Assessing Leakage in Water Supply Networks Using Flowmeters
Flow measurement applications in the water industry are quite varied. They can range from small dosing and treatment flows in pipes just a few millimeters in diameter, to the flow of treated water or wastewater in trunk mains and aqueducts that are two meters in diameter and larger. Often in larger cities, vast interceptor sewers are used to collect huge volumes of effluent and spent water to convey them to wastewater treatment plants. Therefore, flowmeter usage is diverse and central to the entire water cycle control within the industry. The metering process directly or indirectly influences resource management, process control, new works planning, distribution management, leakage detection, financial control and environmental issues.
Managing Water Distribution Systems
Preventing water imbalances is becoming a key focus to increase the reliability and quality of the supply. The complete supply system consists of a number of elements. Each component needs to be effectively managed and controlled if the overall supply cycle is to remain within tight control. A typical system, shown in Figure 1, usually consists of the following main elements.
* Raw water piping system between abstraction and primary treatment plant.
* Piping and components within the water treatment works.
* Transmission mains and supply storage reservoirs.
* Local distribution supply mains.
* Connection pipes to the consumer's dwellings.
* Piping in the consumer's premises after the point of final metering.
Water is drawn from the source, through an abstraction meter (SP1) and into the treatment plant. A second supply meter (SP2) may identify losses in the supply main. From the treatment plant, water passes through the outlet meters (OMs) into the transmission mains and finally into local reservoirs. District meters measure the amount taken out from these reservoirs into each district and finally area meters (AMs) measure flow in smaller supply mains into each subdivision when the supply mains split after the DMAs. In each area, there may be individual meters for shops, apartment complexes and factories. It is vital that all the meters within this system are of known performance and regularly maintained so the water balance within any section can be documented and the uncertainty correctly estimated.
Losses or imbalances can occur in any or all parts of this process. In the initial stage, large volumes of water are pushed through a few key flowmeters. For example, a large Venturi meter might be in a 1,600mm pipe measuring a flow of 250 mld. Over a period of time, the inside surface of the pipes and meter will be altered due to deposits and generally the meter may begin to under-read.
The intrinsic accuracy of an installed meter is fundamental to managing the water balance. If the same type of meter was used throughout the supply system, then they all could drift in the same way and at almost the same rate. In this way, the meters may balance, but the quantitative assessment of the flow is in error. This has been the case in many of the older networks. Invariably these meters have not been properly installed, calibrated or maintained. However, they are still used to measure the amount of water put into supply. From my experience of installing flowmeters, this total is never precisely known. Therefore, estimating leakage becomes extremely difficult.
Sources of Water Imbalance and Leakage
Leakage rates within small areas (or small towns) can be estimated by the direct measurement of flows at times when demand is low. Water companies make regular "soundings" during late night hours to try to estimate these system flows. However, most "macro-level" water balances are made directly from totalled values over longer periods by taking the difference between a measured input volume and the accumulated volume of many estimated or measured outputs over the same period. The uncertainty of metering directly affects the balance result. Thus, accurate flow metering is fundamental in water conservation projects.
The total water balance is made up of a number of components.
* Metered consumption.
* Unmetered consumption, estimated from the population statistics.
* Real losses or leaks.
* Apparent or inferred losses.
Metered consumption is very clear and has an estimated uncertainty. Unmetered consumption is less clear and has a much larger uncertainty. This larger uncertainty is partly due to the fact that the exact number of consumers is never precisely known and the rate of consumption may vary within an area depending on the balance between domestic and/or industrial users. Real losses consist of leaks and losses prior to the point of final measurement and may be measured through the use of system balance meters, district meters or zone flowmeters. More measurements may mean a better estimation of the flow totals, but regular verification is essential to error analysis. Apparent losses are largely due to the errors in the assumptions made on consumer demand, population, etc.
The AWWA Manual "Water Audits and Leak Detection,"1 lists six main sources of leakage that may occur in any section of the system.
* Material defects induced by poor design or insufficient planning at concept stage.
* Pipe breaks caused by poor workmanship in construction phase--laying and support of pipes.
* Operational errors--over-pressure, water hammer valve operation, etc.
* Corrosion due to soil and/or water chemistry effects and groundwater effects (e.g., seawater).
* Leakage from any of the installed fittings (valves, saddles, bends, tees, hydrants, etc.)
* Accidental or deliberate damage to hydrants and line air valves (including unauthorized tappings).
However, there also are hydraulic effects within the pipelines that may bias the readings to a far greater extent than is realized.2 Leakage also can occur in each reservoir or storage tank through evaporation (if open to the elements), by seepage (through defective concrete walls) and because of simple overflows in times of heavy rainfall. Leakage is very time dependant. Experience has shown that small leaks run undetected for long periods and, in older systems these, are a common source of imbalance. Reported studies generally indicate that losses in the water supply mains are less than in transmission mains, which in turn are significantly less than in the water mains. Thus, the majority of leaks are due to poor or non-existent metering in the many smaller water lines in each area of a city. However, most leakage policies concentrate first on the larger diameter lines, then each district and finally each area or individual customer usage.
Flow Metering Techniques
British Standard 7405 (1991) shows many designs and methods for measuring flow in closed conduits. Each design has strengths and weaknesses, and no single technique fully meets the needs of the municipal water engineer. With many ongoing developments, it takes a great deal more skill and knowledge to select the optimum meter for a given application. The operating principle forms a convenient way of classifying flowmeters and BS 7405 documents and names a number of basic groups (10 being used for closed pipes, 1 for solids type metering and the last for open channel meters).
In water supply systems, the common methods used almost always come from differential pressure types (BSI groups 1 and 2), displacement types (group 3), inferential types (group 4), magnetic types (group 6), ultrasonic types (group 7) and open channel types (not discussed in BS 7405). By far, the domestic inferential meters for individual usage in houses, apartments, factories, etc., are the most sold. Millions are used worldwide and have formed the backbone of network supply management for the past three decades.
Bulk supply traditionally has been monitored with Venturi or Dall tubes or large propeller meters. Recently, magnetic meters increasingly are being used in most parts of the world for these applications. Table 1 is a basic application table, largely based on experience and traditions, showing what meters are used for each application. This table should not be considered as the final means of selecting the best meter, but merely give an indication of the current choices.
For the elements shown in Figure 1, the current practice is to use ultrasonic or magnetic meters wherever possible, governed of course by costs. As large meter costs can vary enormously, the balance has to be struck between spending the right amount of money for an economic return. Normally meters are chosen purely on cost grounds, but this may not necessarily be the optimum choice. Life cycle costs are a much better means of judging instrument selection.
Expected and Actual Installed Meter Performance
When purchasing any instrument, the supplier usually includes a written specification for that device. While some manufacturers give detailed performances backed by independent testing, a minority of specifications barely enable the prospective user to determine what they are purchasing. Flowmeters are a little different from other instruments. They are tested in a flow laboratory under reference conditions. This means that standard flow rates are used in long straight pipes under steady flow conditions. Few manufacturers have comprehensive data on installation effects and often are reluctant to part with this data. An installation effect is defined as the variation from laboratory calibration to that obtained under field conditions.
Examples of flowmeter installation effects are
* Differences in pipe characteristics (roughness, ovality, etc.),
* Proximity of fittings (valves/bends) that are not present in reference testing,
* Differences in temperature (fluid and ambient),
* Effect of local RFI that is not present in reference testing, and
* Signal acquisition and processing errors of the local system.
The laboratory data is relevant for the meter in the laboratory setup only. Once the meter is installed in the customer's pipe, other changes may be introduced, such as
* Bore of the mating pipe is usually different to that of the meter,
* The flow rates in the transmission lines may be different from the lab data,
* There may be sediment, or calcification effects within the network, and
* Many other reasons.
All these variables may introduce additional bias into the meter readings from the initial day a meter is installed. This bias can be estimated or quantified only from an in-situ verification. This verification should be a part of regular network management activities. When purchasing flowmeters, many users expect the manufacturer's specification to apply immediately and be stable over time. This is a popular misconception. For example, small mechanical meters always are tested at the manufacturer's premises in accordance with local Weights and Measures regulations or international metrology standards. This is simply a guarantee to the user that when it leaves the factory it is within predetermined calibration limits. If it is installed incorrectly, installed too close to valves or used in a water supply with a high sediment content, its performance may shift. Usually, though not always, it under-records. Over time, with component wear, this under-registration may increase. There also may be a steady but noticeable fall in accuracy of the meter. All flowmeters, regardless of the type or source, show these time dependent effects to varying degrees. The key to successful network management is to estimate the rate of degradation (if present) with time.
It is quite difficult to calibrate flowmeters to much better than 0.2 percent total uncertainty. Therefore, this figure represents the baseline when any meter leaves a manufacturing facility. When the user installs the meter, this 0.2 percent lab figure almost certainly changes (increases).
Special consideration regarding the accuracy and installation should be given to the outlet meters from a water treatment plant because they are handling large volumes of treated water being put into the supply. Any flowmeters of 250mm and above should be carefully selected and even more carefully installed. Poor installation is the greatest source of error. Standards give guidance on the effect of single fittings but little data exists on the effect of multiple fittings close to flowmeters.4 Due to rapid developments, even accepted standards currently in use may not be totally correct.5 This is why site verification is vital to ensure the meter readings are valid.
All flowmeters are affected by a lack of attention to details at the installation stage such as protruding gaskets, pipe bore alignment, the proximity of valves and presence of pipe branches.6 In addition to these hydraulic considerations, close attention must be given to environmental aspects on secondary equipment to avoid excessive vibration, flooding, ambient temperature swings and other effects. These factors can greatly influence the actual measurement uncertainty that can be achieved. D.A. Phair's "Errors Occur Often in Industrial Flowmetering" discusses such effects.7
Taking into account these many installation factors, it often is difficult to demonstrate total installed errors of much better than 2 percent. In order to achieve this long-term figure, it is necessary to specify instruments with an intrinsic uncertainty of around 0.5 percent or better. This then allows more than 1.5 percent for all the other unquantified effects.
More attention should be given to the nature of the meter chosen and the application needs. If water companies would standardize on this 2 percent value, leakage figures would become much more understandable. It is essential that all meters are site verified to ensure that this 2 percent value is maintained, or at the very least any deviation is estimated.
It should be remembered that if each meter has a total installation within the network of 2 percent uncertainty or better, summing all these meter readings may still give total metered system imbalances of 10 percent or more. The performance critically depends on the number of meters, the type of meter, the design of each installation and the degree to which maintenance and field verifications are performed. There also are the uncertainties due to unmetered consumption, losses due to undetected leaks and any assumptions on inferred losses. Combined, these losses can give system imbalance uncertainties of greater than 25 percent.
Installation and Cost Implications
Cost often dictates the meter choice as well as the design of the complete installation. Unfortunately it usually is only the purchase price of the equipment that is the driving factor. The full cost of flow measurement is not simply buying the instrument. Recent independent studies have shown that the purchase price is only about 40 percent of the total cost of owning and operating the meter in the first year. The actual price depends heavily on the meter specification such as pressure ratings, materials of constructions, transmitter functionality, etc. Installing the meter is the single greatest expense in the first year. This covers both mechanical and electrical installation as well as the purchase of piping and ancillaries. As a result, many water supply authorities have focused on reducing installation costs. Previously, meters had been installed in chambers to allow access for maintenance and calibration. However, modern flowmeters require little maintenance, and recent advances in electronic signal processing and fault diagnostics have allowed field devices to be developed that do not require direct access to the meter. These devices use electronic fingerprinting to look for changes in both the meter and transmitter characteristics.
As a result, the direct burial of flowmeters, pioneered in the UK and South Africa, has gathered pace. This practice now has more than 15 years of practical experience from several large water companies in many parts of the world. Several thousand meters of all sizes now are direct buried. The three main benefits are
* Much lower installations costs--building of concrete chambers is more than the meter costs.
* Reduced maintenance costs--magnetic meters offer a mean time before failure in excess of 80 years.
* Greater reliability of the installation--tampering is eliminated.
It is important to understand the local soil chemistry to eliminate external corrosion effects from either acidic or alkaline soil or because of the ingress of groundwater that may contain pollutants or chemicals. Usually the meter is specially coated and surrounded with sand to act as a partial filter. Such techniques have been successfully applied in line sizes up to 2,200mm. Special attention also needs to be given to those cases where cathodic protection is used on the line. The cathodic current may bias readings unless it is correctly insulated. Finally, earthing of the line is essential in those areas where lightning strikes may be common. Despite these considerations, experience has shown that very significant savings can be made through such practices.
Water supply demand can vary greatly during the working day as well as show variations through the working week. Figure 2 shows the flow variations of a modern area meter (AM) installed in one street of a major U.S. city.
Peak demand in this case occurred on Sunday morning--presumably when people were watering gardens and washing their cars. Also notice the variations of peak demand between the average daytime velocity and the average night-time velocity. At this location, reversed flow was actually indicated. For this suburban location, flow velocities within the supply line always are low, putting severe demands on the performance of the flowmeter. Low line velocity is common in many supply systems, but what often is not appreciated is the pipeline dynamic variations between the day and night demands. This is a good example where modern metering has identified operational and system design problems.
Line pressure can affect leakage significantly. The operator has to give sufficient supply pressure while reducing the driving force that may cause water to leak out of a defective pipeline. As demand reduces at night, there is a tendency for line pressure to rise due to the supply pumps running to cope with peak demand. If pressure was reduced as the flow rate reduces, then the variations on stress at the pipe walls will not see a normal cyclic variation. It is this cyclic variation that partly causes leakage at connections, valves, etc. Several cities now implement pressure reduction control during the night hours. Examples of this have been documented in chapter 9 of Lambert's "Managing Water Leakage."8
The financial implications of incorrect metering often are not appreciated. When it comes to project costing and particularly operational costs, the monetary costs due to inaccuracy rarely are considered. Table 2 is part of a study made in a French city four years ago. For the price of water given, the financial uncertainty has been calculated for a range of flowrates occurring in a number of different line sizes at various levels of inaccuracy.
For a 200mm flowmeter running at 0.7m/s with a total installed uncertainty of 2 percent (a very real case), the uncertainty in billing the volume passed is 25,000 Euros/year (approximately, U.S. $27,000). This is almost twice the price of installing a traditional high quality water meter. For a smaller 25mm area meter, the corresponding figure is 400 Euros. Similar calculations for a major transmission line meter of 1,600mm show an uncertainty figure of 1M Euros.
Future Role of Metering
It is clear that accurate metering is fundamental to the future conservation of water resources and to the successful financial and operational management of existing water supply networks. Without metering, the true loss of water cannot be assessed, leading to financial under-recovery and possible errors in capital investment within the industry. An economic reduction in leakage rates can render the investment in new water sources, dams, treatment works or reservoirs unnecessary, or at the least enable a better economic justification to be made. Studies have shown that in major city water supply networks, financial uncertainties in excess of several millions of U.S. dollars per day are easily possible. When set against the cost of good instrumentation and maintenance practices, such leakage costs are very significant. This is why so many cities around the world are investing in long-term projects for water loss reduction.
Figure 3 shows data from Calgary, Alberta, Canada. Investment over a 20-year period has resulted in the saving of more than 100 million liters per day (mld), every year for the past decade. It also means better financial management and more prudent capital investment that has been based on accurate flow metering.
In the near future, as digital electronics and communications protocols become more accepted, flow meters are expected to evolve into small metering systems (pressure, flow temperature and alarm capability at each point). Even the small domestic mechanical meters are undergoing radical development and evaluation. They are evolving into solid state devices, linked to telephone systems, satellite dishes, local cable TV or other lines into properties, so that flowrate, consumption and other data can be automatically transmitted back to central centers for collation and billing.
This article is partly based on a article appearing in the Winter-Spring 2002 ITA Analyzer. The ITA can be reached at 702-568-1445 or www.instrument.org.