By R.M. Zucchetti*, C.E.T. and D. Phair **
* Director, Simcoe Engineering Group Limited
** Senior
Technician, Simcoe Engineering Group Limited, Pickering, Ontario
In an excellent article about flowmeter selection, R.A. Furness provided a detailed technical overview of the many factors that should be considered1. But his first question was Do I need a meter at all? A flow indicator is adequate if all you need to know is that water is flowing. If alarm limits for high or low flow are needed then micro-switches can be added to the indicator. When measurement of flow within 10% is sufficient, then putting pressure taps at line size changes or at an elbow makes a crude differential producer that can be calibrated empirically.
Mostly, municipal water producers need to measure flow more accurately. A meter that is plus or minus 0.5 percent accurate is typical of the advertising literature for flow meters. There may be some fine print associated with the claim, but you go ahead and select the flowmeter, install it and then get disappointing results. What went wrong? There is a great deal of published literature explaining installation, type selection, and other details of flow measurement for water service. But despite the many articles and texts that address these matters the question: How accurate is my flowmeter? always remains.
When a metering system is used for revenue or billing purposes, a flow meter that indicates too low results in lost revenue. Meters that indicate too high may cause customer dissatisfaction! Chemical feed systems at treatment plants and pumping stations are often paced to flow. If the flow meter is inaccurate then chemical doses may be too low to be effective, or expensive overdosing may occur.
Inaccurate metering can also affect the planning and engineering processes. Often, decisions are taken on whether to expand a facility based on historical flow records. At one water pumping station, the meter recording was found to be almost 60% higher than the true flow. At another station the indicated flow was 45% lower than the true flow. Obviously, using these records as a basis for deciding whether to increase capacity would lead to the wrong conclusion!
There are many types of meter used for measuring water flow. The most common used in municipal service are venturi tubes, magnetic flowmeters, and positive displacement meters (on smaller services). This article deals with some practical results and observations that came from many flowmeter calibration and inspection exercises that Simcoe completed over the past three years.
Figure 1 summarizes the accuracy findings noted during inspections of over 200 meters ranging from 3-inch to 96-inch diameter. The results from venturi meters, positive displacement meters, and magmeters are all included. Almost half the systems tested had errors greater than 5%. The data excludes the occasional case where a meter was originally calibrated in US gallons and had been converted to metric as if they were Imperial gallons - an easy way to make a 15% error. A typical cost for water is $1.50 per thousand gallons. A 5% error in measuring a flow of ten million gallons a day could result in lost revenue of over $250,000 per year, or irate customers who have been overcharged. Figure 2 summarizes the generic problems found that affected accuracy. The percentages total more than 100% because some metering systems tested had more than one source of error.
Over 20% of the meters tested had installation problems that affected accuracy. Flowmeters of all types are designed to produce a response in secondary instruments from the effect of fluids passing through the primary device. This response is normally linearized by secondary instrumentation. The response of the primary device is dependant on the flow profile of the fluid as it enters and exits the meter. Manufacturers normally suggest the minimum upstream and downstream distance (in terms of pipe diameters) without direction or cross-section changes required to minimize the errors associated with flow disturbances. Often the installation of the flowmeter primary is a compromise between the added cost of having the ideal piping configuration and the added error caused by the flow disturbances. Other problems that were observed include having intrusive devices like flow switches near the metering section, and one case where a gasket was protruding over an inch into the metering section. Accessibility of the primary device is an important design criterion - if you can't reach it, you can't clean it.
The response to flow in the primary is communicated to the secondary instrumentation by sense lines. Over 25% of the metering systems had problems in the sensing lines. Sense lines can be either hydraulic lines for differential producers such as venturi meters, or electrical probes for magnetic flowmeters. Hydraulic sense lines may trap gas and sediments resulting in flow measurement error. Sense line installation is site specific and often results in high spots where gas may accumulate and low spots where sediments may be trapped. Electrical signal lines must be properly isolated and protected so that interference is minimized.
More than 50% of the metering systems tested showed errors in overall calibration. Accuracy of the primary device can be affected in two ways - buildups of sediment or growths, and dimension changes due to wear. The magnitude of these errors can only be determined by a flow test in a laboratory setting. For important cases like revenue metering or custody transfer, one advantage of the venturi meter is that the response of the differential producer can be calculated from internal measurements and first principles.
The secondary instruments include the flow convertor, transmitter, and/or linearizer and any signal loads such as recorder or computer inputs. When calibration is performed on a flowmeter system, it is usually the secondary instruments that get the service. The accuracy of the calibration of these secondary instruments is often erroneously used as the accuracy of the complete flowmeter system.
Secondary system accuracy for differential producing meters can be assessed by applying a known differential to the sense lines, i.e., simulating the differential producer, and comparing the anticipated system response and the observed response.
To minimize the errors in a flow metering system, all components within the system should be calibrated and evaluated separately. The installation, piping and primary elements should be internally inspected, maintained and cleaned regularly to minimize errors due to sediment, growths, and wear. The primary element operating error can be established for use in calculating system error. Hydraulic sense lines should be inspected and cleaned or flushed regularly to remove gas or sediment buildup.
Another consideration when evaluating an existing installation is the age of the equipment. Many flowmeters manufactured recently have greatly improved electronic technology that usually translates into improved accuracy when compared with older installations.
Finally, the system error for the flow metering system can be obtained from the individual errors found from the inspection, measurement, and testing work. Figure 3 is an example of overall metering system accuracy when all errors are considered. One of the most interesting things this shows is that there is often a range of flows where the error is low (and accuracy is high), but that range does not necessarily coincide with the range of flows that is used. Frequently, municipal design practice allows for systems to be in service many years, and the design flow is often much greater than the average flow in the early years. Sizing a metering system for flow in the year 2040 and expecting it to be accurate when the flow is only 10% of full scale is overly optimistic.
The fourth problem noted was the internal state of the primary device. A clean meter is better than a dirty one. Almost 30% of the metering systems tested had cleanliness problems that affected accuracy. To assess the effect of meter cleanliness, one 16-inch venturi meter was tested at Ontario Hydro's flow metering laboratory before and after cleaning. The growths inside the tube averaged 0.75 inches high, and caused an apparent flow that was almost 3% higher than after cleaning. The first photograph shows the sort of bizarre growth found in a large venturi tube used for revenue metering. The tubercles ranged from 0.5 inches to 1.75 inches. Filtered, treated, water can also produce tuberculation and growths in watermains. The second photograph shows a 36-inch venturi that had been used for metering raw water for twenty years without cleaning.
Inspection of large flow meters can bring you into unusual situations that the standard textbooks rarely mention. For example, open the manhole and remove the snakes resting peacefully in a nice dry cool chamber. Or open the manhole to find that the meter is under 30 feet of water.
Removing a meter for inspection and/or cleaning is preferable to in-situ inspection. Normally a one or two-person job for a small meter, the in-situ inspection of a large distribution system flowmeter can be a logistical nightmare, and an expensive and time-consuming exercise. First, block off part of the road or intersection at the meter chamber. Remove the manhole cover, pump out the chamber, ventilate it and test the air to ensure workers can safely enter. Isolate the piping section to ensure no leakage will occur when the metering section is removed. Remove bolts that have been in place for many years, and roll out the meter section for inspection. After inspection and measurement, clean the meter; install new gaskets; disinfect the line; pressure test the replacement; and finally put it back into service. The typical time taken for this exercise was 7-10 days because work could only begin after the morning rush hour, and had to be completed each day before the evening rush hour.
Older metering installations may pose their own special problems where the flowmeter has to be removed for inspection and cleaning. In one case, there were concerns that removing the meter from the piping could cause the meter chamber to collapse as the pipe appeared to be an integral part of the meter chamber structure.
Providing a bypass line so that a meter can be taken out of service for inspection and cleaning is well worth considering. The alternative may mean shutting down the only pipeline servicing an area at 3:00 a.m. on a Sunday morning when flow interruptions cause minimum problems. These are not popular assignments for the service crews!
To summarize, don't be surprised if your flowmeter is less accurate than you expect. The basic rules for getting value from a flow metering installation are simple:
Happy metering.
Reference 1: Furness, R.A., 'BS7405: The Principles of Flowmeter Selection', Flow Meas. Instrum. Vol. 2 October, 1991.