000), Salvador, Brazil, June 5-8, 2000.
[10]
R. Engel, H.J. Baade, and A. Rubel, Performance improvement of liquid flow calibrators by applying special measurement and control strategies, in: Proc. of the
11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003. [11]
R. Engel and H.J. Baade, New-design dual-balance gravimetric reference systems with PTB’s new hydrodynamic test field,
in: Proc. of the 11th International
Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003. [12]
R. Engel, PTB’s hydrodynamic test field – investigations to verify the measurement uncertainty, in: Proc. of the 12th International Conference on Flow Measurement (FLOMEKO 2004), Guilin, China, September 14-17, 2004.
[13]
R. Engel, Modeling the uncertainty in liquid flowmeters calibration and application – requirements and their technical realization for PTB’s national water flow standard, in: Proc. of the SENSOR Conference 2007, Nurnburg, Germany, May 22-24, 2007.
[14]
R. Engel and H.J. Baade, Model-based flow diverter analysis for an improved uncertainty determination in liquid flow calibration facilities, Measurement Science and Technology, vol. 21 (2010) 1-11.
[15]
T.T. Yeh, N.P. Yende and P.I. Espina, Theoretical self-error cancelling diverters for liquid flow calibration facilities, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[16]
V. Gowda, T.T. Yeh and P.I. Espina, The new NIST water flow calibration facility, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[17]
N.P. Yende, T.T. Yeh and P.I. Espina, Performance of an error free liquid flow diverter design, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[18]
I. Marfenko, T.T. Yeh and J. Wright, Diverter uncertainty less than 0.01% for water flow calibrations, in: Proc. of the 6th International Symposium for Fluid Flow, Querétario, Mexico, May 16-18, 2006.
[19]
T. Shimada, S. Oda, Y. Terao and M. Takamoto, Development of a new diverter system for liquid flow calibration facilities, Flow Measurement and Instrumentation, vol. 14 (2003) pp 89–96.
[20]
R. Doihara, T. Shimada, Y. Terao and M. Takamoto, Development of weighing tank system employing rotating double wing diverter, Flow Measurement and Instrumentation, vol. 17 (2006) pp 141–152.
[21]
J.N. Som, R. Singh, V. Babu and I.S. Taak, Environmental conditions required for Fluid Flow Measurement Laboratory with special reference to Water Flow Measurement set-up at NPL, MAPAN-Journal of Metrology Society of India, 13 (1998) 3-8.
[22]
J.N. Som, R. Singh, V. Babu and I.S. Taak, Study on the error analysis and uncertainty parameters involved on the flow rate measurement with the NPL primary water flow measurement facility, in: Proc. of the 4th International Conference on Advances in Metrology (AdMet-2003), Delhi, Feburary 5-7, 2003.
[23]
S.K. Jaiswal, I.S. Taak, D. Sharma, C. Singh and A.K. Bandyopadhyay, The fluid flow measurement standard at National Physical Laboratory, New Delhi, in: Proc. of NCSLI 2010 Workshp and Symposium, Providence, USA, July 25-29, 2010.
[24]
S.K. Jaiswal, S. Yadav, and R. Agarwal, Multiple weighing based method for realizing flow, MAPAN – Journal of Metrology Society of India, 30 (2) (2015) 119123.
[25]
D.W. Spitzer, Flow Measurement: Practical Guides for Measurement and Control, 2nd Edition, Research Triangle Park, NC: ISA, 2001.
[26]
NABL 162, Guidelines for proficiency testing program for testing and calibration laboratories, issue no. 4, 2008.
[27]
S. Yadav, V.K. Gupta and A.K. Bandyopadhyay, Standardisation of pressure calibration (7-70 MPa) using digital pressure calibrator, Journal of Scientific and Industrial Research, 69 (2010) 27-33.
[28]
S. Yadav, V.K. Gupta and A.K. Bandyopadhyay, Standardisation of pressure measurement using pressure balance as transfer standard. MAPAN–Journal of Metrology Society of India, 26 (2011) 133-151.
[29]
ISO / IEC 17043, Conformity assessment - General requirements for proficiency testing, 2nd edition, 2010.
[30]
NABL 164, Guidelines for inter-laboratory comparison for calibration laboratories where formal PT programs are not available, issue no. 2, 2012.
List of Figures
Fig.1. Double weighing tank based water flow measurement system, (a) block diagram (b) photograph showing weighing tanks with diverter and fishtail Fig. 2. Operating principle of double weighing tank based WFMS. Fig. 3. Schematic diagram of diverter with design parameters. Fig. 4. Screen shots of the Automation Software developed for single weighing. Fig. 5. Data acquisition and control chart of prototype WFMS. Fig. 6. Calibration of weighing tank using mass standards method. Fig. 7. Plot of diverter error verses flow rate without adjustment of proximity switch. Fig. 8. Plot of diverter error verses flow rate after adjustment of proximity switch. Fig. 9. Plot of interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Fig. 10. Screen shots of the Automation Software developed for multiple weighing. Fig. 11. Comparison of single weighing and multiple weighing measurement results.
Figure 1(a)
Figure 1(b)
Figure 2
Figure 3
Figure 4(a)
Figure 4(b)
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10(a)
Figure 10(b)
-0.94
Figure 11 (a)
-0.94
Figure 11 (b)
List of Tables Table 1 Uncertainty budget of new water flow measurement system of size DN100 Table 2 Uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h in current
output mode Table 3 Interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Table 4 Comparison of single weighing and multiple weighing data where diverter error is
uncorrected and proximity switch is not adjusted.
Table 1 Collected water mass
Sources of uncertainty
[Relative standard uncertainty (%)]
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
2000 kg 0.005 0.0058 0.005
1500 kg 0.0067 0.0077 0.005
1000 kg 0.01 0.0116 0.005
500 kg 0.02 0.0231 0.005
0.0001 0.025 0.005 0.0271 0.055
0.0001 0.025 0.005 0.0279 0.056
0.0001 0.025 0.005 0.0302 0.061
0.0001 0.025 0.005 0.0401 0.081
Table 2 Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Digital multimeter calibration Repeatability (Type A uncertainty) of flowmeter Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 0.005 0.0058 0.005 0.0001 0.025 0.005 0.005 0.031 0.0415 0.083
Table 3 Sr. No.
Nominal Flow Rate (m3/h)
1. 2. 3. 4. 5.
10 30 50 70 90
NPLI assigned value Average Error (%) -0.792 -0.865 -1.012 -0.783 -0.891
Type A Uncertainty (%) 0.029 0.033 0.025 0.027 0.031
Expanded Uncertainty (%) 0.088 0.098 0.079 0.081 0.092
FCRI assigned value Average Error (%) -0.880 -0.745 -0.952 -0.873 -0.966
Type A Uncertainty (%) 0.010 0.007 0.010 0.014 0.009
Expanded Uncertainty (%) 0.080 0.080 0.080 0.080 0.080
En value
0.74 -0.95 -0.53 0.79 0.62
Table 4
S. No.
Type (single/ multiple weighing)
1.
Single weighing
Collected Mass
Flowmeter *Expanded Error uncertainty Standard Estimate Standard % at k=2 uncertainty (ρ) uncertainty U (flow) (ut) (uρ) % kg/m3 %
Collection Time
Water Density
Estimate (M) kg
Standard uncertainty (uM) %
Estimate (t) s
1000
0.02
90
0.0951
994
0.005
-0.845
0.205
2.
2000
0.02
180
0.0476
996
0.005
-0.856
0.121
3.
3000
0.02
270
0.0317
996
0.005
-0.864
0.099
4.
4000
0.02
360
0.0238
996
0.005
-0.858
0.089
5000
0.02
450
0.0190
996
0.005
-0.867
0.085
6000
0.02
540
0.0159
996
0.005
-0.859
0.082
7.
7000
0.02
630
0.0136
996
0.005
-0.888
0.080
8.
8000
0.02
720
0.0119
996
0.005
-0.933
0.079
9.
9000
0.02
810
0.0106
996
0.005
-0.961
0.078
10.
10000
0.02
900
0.0095
996
0.005
-0.986
0.078
5. 6.
Multiple weighing
*Water density uncertainty of 0.01% (at k=2) and Type A uncertainty of 0.03 % have been considered for the expanded uncertainty estimation.
Highlights
•
An economical water flow measurement system (WFMS) is designed and developed.
•
Weighing tanks of conical design for faster draining are presented.
•
It does not require cumbersome process of diverter error evaluation.
•
Uncertainty of measurement is improved significantly in case of multiple weighing.
Graphical abstract
S0263-2241(17)30226-9 http://dx.doi.org/10.1016/j.measurement.2017.04.018 MEASUR 4699
To appear in:
Measurement
Received Date: Revised Date: Accepted Date:
31 December 2015 1 April 2017 10 April 2017
Please cite this article as: S.K. Jaiswal, S. Yadav, R. Agarwal, Design and Development of a Novel Water Flow Measurement System, Measurement (2017), doi: http://dx.doi.org/10.1016/j.measurement.2017.04.018
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Design and Development of a Novel Water Flow Measurement System Shiv Kumar Jaiswal1, Sanjay Yadav1# and Ravinder Agarwal2 1
CSIR-National Physical Laboratory, New Delhi -110012 India 2
#
Thapar University, Patiala -147004, India
Author for correspondence; Email: [email protected]
Abstract
The present paper reports the design and developmental aspects of a prototype of an automatic primary water flow standard. The system is based on a unique technique of carrying out multiple weighing using double weighing tanks. The multiple weighing technique has advantages of increasing the collection time by collecting large water mass thereby reducing the diverter error, the main source of measurement uncertainty, consequently improving measurement uncertainty of the system. The longer collection time minimizes the effect of diverter error. This makes cumbersome process of diverter error evaluation insignificant and may be avoided. Another advantage is to use either of the two tanks in case of any one of the tanks becomes out of order and using another tank as stand alone in a single weighing compromising bit measurement uncertainty. The system thus developed is capable of measuring water flow rate and / calibrating different water flow meters having size (15 to 100) mm, in the flow range from 0.3 m3/h to 100 m3/h. The estimated relative expanded uncertainties of the system for collected water mass of 2000 kg, 1500 kg, 1000 kg and 500 kg are found as 0.055 %, 0.056 %, 0.061 % and 0.081 % (at k=2), respectively which is sufficient for industrial dissemination. The present designed and developed prototype system is a low-cost solution to users and industries which is easier to operate and maintain with comparable existing systems.
Keywords
Uncertainty, Volume flow rate, Water flow measurements, Diverter error, Multiple weighing, Double weighing tanks.
1. Introduction
Water is one of the most important fluids essentially needed for survival of living things. Although, at present, the availability of water may be considered plenty but industrialization and alarming climatic changes,
have urged the researchers to use water effectively,
judicially, adequately to save future generations. In the context, the role of metrologists, manufacturers and associated agencies is very vital to device instrumentation and techniques for the accurate and precise measurement of quantity of water for its optimum use and also for exact quantification, estimating appropriate tariff and curbing misuse. Normally, water flow measurement system (WFMS) based on static weighing method is used worldwide for measurement and calibration of different types of flowmeters [1-2]. In a static gravimetric method, various sources of errors and uncertainties are collected mass, collection time and liquid density. Most of these errors and uncertainties contributions are improved by the use of advanced instrumentation but improvement in the diverter error (part of collection time) is a slightly complex process. It is not possible to improve diverter error only using better instrumentation but special techniques are required to this effect. Several researchers have carried out investigations on improvement of diverter error at various NMIs i.e. National Institute of Standards and Technology (NIST), USA; Physikalisch-Technische Bundesanstalt (PTB), Germany and National Metrology Institute of Japan (NMIJ), Japan [36]. During 2007, PTB, Germany established high accuracy WFMS with an expanded uncertainty of total flow as 0.02 % which is a result of several studies carried out over the years starting from 1998 [7-14]. In this system, a new diverter design having an angular encoding transducer and electric diverter actuator with suitable electronics and software has been used to achieve targeted diverter timing error. The NIST, USA has also worked during the same period on reducing the diverter error and developed a new WFMS based on self canceling diverter (uni-diverter) achieving diverter error less than 0.01 %. The uni-diverter moves in the same direction against liquid jet both at the beginning and at the end of the water collection. Uni-directional travel of the divider reduces the error due to asymmetry in the divider actuated motion, the liquid jet velocity profile and the position of the diverter trigger [15-18]. The WFMS developed by NMIJ, Japan having diverter error as 0.003 % is based on double wing method [19] wherein both the wings of the diverter move in the same direction against the liquid jet at the beginning and end of the measurement. This design has limitation of
adjusting the start position for the next diversion. To address the issue, further, a novel rotating double wing (RDW) type diverter has been developed achieving diverter error <0.002 % which is the best in the world [20]. In India, the responsibility for establishment, maintenance, upgradation and dissemination of national flow standards lies with National Physical Laboratory, New Delhi (NPLI) as the National Metrology Institute (NMI) for India. Consequently, NPLI established water measurement laboratory in joint collaboration with PTB, Germany during 1992-1998 in the flow range of 0.1 m3/h to 600 m3/h using static-gravimetric method as per ISO 4185 [1-2, 2123]. There are two test rigs, namely, DN50 and DN200 in this facility for calibration of water flowmeters of different sizes ranging from 10 mm to 200 mm. This highly complex automated system comprised of lot of electronics and controls, required highly skilled manpower and resources to maintain. Although several researchers have carried out a lot of works as reported above but reducing diverter error is still a challenge. This challenge can be addressed by faster actuation, symmetric triggering and longer collection time. This motivated the authors to develop a simpler, easy to maintain, cost effective system having improved diverter error. In this paper, a novel method based on multiple weighing utilizing two weighing tanks has been designed and developed for achieving longer collection time to improve diverter error for which theoretical aspects have already been reported elsewhere [24]. . 2. Design Features of Water Flow Measurement System
The design of the developed WFMS is based on gravimetric method as per ISO 4185 [1]. This method is also considered as primary method in flow metrology wherein flow is realized in fundamental units of mass, length and time. A block diagram and photograph of the presently designed and developed system are shown in Fig. 1(a) and (b), respectively. The schematic diagram showing operation of the system is depicted in Fig. 2. The system based on a unique multiple weighing technique, is mainly comprises of a flow generation system, test section and a timed weighing system [9]. The detailed description of each section is as follows;
(a)
(b) Fig. 1. Double weighing tank based water flow measurement system, (a) block diagram (b) photograph showing weighing tanks with diverter and fishtail.
Fig. 2. Operating principle of double weighing tank based WFMS.
2.1. Flow generation system The WFMS is a closed loop flow system. A underground storage tank made of a cemented concrete structure [size (25×3×3) m3] is having capacity of storing 225 m3 water. The temperature stability of the storage tank is found to be better than ±1 °C/h for calibrating medium i.e. water. Three constant velocity pumps, each of 37 kW (50 hp), are used. These can be operated individually or in combination as per requirement, to pump water from the storage tank to constant level over head tank (CLOHT) through a customized manifold (DN300). The head and capacity of the each pump are 30 m and 240 m3/h, respectively. Therefore, the total capacity of 3 pumps is 720 m3/h. The CLOHT at 25 meter height, with effective capacity of 21 m3 is in a position to supply water to the test rig up to a maximum flow so that any type of flow meter can be calibrated. The effective capacity and height are adequate to take care of all losses due to various fittings and providing the desired flow rate of 600 m3/h. There exists a weir for maintaining the water level constant. The main features of the CLOHT are maintaining the water level constant, water surface free from ripples, restricting the formation of swirl, highly stable flow and constant head to the test rigs. The diameter of pumping line is 300 mm. There is an overflow line of 200 mm diameter. From the CLOHT, a separate manifold (DN400) is derived which splits the flow into two separate pipelines of 50 mm and 200 mm [3-4]. The prototype system has been developed by deriving/ taping a separate line of 100 mm from the 400 mm manifold. Admittedly, the flow generation system is of the previous water flow measurement system, however, the 100 mm test rig is the new development. A 15 kW air compressor with refrigerated type dryer is used for
pneumatically operated valves. The control panel of the system is now upgraded with latest electronics, instrumentation and controls.
2.2 Test section Test section is comprised of a 100 mm piping system (110D upstream and 30D downstream lengths, where D is diameter of the pipeline) for obtaining optimum and fully developed pipe flow and better flow stability. The Meter under Test (MUT) is installed in the test line. Since CLOHT has been used for better flow stability, the use of flow conditioner is avoided in the line. The adopters of different sizes have been fabricated to install MUT. The diaphragm type 100 mm flow control valve has been used in downstream side for control of flow of the line. The current to pressure converter has been integrated with flow control valve which operates on (4 to 20) mA.
2.3 Timed weighing system The weighing system containing 4 load cells (1000 kg capacity each) are used for the weighing purpose. The maximum limit set on load cell indicator is 2500 kg to accommodate the weight of weighing tank. Although, the capacity of weighing tank is 2500 kg but actual capacity used is up to 2000 kg in the present investigation to avoid any overflow. Two such identical weighing systems have been developed and used. The conical (V) shape design of the weighing tanks at the bottom end results into faster draining which is required in case of multiple weighing method. A plate has been suspended beneath the weighing tank as a platform for keeping the dead weights for calibration of the weighing system. The each weighing tank is connected with 200 mm drain valve for draining the water to sump tank. A simpler fishtail type diverter has been designed and adopted for easier maintenance. The fishtail changes the flow from circular to the rectangular cross-section of the nozzle. The flow nozzle has aspect ratio of 12.5. During calibration, the diverter valve switches the flow to either bypass line or collection tank. The diverter error is the main uncertainty contribution in collection time. For the timing error measurement, the use of a single proximity switch is advantageous for easy triggering of frequency counter. However, in the previous system, 3 proximity switches were used. These switches require automation for triggering. The schematic diagram of the diverter is depicted in Fig. 3. The electronic circuitry for automation of diverter and drain valves have been designed and developed. The functions, operations and data acquisition for load cell indicators, frequency counter/timer, digital multimeter, diverter and drain valves have been made automatic in LabVIEW software. Fig.
4 shows the screen shot of software developed for single weighing. The flow diagram of the system showing data acquisition and control is shown in Fig. 5.
Fig. 3. Schematic diagram of diverter with design parameters.
(a)
(b) Fig. 4. Screen shots of the Automation Software developed for single weighing.
Fig. 5. Data acquisition and control chart of prototype WFMS.
3. The Measurement Principle
In a conventional water flow measurement system, diverter valve is used to direct flow either into the weighing system or into a bypass, which is ultimately returned to sump tank due to being a closed loop system. The measurement of flow is determined by collecting a prescribed mass (m) of steadily flowing water over a measured time interval (t). Since the flow is dynamic, therefore, the output of the flowmeter (i.e. frequency/ current/ other output form) is averaged during collection interval. During the calibration, other quantities (e.g., water temperature; air temperature, humidity and barometric pressure) are measured to determine air density and water density (ρ) for determining air buoyancy correction factor as per NIST formula [16]. The buoyancy correction is applied to the mass indicated by the weighing system. The mass flow rate is calculated by dividing corrected collected mass by the collection time. The volume flow rate is determined by dividing mass flow rate by density of the water. The volume flow rate indicated by MUT ( qvMUT ) is compared against calculated by the standard ( q vSTD ) and error (e) of the MUT is determined from the following equation:
e=
(qvMUT − qvSTD) ×100 qvSTD
The flowmeter error is also determined in terms of standard volume as follows:
(1)
e=
[(qvMUT × t ) − V STD] ×100 VSTD
(2)
Some flowmeters are calibrated in terms of the meter K-factor, which is defined as follows:
meter K - factor =
N
=
V STD
f
(3)
qvSTD
where, N is the no. of pulses output by MUT, f is the frequency output of the MUT, VSTD and q vSTD are the volume and volume flow rate measured by WMFS.
The relative standard uncertainty of flowmeter error is determined by the following equation [3-4, 13, 16]: 2
2
2
2
ue u u u u uρ = Ae + I + m + t + e e I m t ρ
2
(4)
where, ue is the combined standard uncertainty in flowmeter error determination, uAe is the repeatability (Type A uncertainty) in the flowmeter error, uI is the standard uncertainty of digital multimeter in current measurement, um is the standard uncertainty in collected water mass (inclusive of all the components of weighing system), ut is the standard uncertainty in collection time (inclusive of all the components of collection time measurement) and uρ is the standard uncertainty in water density measurement.
The relative standard combined uncertainty of meter K-factor is determined as follows: 2
u K −factor u u = AK −factor + f K − factor K − factor f
2
2 2 u m u t uρ + + + m t ρ
2
(5)
where, uK-factor is the standard combined uncertainty in flowmeter meter K-factor determination, uAK-factor is the repeatability in the meter K-factor measurement and uf is the standard uncertainty of frequency counter in frequency measurement.
4. Results and Discussions
The evaluation of the individual uncertainty contributions as described by earlier researchers [13, 16], is carried to establish the traceability of the system and assign the associated measurement uncertainty. All the instruments used in the present investigations i.e. load cells with weighing tank, universal counter, digital multimeter, barometric pressure indicator, temperature and humidity indicator, PT-100 sensor & indicator were calibrated against appropriate national standards of respective parameters. The evaluation of uncertainty contributions is described herein;
4.1.1 Collected Water Mass
The uncertainty of the collected water mass depends on the balance indication, balance drift, balance calibration, buoyancy correction, leaks and splashes, storage effects and the evaporation. The weighing system was calibrated using 20 weights of 50 kg each up to 1000 kg. Beyond 1000 kg and up to 2000 kg, a calibration technique of equivalent mass of water was used (Fig. 6). The measurement uncertainty estimated with weighing system is found to be 0.2 kg (at k=2) from 500 kg to 2000 kg. The buoyancy correction is applied to the collected mass values and its uncertainty contribution is almost negligible. The drift of the balance/ weighing system is not included in the uncertainty budget but may be considered later on after acquiring sufficient data over the time. The effect of leaks and splashes are assumed negligible as there were no leaks and splashes observed in the system during measurement. The storage and evaporation effect is not studied in the present investigation. However, combined contribution of buoyancy correction uncertainty, storage and evaporation effect is taken as 0.005 % (at k=1) as reported by other researchers [13, 16]
Fig. 6. Calibration of weighing tank using mass standards method.
4.1.2 Collection Time
The main uncertainty contributions for the collection time are timer calibration and diverter error. For the time measurements, a frequency counter/ timer of nano second resolution was used and its relative uncertainty is estimated to be better than 0.0001% (at k=1). The diverter error was evaluated as per ISO 4185 standard [1] using eq.(6);
n n ∆ ∑ ∑ m t q i i i =1 ∆tm = t m 0 . i =1 − 1 n −1 q m0 t m n
(6)
where, ∆tm is the diverter timing error, q0 is average flow rate during the single-burst filling of the weighing tank with tm being the metered time of flow, qn is the average flow rate during the n bursts, m0 is the mass of collected water during single-burst filling of the weighing tank, ∆mi is the incremental change in balance read out in ti time interval during collection in the weighing tank. The time ti shows individual time measurement of n increment. The diverter error thus evaluated is shown in Fig. 7 without proximity switch adjustment. The PC clock was also used as timer along with frequency counter for improvement in the error. The PC clock is also used as timer in multiple weighing technique for improvement of diverter error. The combined use of PC clock and frequency counter has the reduced the diverter error considerably and it is found to be within 0.025% (at k=1) for the flow range 30
% to 100 %. This method is suitable where there is no provision for adjustment of proximity switch. The proximity switch was adjusted for optimizing the diverter error. The diverter error is found to be within 0.02 % (at k=1) in the flow range 10 m3/h to 100 m3/h (Fig. 8).
Fig. 7. Plot of diverter error verses flow rate without adjustment of proximity switch.
Fig. 8. Plot of diverter error verses flow rate after adjustment of proximity switch.
4.1.3 Water Density
The routine municipal water supply water is used in the experiments. The water density is determined using weighing method and its relative uncertainty is estimated as 0.01% (k=2). The uncertainty due to repeatability of MUT and digital multimeter / frequency counter (depending upon current or frequency output) is not included. Accordingly, the overall
system uncertainties at 2000 kg, 1500 kg, 1000 kg and 500 kg collected water mass are estimated as 0.055 %, 0.056 %, 0.061 % and 0.081 % (at k=2), respectively. The uncertainty budget of the new WFMS is summarized in Table 1.
Table 1 Uncertainty budget of new water flow measurement system of size DN100
Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 1500 kg 1000 kg 500 kg 0.005 0.0067 0.01 0.02 0.0058 0.0077 0.0116 0.0231 0.005 0.005 0.005 0.005
0.0001 0.025 0.005 0.0271 0.055
0.0001 0.025 0.005 0.0279 0.056
0.0001 0.025 0.005 0.0302 0.061
0.0001 0.025 0.005 0.0401 0.081
The calibration of an electromagnetic flow meter requires data acquisition and evaluation of current output and pulse/ frequency output using digital multimeter and frequency counter respectively. The estimated relative combined measurement uncertainty including all the uncertainty components was found to be 0.083 % at k=2. The detailed uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h using prototype WFMS is given in Table 2. Table 2 Uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h in current
output mode
Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Digital multimeter calibration Repeatability (Type A uncertainty) of flowmeter Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 0.005 0.0058 0.005
0.0001 0.025 0.005 0.005 0.031 0.0415 0.083
In the WFMS, the installation effect is one of significant uncertainty contributions apart from the instrumentations used to achieve traceability [25]. Therefore, an interlaboratory comparison (ILC) is required to prove the performance of system [26-30]. An interlaboratory comparison (ILC) was carried out with Fluid Control Research Institute (FCRI), India in the flow range (10 to 90) m3/h using an electromagnetic flowmeter of DN80 size. The flowmeter was provided several times more than the minimum upstream and downstream length requirement of the flowmeter. The flowmeter was calibrated in current output mode by both the laboratories. Table 3 and Fig. 9 show the results of interlaboratory comparison. From the Table 3 and Fig. 9, it is observed that the normalized error value, En, is found to be < ±1, which proves the satisfactory performance of the system. Table 3 Interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Sr. No.
Nominal Flow Rate (m3/h)
1. 2. 3. 4. 5.
10 30 50 70 90
NPLI assigned value Average Error (%) -0.792 -0.865 -1.012 -0.783 -0.891
Type A Uncertainty (%) 0.029 0.033 0.025 0.027 0.031
Expanded Uncertainty (%) 0.088 0.098 0.079 0.081 0.092
FCRI assigned value Average Error (%) -0.880 -0.745 -0.952 -0.873 -0.966
Type A Uncertainty (%) 0.010 0.007 0.010 0.014 0.009
Expanded Uncertainty (%) 0.080 0.080 0.080 0.080 0.080
En value
0.74 -0.95 -0.53 0.79 0.62
Fig. 9. Plot of interlaboratory comparison results of electromagnetic flowmeter of DN80 size.
Both the weighing systems are used in case of multiple weighing. The PC clock is used for time measurement. This PC clock was calibrated using universal frequency counter / timer
using comparison method. Fig. 10 shows the screen shot of the software developed for multiple weighing.
(a)
(b) Fig. 10. Screen shots of the Automation Software developed for multiple weighing.
For prototype WFMS, the concreted drain line of old DN50 Test Rig has been used which can handle the maximum flow up to 70 m3/h. Although, higher drain capacity is achieved using 2500 kg weighing tank with 200 mm drain valve but due to limitation of concrete drain line, the diameter of the connecting metallic drain pipe was reduced from 200 mm to about 100 mm. This process increases the draining time from 18 s to 55 s for 2500 kg water. Due to this reason, the multiple weighing was carried out at flow rate of 40 m3/h using a weighing capacity of 1000 kg. A comparative results of error and measurement uncertainty in single
and multiple weighing is given in Table 4 and Fig. 11. The flowmeter was calibrated in a single weighing mode using a universal frequency counter and its error was determined as -0.94 % with expanded uncertainty of ±0.11 %. Here the corrected diverter error was used to show the difference of flowmeter error and uncertainty due to diverter error. It is clearly evident from the data that measurement uncertainty improves considerably in case of multiple weighing (Fig. 11). The effect of diverter error decreases as the collected mass (& collection time) increases. This shows if collected mass and collection time are sufficiently high, then the effect of diverter error is insignificant and it is not required to take into consideration while estimating measurement uncertainty. This enables users to avoid the cumbersome process of evaluation of diverter error.
Table 4. Comparison of single weighing and multiple weighing data where diverter error is
uncorrected and proximity switch is not adjusted.
S. No.
Type (single/ multiple weighing)
1.
Single weighing
Collected Mass
Flowmeter *Expanded Error uncertainty Standard Estimate Standard % at k=2 uncertainty (ρ) uncertainty U (flow) (ut) (uρ) % % kg/m3 %
Collection Time
Water Density
Estimate (M) kg
Standard uncertainty (uM) %
Estimate (t) s
1000
0.02
90
0.0951
994
0.005
-0.845
0.205
2.
2000
0.02
180
0.0476
996
0.005
-0.856
0.121
3.
3000
0.02
270
0.0317
996
0.005
-0.864
0.099
4.
4000
0.02
360
0.0238
996
0.005
-0.858
0.089
5000
0.02
450
0.0190
996
0.005
-0.867
0.085
6000
0.02
540
0.0159
996
0.005
-0.859
0.082
7.
7000
0.02
630
0.0136
996
0.005
-0.888
0.080
8.
8000
0.02
720
0.0119
996
0.005
-0.933
0.079
9.
9000
0.02
810
0.0106
996
0.005
-0.961
0.078
10.
10000
0.02
900
0.0095
996
0.005
-0.986
0.078
5. 6.
Multiple weighing
*Water density uncertainty of 0.01% (at k=2) and Type A uncertainty of 0.03 % have been considered for the expanded uncertainty estimation.
-0.94
(a)
-0.94
(b) Fig. 11. Comparison of single weighing and multiple weighing measurement results.
The developed WFMS is simpler as compared to the WFMSs of NIST, PTB and NMIJ. The automation has been kept to the optimum level. The cheaper instrumentation has been used
in this system. Hence, in overall, our present developed system is simpler, easy to use and low-cost solution to industries.
5. Conclusion
The following conclusions are drawn from the studies; i)
A new Water Flow Primary Standard (size DN100) is designed and developed indigenously.
ii)
The standard thus developed is based on gravimetric method which is suitable to use in flow range 0.3 m3/h to 100 m3/h.
iii)
The main features of the developed system are that it can be used either in single or multiple weighing modes, easy to maintain and low-cost industrial solution. The minimum and optimum automation is used to make it ease in operations.
iv)
The single weighing is a commonly used method but multiple weighing is advantageous in case of high flow where collected mass is limited by weighing tank size.
v)
The relative expanded measurement uncertainties associated with a use of single weighing tank are found to be 0.055 %, 0.056 %, 0.061 % and 0.081 % at 2000 kg, 1500 kg, 1000 kg and 500 kg, respectively. The measurement uncertainties of this order are considered to be excellent for such a primary standard.
vi)
An interlaboratory comparison with FCRI was carried out to check the performance of the developed system and it was found satisfactory.
vii)
The novelty of the system is that it does not require the cumbersome process of diverter error evaluation using multiple weighing. The overall measurement uncertainty in multiple weighing is found to be 0.08 % in comparison to 0.21 % in single weighing which is a significant improvement.
viii)
There is a scope of further improvement if better flow stability and flow velocity profile are achieved using flow conditioner with CLOHT. Using high accuracy weighing system, densitometer and timer will further supplement to the improvement of uncertainty of WFMS.
Acknowledgement
The authors are thankful to Director, CSIR-National Physical Laboratory, New Delhi and Director, Thapar University, Patiala for their constant encouragement. Two of the authors
express their gratitude towards Dr. Ranjana Mehrotra, Head, Physico-Mechanical Metrology Division, CSIR-National Physical Laboratory and Dr. A K Bandyopadhyay for their support and encouragement. Authors are thankful to Council of Scientific and Industrial Research (CSIR) for granting fund to CSIR–National Physical Laboratory under project PSC-0111 for progress of the research work.
References
[1]
ISO 4185:1980, Measurement of liquid flow in closed conduits – Weighing methods, 1st edition, 1980.
[2]
ISO 9368-1:1990, Measurement of liquid flow in closed conduits by weighing methods – Procedure for checking the installations, Part 1: Static weighing system, 1st edition, 1990.
[3]
ISO 5168:2005, Measurement of liquid flow in closed conduits – Procedure for the evaluation of uncertainties, 2nd edition, pp 1-65, 2005.
[4]
JCGM 100:2008, Evaluation of measurement data – Guide to the expression of uncertainty in measurement, 1st edition, 2008.
[5]
NIST Special Publication 250, NIST calibration services for water flowmeters, water flow calibration facility, NIST, Gaithersburg, August 2006.
[6]
S.K. Jaiswal, S. Yadav, A.K. Bandyopadhyay and R. Agarwal, Global water flow measurement and calibration facilities: Review of methods and instrumentations, MAPAN-Journal of Metrology Society of India, 27(2) (2012) 63-76.
[7]
W. Poeschel and R. Engel, The concept of a new primary standard for liquid flow measurement at PTB, Braunschweig, in: Proc. of the 9th International Conference on Flow Measurement (FLOMEKO ‘98), Lund, Sweden, June 15-17, 1998, pp 7-12.
[8]
R. Engel and U. Klages, A novel approach to improve diverter performance in liquid flow calibration facilities, in: Proc. of the 10th International Conference on Flow Measurement (FLOMEKO 2000), Salvador, Brazil, June 5-8, 2000.
[9]
W. Poeschel, R. Engel, D. Dopheide, H. J. Baade, H. J. Kecke, R. Praetor, N. Weist and E. Kurras, A unique fluid diverter design for water flow calibration facilities, in: Proc. of the 10th International Conference on Flow Measurement (FLOMEKO 2000), Salvador, Brazil, June 5-8, 2000.
[10]
R. Engel, H.J. Baade, and A. Rubel, Performance improvement of liquid flow calibrators by applying special measurement and control strategies, in: Proc. of the
11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003. [11]
R. Engel and H.J. Baade, New-design dual-balance gravimetric reference systems with PTB’s new hydrodynamic test field,
in: Proc. of the 11th International
Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003. [12]
R. Engel, PTB’s hydrodynamic test field – investigations to verify the measurement uncertainty, in: Proc. of the 12th International Conference on Flow Measurement (FLOMEKO 2004), Guilin, China, September 14-17, 2004.
[13]
R. Engel, Modeling the uncertainty in liquid flowmeters calibration and application – requirements and their technical realization for PTB’s national water flow standard, in: Proc. of the SENSOR Conference 2007, Nurnburg, Germany, May 22-24, 2007.
[14]
R. Engel and H.J. Baade, Model-based flow diverter analysis for an improved uncertainty determination in liquid flow calibration facilities, Measurement Science and Technology, vol. 21 (2010) 1-11.
[15]
T.T. Yeh, N.P. Yende and P.I. Espina, Theoretical self-error cancelling diverters for liquid flow calibration facilities, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[16]
V. Gowda, T.T. Yeh and P.I. Espina, The new NIST water flow calibration facility, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[17]
N.P. Yende, T.T. Yeh and P.I. Espina, Performance of an error free liquid flow diverter design, in: Proc. of the 11th International Conference on Flow Measurement (FLOMEKO 2003), Groningen, Netherlands, May 12-14, 2003.
[18]
I. Marfenko, T.T. Yeh and J. Wright, Diverter uncertainty less than 0.01% for water flow calibrations, in: Proc. of the 6th International Symposium for Fluid Flow, Querétario, Mexico, May 16-18, 2006.
[19]
T. Shimada, S. Oda, Y. Terao and M. Takamoto, Development of a new diverter system for liquid flow calibration facilities, Flow Measurement and Instrumentation, vol. 14 (2003) pp 89–96.
[20]
R. Doihara, T. Shimada, Y. Terao and M. Takamoto, Development of weighing tank system employing rotating double wing diverter, Flow Measurement and Instrumentation, vol. 17 (2006) pp 141–152.
[21]
J.N. Som, R. Singh, V. Babu and I.S. Taak, Environmental conditions required for Fluid Flow Measurement Laboratory with special reference to Water Flow Measurement set-up at NPL, MAPAN-Journal of Metrology Society of India, 13 (1998) 3-8.
[22]
J.N. Som, R. Singh, V. Babu and I.S. Taak, Study on the error analysis and uncertainty parameters involved on the flow rate measurement with the NPL primary water flow measurement facility, in: Proc. of the 4th International Conference on Advances in Metrology (AdMet-2003), Delhi, Feburary 5-7, 2003.
[23]
S.K. Jaiswal, I.S. Taak, D. Sharma, C. Singh and A.K. Bandyopadhyay, The fluid flow measurement standard at National Physical Laboratory, New Delhi, in: Proc. of NCSLI 2010 Workshp and Symposium, Providence, USA, July 25-29, 2010.
[24]
S.K. Jaiswal, S. Yadav, and R. Agarwal, Multiple weighing based method for realizing flow, MAPAN – Journal of Metrology Society of India, 30 (2) (2015) 119123.
[25]
D.W. Spitzer, Flow Measurement: Practical Guides for Measurement and Control, 2nd Edition, Research Triangle Park, NC: ISA, 2001.
[26]
NABL 162, Guidelines for proficiency testing program for testing and calibration laboratories, issue no. 4, 2008.
[27]
S. Yadav, V.K. Gupta and A.K. Bandyopadhyay, Standardisation of pressure calibration (7-70 MPa) using digital pressure calibrator, Journal of Scientific and Industrial Research, 69 (2010) 27-33.
[28]
S. Yadav, V.K. Gupta and A.K. Bandyopadhyay, Standardisation of pressure measurement using pressure balance as transfer standard. MAPAN–Journal of Metrology Society of India, 26 (2011) 133-151.
[29]
ISO / IEC 17043, Conformity assessment - General requirements for proficiency testing, 2nd edition, 2010.
[30]
NABL 164, Guidelines for inter-laboratory comparison for calibration laboratories where formal PT programs are not available, issue no. 2, 2012.
List of Figures
Fig.1. Double weighing tank based water flow measurement system, (a) block diagram (b) photograph showing weighing tanks with diverter and fishtail Fig. 2. Operating principle of double weighing tank based WFMS. Fig. 3. Schematic diagram of diverter with design parameters. Fig. 4. Screen shots of the Automation Software developed for single weighing. Fig. 5. Data acquisition and control chart of prototype WFMS. Fig. 6. Calibration of weighing tank using mass standards method. Fig. 7. Plot of diverter error verses flow rate without adjustment of proximity switch. Fig. 8. Plot of diverter error verses flow rate after adjustment of proximity switch. Fig. 9. Plot of interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Fig. 10. Screen shots of the Automation Software developed for multiple weighing. Fig. 11. Comparison of single weighing and multiple weighing measurement results.
Figure 1(a)
Figure 1(b)
Figure 2
Figure 3
Figure 4(a)
Figure 4(b)
Figure 5
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10(a)
Figure 10(b)
-0.94
Figure 11 (a)
-0.94
Figure 11 (b)
List of Tables Table 1 Uncertainty budget of new water flow measurement system of size DN100 Table 2 Uncertainty budget of electromagnetic flowmeter calibration at 90 m3/h in current
output mode Table 3 Interlaboratory comparison results of electromagnetic flowmeter of DN80 size. Table 4 Comparison of single weighing and multiple weighing data where diverter error is
uncorrected and proximity switch is not adjusted.
Table 1 Collected water mass
Sources of uncertainty
[Relative standard uncertainty (%)]
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
2000 kg 0.005 0.0058 0.005
1500 kg 0.0067 0.0077 0.005
1000 kg 0.01 0.0116 0.005
500 kg 0.02 0.0231 0.005
0.0001 0.025 0.005 0.0271 0.055
0.0001 0.025 0.005 0.0279 0.056
0.0001 0.025 0.005 0.0302 0.061
0.0001 0.025 0.005 0.0401 0.081
Table 2 Sources of uncertainty
Weighing tank calibration Resolution (indication) of weighing system Buoyancy correction, storage and evaporation effect Timer calibration Diverter timing error Water density determination Digital multimeter calibration Repeatability (Type A uncertainty) of flowmeter Relative combined uncertainty (uc) Relative expanded uncertainty (U) at k = 2
Collected water mass [Relative standard uncertainty (%)] 2000 kg 0.005 0.0058 0.005 0.0001 0.025 0.005 0.005 0.031 0.0415 0.083
Table 3 Sr. No.
Nominal Flow Rate (m3/h)
1. 2. 3. 4. 5.
10 30 50 70 90
NPLI assigned value Average Error (%) -0.792 -0.865 -1.012 -0.783 -0.891
Type A Uncertainty (%) 0.029 0.033 0.025 0.027 0.031
Expanded Uncertainty (%) 0.088 0.098 0.079 0.081 0.092
FCRI assigned value Average Error (%) -0.880 -0.745 -0.952 -0.873 -0.966
Type A Uncertainty (%) 0.010 0.007 0.010 0.014 0.009
Expanded Uncertainty (%) 0.080 0.080 0.080 0.080 0.080
En value
0.74 -0.95 -0.53 0.79 0.62
Table 4
S. No.
Type (single/ multiple weighing)
1.
Single weighing
Collected Mass
Flowmeter *Expanded Error uncertainty Standard Estimate Standard % at k=2 uncertainty (ρ) uncertainty U (flow) (ut) (uρ) % kg/m3 %
Collection Time
Water Density
Estimate (M) kg
Standard uncertainty (uM) %
Estimate (t) s
1000
0.02
90
0.0951
994
0.005
-0.845
0.205
2.
2000
0.02
180
0.0476
996
0.005
-0.856
0.121
3.
3000
0.02
270
0.0317
996
0.005
-0.864
0.099
4.
4000
0.02
360
0.0238
996
0.005
-0.858
0.089
5000
0.02
450
0.0190
996
0.005
-0.867
0.085
6000
0.02
540
0.0159
996
0.005
-0.859
0.082
7.
7000
0.02
630
0.0136
996
0.005
-0.888
0.080
8.
8000
0.02
720
0.0119
996
0.005
-0.933
0.079
9.
9000
0.02
810
0.0106
996
0.005
-0.961
0.078
10.
10000
0.02
900
0.0095
996
0.005
-0.986
0.078
5. 6.
Multiple weighing
*Water density uncertainty of 0.01% (at k=2) and Type A uncertainty of 0.03 % have been considered for the expanded uncertainty estimation.
Highlights
•
An economical water flow measurement system (WFMS) is designed and developed.
•
Weighing tanks of conical design for faster draining are presented.
•
It does not require cumbersome process of diverter error evaluation.
•
Uncertainty of measurement is improved significantly in case of multiple weighing.
Graphical abstract