alt text Home  
alt text Products  
Library  
alt text Company  
Contact Us
Library
Application: Using Demand Profiles for Meter Sizing

1. Introduction


Model 100EL in pit
A demand profile consists of rate-of-flow data describing water use vs. time. Such data is typically gathered directly from a utility customerís existing meter installation using specialized flow recorders that attach to meters and log water usage per unit of time. This document explains what a flow profile consists of, how to collect a flow profile, and how to analyze the data in order to determine whether specific meters are properly sized, or should be replaced with smaller meters in order to reduce unaccounted for water.

Demand profiles generated from existing meters provide data essential for making a variety of critical decisions. Data logged from water meters is more accurate than other means because a water meter represents the most precise means to measure actual water use. Flow recorders accomplish their mission without interrupting the accurate registration of the water meter and, typically, without altering the existing meter configuration. In a small number of cases, adapters are required and easily installed.

Applications for customer demand profiles may be grouped into four specific categories: (1) checking on whether existing meters are properly sized as part of a routine maintenance and replacement program, (2) performing water use audits and leak detection programs, (3) collecting peak demand information for cost of service studies, and (4) collecting peak instantaneous demand data for creation of demand curves used for sizing new and existing service lines and meters. While only the first application is discussed in detail in this document, it is worth remembering that the same data gathered for meter maintenance and replacement purposes has other important applications and can benefit a variety of utility divisions, including distribution, metering, conservation, customer service, engineering, and finance. In the case of water use audits, demand profiles assist with conservation programs, leak detection, customer service, and hydraulic modeling. In the case of cost of service studies, demand profiles are used to obtain data regarding the variability of peak use by residential, commercial, industrial, and wholesale customer class groups that can be used for more equitable rate structures.

In order to determine if an existing meter is properly sized, a demand profile should accurately provide peak flow data and the percentage and volume of water flowing through the meter over the entire range of observed flows, from zero to the maximum recorded flow. Critical flow ranges include, as a minimum, flow below the specified accuracy range of a meter, flow at the cross-over range in a compound meter setting, and high flow. The objective is to properly size the meter for maximum accountability and revenue recovery without adversely effecting pressure levels or fire flow requirements. It is also important to consider meter maintenance costs. It may be that a 6Ē turbine meter could better serve a customer with constant flows of 600 gpm than a 4Ē turbine meter because, while both would accurately measure the flows, the 6Ē turbine would experience less degradation from wear and tear. The most obvious direct benefit of proper meter sizing is the accurate measurement of water use; the more closely a meter is matched to a customerís usage pattern, the more water will be accounted for and billed. What is often not quite so obvious is the potential size of revenue gains associated with proper meter sizing.

Tim Edgar, in The Large Water Meter Handbook, First Edition, illustrates this potential revenue gain with the case of a 100-unit apartment building with a 4Ē turbine meter. The actual monthly consumption was 500,000 gallons, but much of that volume was at low flow rates. Because the turbine meter was not accurate at flow rates less than 12 gpm, 15% of the volume went unrecorded and unbilled in both water and, as is very often the case, sewer charges. The result was a revenue loss of $1,700.00 per year (at $3/1,000 gallons for combined water and sewer). As Edgar points out, if a utility has 100 such incorrectly sized meters, those 100 meters would cost a utility over one million dollars in lost revenue over six years.



Model 100EL, close up
As an example, the Boston Water and Sewer Commission began a downsizing program in 1990. John Sullivan, Bostonís Director of Engineering, reported in presentations to the American Water Works Association that, between August 1990 and April 1992, the city had accounted for an additional 113,784 cubic feet of water per day (0.8 mgd). With just the meters downsized in the first year of the program, Boston anticipated the total increase in revenue over 5 years from combined water and sewer billings to be $6.8 million (1991 dollars). This level of revenue enhancement would only be realized in systems with many oversized turbine meters.

While the most direct benefit of proper meter sizing is increased revenue and accountability, meters offer a distribution system much more value than just revenue enhancement. Any decision made by a utility related to water usage can only be as good as the consumption data collected from meters. In general, demand profiles provide valuable data to improve distribution system design, performance, and management. In addition to finding ways to increase accounted-for water levels and revenue, demand profiles help to identify service size requirements, clarify meter maintenance requirements, define water use characteristics for conservation programs, enhance customer satisfaction and awareness, improve hydraulic models, and establish equitable and justifiable rate structures. Additionally, with increased water scarcity and cost, conservation has become an important industry issue. For many utilities, conservation has become the most cost effective means to improve water resource availability. All of these distribution system design, performance, and management objectives are dependent on the capability of a systemís meters to account for usage as accurately as possible, which can only occur as a consequence of sizing meters properly for each and every application.


2. Recorder Design


Another installation close up
Theory of Operation. Demand profiles are generated with electronic flow recorders. The portable flow recorders discussed herein are also referred to as demand profilers, demand recorders, and data loggers. The devices pick up data from the meterís internal drive magnets and store the data for later downloading into a desktop or handheld computer for analysis. These recorders can be moved from one meter site to the next with minimum effort and operate with standard meters, thereby eliminating the need for special registers. Typically, the magnetic sensor is either strapped to the outside of a meter using velcro or heavy-duty tape or is integral to an adapter located between the meter body and the existing register.

Because of the adverse operating conditions (meter pits, temperature extremes, rough handling, public access), recorders should be submersible, durable, and securable. In order to provide extended data storage capability in remote locations, recorders should also offer substantial battery life. This section describes current technology for demand profiling. As new technologies evolve in this field, they should be evaluated in order to promote this area of knowledge and capability.


Recording Methods. Flow recorders using magnetic pick-ups sense the magnetic field generated by the magnetic coupling of a water meterís internal drive magnets and convert the magnetic flux change into a digital pulse that is logged into memory and later downloaded into a PC for analysis. Each pulse is associated with a known volume of water. The principal advantage of a magnetic pick-up is the higher resolution of data made possible by the rotation speed of a meterís magnets. In almost all cases, the drive magnets inside a meter rotate much faster than the sweep hand (pointer) on the registerís dial face. In small meters, the number of magnet rotations per unit of time can be as high as approximately 30 per second at 20 gpm. At this rate, the magnets are rotating 900 times as fast as the sweep hand. In the case of turbine meters, the rotation speed of the magnets can vary greatly, from approximately 800 times the speed of the sweep hand to the same speed as the sweep hand. Available adapters can substantially increase the resolution of the data on many of the slower magnet speed meters by isolating an additional magnet with a higher rotation speed. Mechanical adapters are available to enable compatibility with the older gear-driven meters which preceded magnetic-drive meters.


Flux sensor on side of register
Installing Magnetic Sensors. Because most meters have the magnetic coupling directly under the register, it is typically easy to pick up a reliable signal by placing the sensor on the side of the register. Almost without exception, the magnetic coupling is directly under the register in the case of all 2Ē and smaller positive displacement and multi-jet meters. If the magnetic coupling is not directly under the register, it is typically in the center of the turbine rotor in the middle of the flow. In this case, the magnetic sensor must be placed on the side of the meter body in order to be as close to the drive magnets as possible. As discussed above, adapters are required for some meters, such as gear-driven meters.

If the magnetic coupling is under the register, but the register has shielding on the sides, the sensor may have to be located directly on top of the register in order to circumvent the shield. Because the recorderís magnetic sensor is essentially picking up the electromagnetic noise generated by a water meter, the sensor can be susceptible to picking up noise generated by other sources of electromagnetic noise such as motors, generators, and alarm systems. The recorderís sensing circuitry should be designed to consistently pick up the magnetic signal generated by a water meterís drive magnets, while minimizing the potential for picking up electromagnetic noise from other sources.


The Recorderís Data Storage Capacity. It is essential that a recorder have adequate data storage capacity in order to enable the recorder to store a substantial amount of data. As discussed in greater detail below (see The Recorderís Data Storage Interval), flow data must be logged into memory in small time increments if accurate maximum and minimum flow rate data is to be ensured. The potential factor of difference in the observed maximum flow rate between a 10 second and a 60 second data storage interval monitoring the exact same flow is 6:1. The potential factor of difference in the observed maximum flow rate between a 10 second and a 300 second (5 minute) data storage interval monitoring the exact same flow is 30:1. In other words, if a solitary flow usage of 200 gallons occurred for just 10 seconds at a rate of 1,200 gpm, whereas the 10 second data storage interval could detect this high flow rate of 1,200 gpm, the 300 second data storage interval would observe a maximum flow rate of just 40 gpm because the 200 gallons would get averaged over 5 minutes rather than averaged over 10 seconds.

Obviously, this difference could have serious ramifications for a meter size selection. Frequently, users choose to store data for one week when assessing the size of a commercial/industrial userís meter in order to ensure that a representative sample of flow data is gathered. If a user is to store 10 second data for one week, the recorder must be able to continuously store a minimum of 60,480 intervals of data. For other applications, such as cost of service studies and hydraulic modeling, a smaller data storage capacity is required than for meter sizing; however, if the same data is to be used most efficiently, the storage capacity should provide for high resolution data so that the same data may be used effectively for the various applications.


3. Recording Data

Length of Record. As discussed above, many recorder users choose to store data from commercial/industrial sites for one week because certain high rate water uses (e.g., a cleaning operation at a factory) may only occur on a particular day each week. It is important to discuss water usage with a customer prior to storing data, if possible, to ensure that the duration of the recording period is sufficient to get a representative sample of flow data. In the case of multi-tenant residential or hotels/motels, 24 or 48 hours of data may be sufficient as long as the data is collected during hot weather in the case of residential and high occupancy in the case of hotels/motels. Essentially, it is best to make some effort to understand a userís water use characteristics in order to select the optimum length of the data storage period. Experience with different types of users over time will also provide an indication as to the optimum record length for different classes of users. The record length is critical and should be determined on a case by case basis.

Customerís Water Use Habits. Data should be recorded during a period in which the user experiences typical peak, average, and minimum flow rates and for a duration sufficient to capture those rates. For example, it would not be appropriate to record data at a school or factory during a vacation period. Similarly, as mentioned above, you would want to record data for at least a week at an industrial site if there was evidence that the customer performed different operations on different days of the week. Seasonal cycles are as important to consider as weekly ones. Weather at different times of the year may substantially alter demand patterns. If a user uses a lot more water on a hot summer day, it is important to record data on such a day in order to capture peak flow data.

The personnel performing an analysis should anticipate potential changes in demand patterns. At a residential development, it would be important to consider the number of additional units currently under construction. It is also important to survey a user if the type of use changes. Commercial lease space can have a high rate of turnover. A bottling company could be replaced by a warehousing or distribution company with substantially lower water usage. If the meter is not resized, the new user will be the beneficiary of a lot of free water.


The Recorderís Data Storage Interval. This interval is the period of time over which a flow recorder counts pulses before that interval's pulse count is logged into memory. The interval determines the resolution of the raw data file from which all subsequent graphs and reports are generated: the shorter the interval, the greater the detail possible in subsequent graphs and reports. For example, a data storage interval of 10 seconds allows accurate data analysis for periods of 10 seconds or longer. The data storage interval is selected by the user before the recorder goes into the field. As long as the graph/report generating software allows for adjustment of the time interval over which maximum and minimum flow rates are calculated (see the section below entitled, Data Resolution and the ďMax-MinĒ Interval), the data storage interval should be kept short, e.g., 10 seconds.

Keeping the data storage interval small is particularly important in order to provide sufficient data resolution for accurately determining maximum flow rates. In order to ensure the accurate identification of a maximum flow rate, the data storage interval cannot exceed 50% of the duration of a maximum flow event. For example, if an industrial customer has a particular operation which occurs just once each 30 minutes, lasts 30 seconds, and uses 500 gallons of water (i.e., a demand of 1,000 gpm), identification of the 1,000 gpm flow rate can only be assured if data is logged into memory at least once each 15 seconds. If the data storage interval is between 15 seconds and 30 seconds, there is an increasing likelihood that the maximum flow rate will be understated due to the possibility that no data storage interval both begins and ends within the 30 second event. If the data storage interval is more than 30 seconds, the likelihood becomes a certainty. In this particular example, a data storage interval of 15 or less seconds would show the 1,000 gpm flow rate. On the other hand, if the data storage interval is 15 minutes (900 seconds), the maximum flow rate would appear as only 33 gpm because all that is known is that a total of 500 gallons was used during a 15 minute period, and 500 gallons divided by 15 minutes is 33 gallons per minute. If the data storage interval is 5 minutes, a maximum flow rate of 100 gpm would be indicated. A lower maximum flow rate would be indicated if the 500 gallon usage was divided between two 5 minute data storage intervals. As can be seen, a serious meter sizing error can easily be made if the recorded data is not stored at a level of resolution sufficient to capture the actual maximum flow rate.

As another example, letís say that a small manufacturing company has an operation which periodically uses 250 gallons of water for 10 seconds (which equates to a rate of 1,500 gpm) in addition to its other uses. This scenario is graphically simulated in the following 3 graphs. In each case, the exact same data was used to create each graph; the only difference is the data storage interval which, in this case, is also the interval used for maximum and minimum flow rate calculations. In the first case, a data storage interval of 10 seconds is used. In the second case, the data storage interval is 60 seconds. In the third case, the data storage interval is 300 seconds.



Graph using 10 second interval
 Meter-Master Model 100 Program



10 Second Data Storage Interval:
With a 10 second interval, the true maximum flow rate of 1,520 gpm is identified.











Graph using 60 second interval
 Meter-Master Model 100 Program



60 Second Data Storage Interval:
With a 60 second interval, the calculated maximum flow rate is reduced to 280 gpm.











Graph using 5 minute interval
 Meter-Master Model 100 Program



300 Second Data Storage Interval:
With a 300 second (5 minute) interval, the true maximum flow rate disappears into the rest of the data.













Although the above examples exaggerate normal circumstances, they are intended to clearly illustrate the potential for meter sizing errors if one ignores the importance of data resolution.

It should be noted that there are possible disadvantages to making the data storage interval too small. This interval defines the size of the downloaded data file and the length of time you can record before running out of memory. The same test recorded with a 5 second interval will take up six times more memory than one stored with a 30 second interval. Furthermore, larger files take longer to download and to generate graphs and reports. Generally, a 10 second interval provides adequate detail and recording time for most applications. If you are making a long recording and a 10 second interval would use up all of the logger's memory before the recording is completed, lengthen the data storage interval. Another problem with too short an interval is discussed in the next section, The Meter's Pulse Resolution, and in the section below entitled Data Resolution and the "Max-Min" Interval, under Creating Reports/Graphs. Briefly, if too short an interval is used on a meter with slow moving drive magnets, skewing (exaggeration) of maximum and minimum flow rates can occur because there is too little data for accurate calculations. A recorder's operating instructions should identify such meters so that care is taken when selecting intervals for data presentation. Software design can improve the integrity of downloaded data by intelligently interpreting pulse data in order to minimize the potential for exaggerated maximum and minimum flow rates. The Model 100 Program includes a feature called "AccuRate," which eliminates exaggerated maximum and minimum flowrates due to slow moving magnets.


The Meter's Pulse Resolution. This resolution is defined as the number of pulses generated by a meter that equate to a unit of liquid measure. For magnetic pick-ups, the resolution is the number of meter magnet poles (as the magnets rotate) which equate to a unit of liquid measure. It is desirable that the internal magnets revolve as fast as possible without degrading the reliability of the meter; accordingly, the higher the number of magnet poles per unit of measure, the better. Faster magnets generate more pulses which translates into greater data accuracy. Therefore, it is important to have some knowledge concerning the speed at which a meter generates pulses. A flow recorder's operating instructions should provide guidance in this area.

The pulse resolution (or factor) is especially important when determining maximum and minimum flow rates. The issues are very similar to those discussed in the preceding section, The Recorder's Data Storage Interval. Concerning maximum flow rates, if a magnet is rotating slowly, it is possible that a large, short term usage could take place without any evidence of its occurrence. For example, if a 6" turbine meter (meter "a") generates just one magnetic pulse for each 500 gallons while another 6" turbine (meter "b") generates one pulse for each 2 gallons, the 250 gallon usage at 1,500 gpm described in the preceding section might not even be identified at all by a recorder attached to meter "a", while meter "b", with the fast moving magnets, would have provided 125 pulses to the recorder. Furthermore, if the recorder attached to the meter with the slow moving magnets did detect one pulse within a 10 second interval, it might be erroneously assumed that 500 gallons were used during that 10 second interval which would equate to a flow rate of 3,000 gpm. It equates to a flow rate of 3,000 gpm because, if one pulse is logged in 10 seconds, this is the equivalent of 6 pulses per minute, and 6 pulses/minute multiplied by 500 gallons/pulse equals 3,000 gallons/minute. Accordingly, a meter with fast moving magnets can provide continuously accurate data throughout the flow ranges, whereas, a meter with slow moving magnets cannot.

Minimum flow rates identify leakage rates and impact the selection of turbine vs. compound meters in larger applications. In order to ensure the accurate identification of minimum flow rates, as with maximum flow rates, a user must know which meters have slow moving drive magnets. For example, if a meter's magnets are providing just one pulse for each 20 gallons, and the current flow rate is a steady rate of just 5 gpm, only 1 pulse will be generated each 4 minutes. If one observes the data in time increments smaller than once each 4 minutes, the flow rate will appear to vary between zero and some amount greater than the actual flow rate of 5 gpm. As an illustration, if a 1 minute time interval is used for observing the data, the flow rate will appear to equal zero for 3 of each 4 minutes and 20 gpm for 1 of each 4 minutes because each pulse, equaling 20 gallons apiece, will appear just once each 4 minutes when a steady flow rate of 5 gpm is occurring. If a 4 minute time interval is used to observe the data, it will appear as if a steady flow rate of 5 gpm is occurring.



Graph with 1 minute grid interval

The two adjacent graphs represent the scenario just described. Both graphs were generated from the exact same data, but the time increments used to view the data are 1 and 4 minutes, respectively. Each pulse from the meter equals 20 gallons, and they were spaced 4 minutes apart (except during the initial interval shown). Software design can help by evaluating the data to determine the likelihood that raw pulse data should be averaged over longer periods of time because the pulse distribution indicates the presence of a constant flow rate. The Model 100 Program automatically performs this task.



Graph with 4 minute grid interval
 Meter-Master Model 100 Program


Unless you are actually at the meter site watching the meter at the time of the event, it is not possible to know with certainty whether a periodic use of 20 gallons is occurring or a steady flow rate of 5 gpm is occurring. If each pulse from the meter equaled a smaller amount of water, such as one gallon, the true picture would be much clearer.

The key to getting accurate flow data is generating a sufficient number of pulses per time interval. In the case of magnetic pulses, all 2" and smaller positive displacement and multi-jet meters provide a good pulse resolution, such that the data can reasonably be observed in time increments as small as 10 seconds. Adapters which increase the magnetic pulse resolution are useful in determining accurate flow rate data because the flow data may be accurately viewed in smaller time increments which minimizes the need for interpreting the data with potentially inaccurate assumptions.

Meter Accuracy. When one uses a flow recorder, it is assumed that the meter it is attached to is accurate. A flow recorder can not determine meter accuracy, but it can determine the accurate meter type and size for a meter site. Because a flow recorder is only as accurate as the meter it is attached to, routine meter testing is important when using recorders to determine the appropriate meter size. Because most meter inaccuracy involves under-registration of usage, a flow record on an under-registering meter can result in the selection of an undersized meter.

Ideally, a meter should be tested for accuracy, and repaired/recalibrated if testing indicates that it is not accurate, prior to recording data for meter sizing purposes. As discussed under Meter Maintenance Considerations later in this document, a demand profile performed in conjunction with a flow test may indicate that all of the flow is occurring in an accurate range of the meter, even though the meter is not accurate throughout the flow ranges. If this is the case, the meter does not need to be repaired/recalibrated because no accountability or revenue is currently being lost.

Flow recorders should be considered a valuable companion tool as part of a meter test program. As referred to in the previous paragraph, a flow recorder can identify the percentage of flow in low, medium, and high flow ranges. With this information, testing can be focused on the ranges in which most of the usage is occurring and unnecessary and costly repairs can sometimes be avoided. If a flow record indicates that all of the flow at an oil refinery or brewery is occurring in a high flow range, it is not relevant whether or not the meter is accurate at low and medium flows.


4. Creating Reports/Graphs

Verifying Data Accuracy. One of the principal advantages of recording flow data directly from water meters rather than using alternative technologies, such as ultrasonic devices, is that the resultant flow data is based on and may be verified against the meter's registration. Graphs and reports generated from the data may be used with confidence because the accuracy is based on the premise that a water meter is the most accurate and reliable means to measure potable water use. However, if the accuracy of the data generated with a flow recorder is not verified by comparing the total volume observed by the flow recorder to the total volume registered by the water meter itself during the data storage period, this key advantage is lost.

Verification of data accuracy is critical and is accomplished: a) by requiring the user to enter the beginning and ending meter readings when downloading data and b) by having an accurate meter magnetic pulse factor database so that the total volume registered by the meter may be compared to the total volume registered by the flow recorder. This procedure also requires the operator to take special care, when making a record of the meter readings, that the numbers are accurate and include digits down to the decimal. In order to read a meter down to the decimal, a digit for all rotating dials and painted on "zeros" must be read.



Customer Information screen
 Meter-Master Model 100 Program
The sample software screen shown requires the user to compare the meter's register volume to the flow recorder's observed volume. The numbers should either be extremely close or differ by an explicable margin. In this case, the electronically recorded total of 1,291.774 gallons compares favorably with the water meter's registered volume of 1,295 gallons during the same period. The software calculates register volume by subtracting the meter's beginning register reading from the ending register reading. The software calculates the recorder volume by multiplying the total magnetic pulse count for the entire recording period times the magnetic pulse factor for that meter in the software's database. An explicable difference between the two total volumes would include differences due to change gears used in some meters for calibration purposes. Because the change gears are used to speed up and slow down the register to match the activity below in the meter's chamber, the recorder's volume could differ from the register's volume by as much as 15% even though both the meter and the recorder may have functioned 100% accurately. The software screen shown includes an automatic "Data Conversion Factor" option so that the recorder's volume can automatically be calibrated to match the meter's volume 100% in such cases.


Data Grid
 Meter-Master Model 100 Program

Data Resolution and the "Max-Min" Interval. Data resolution refers to the time intervals over which volume and maximum, average, and minimum flow rates are calculated. The sample software screen shown displays volume, maximum, minimum, and average flow rate data in a grid format. In this case, the volume interval is the time interval represented by each line of data. A volume interval of 300 seconds will provide volume data as well as maximum, minimum, and average flow rates for each five minutes of the survey. When creating a report or graph, the longer (larger) the volume interval, the shorter the report and the fewer the points plotted on a graph.

To compute a flow rate, the software calculates the number of pulses per unit of time selected. For example, if a 10 second Max-Min interval is selected, each volume interval is divided into 10 second increments, and the increments with the largest and the smallest pulse counts are converted to per-minute maximum and minimum flow rates, respectively. This represents one widely used method for calculating flow rate data. The considerations for selecting the Max-Min interval are similar to those related to the data storage interval. You want to ensure that the interval is sufficiently short for accurate flow rate calculations. In general, the maximum flow rate gets both larger and more accurate as the Max-Min interval gets smaller. Similarly, the minimum flow rate gets both smaller and more accurate as the Max-Min interval gets smaller.

Selection of the time intervals for viewing the data depends, in part, on the application. As discussed in section C above, Recording Data, it is important to consider the type of usage profile typically generated by each class of user. Usage at multi-family residential locations, for example, typically does not differ substantially in small time increments. Demand typically ramps steadily up and down in the morning and evening which allows for longer time intervals when viewing the data. On the other hand, an industrial user may have high volume wash cycles with short durations, requiring shorter time intervals for accurate maximum flow rate calculations.



Graph of customer's max/avg/min flow rates

Graph/Report Presentation Options. Software can present data in endless formats and styles. Generally, the software should provide options to view volume data, max/avg/min flow rate data, and rate vs. volume data. The 2 sample graphs shown display max/avg/min flow rate data and rate vs. volume data. The max/avg/min graph is useful for identifying instantaneous maximum and minimum flow rates and the duration of events. The rate vs. volume graph is useful for meter sizing and maintenance programs because it shows the percentage and volume of water being used in various flow ranges.



Rate vs. Volumn Graph
 Meter-Master Model 100 Program











5. Using Demand Profiles To Size And Maintain Meters

Summary of Meter Sizing Benefits. The use of demand profiles for meter sizing applies to all users. Although relatively standard meter size and water use patterns characterize single family residential customers, outdoor residential water use can differ substantially requiring meter sizes larger than the norm. With users other than single family residential, each customer generates a unique demand profile, and the meter should be sized accordingly. Although generic demand data can be developed for various customer class groups based on demographic and business type information, the cost of gathering customer-specific demand data is minimal when compared to the revenue and community relations benefits associated with maximizing meter accuracy and water use accountability.


Graph of an 8'' wholesale connection

The graph displayed is from an 8" wholesale connection serving a small residential community. Although the specified accuracy range for an 8" turbine meter is approximately 40 gpm to 3500 gpm, the flow rate never exceeded 40 gpm at this site. Accordingly, the customer received a lot of free water. Replacement of the meter with one that is properly sized and configured will substantially increase both accounted-for water levels and revenue. Although, in some cases, a smaller customer surcharge based on the meter size means less revenue to a utility in the short run, the overall water service cost to a community is reduced as a consequence of the lower capital costs associated with smaller meters.

Proper meter sizing has positive spill over effects with other programs mentioned herein. For example, a cost of service study in support of a rate structure design can only be fair and equitable if all of the sample sites have properly sized meters. Leak detection efforts are undermined if a meter is oversized because low flows are needlessly undetectable and the meter's pulse resolution is less than it would be with a smaller meter. Similarly, hydraulic models, conservation efforts, and other programs all benefit from accurate use registration which is dependent on proper meter sizing.

Compound vs. Turbine Decisions. Many utilities experience shifting philosophies concerning the application of compound vs. turbine meters. Compound meters are more expensive and have higher maintenance costs, but they register accurately through a broader range of flows. By comparison, turbine meters are less expensive to purchase and maintain, but offer a smaller accuracy range. For each meter application, there is an optimum solution, and a demand profile will enable you to make the correct decision in each instance. If a compound meter is installed when a turbine is more appropriate, excessive maintenance costs and problems can be expected, and the utility will unnecessarily lose money. Conversely, if a turbine is installed when a compound is more appropriate, registration will be lost, and, once again, the utility will unnecessarily lose money and accountability.

The rate vs. volume graph displayed above (bar graph) enables a user to determine the amount of flow occurring in the cross-over range of a compound meter setting. In a compound meter's cross-over range, there is a substantial drop in the level of accurate use registration because the turbine side of the compound setting is just starting to move, and, consequently, all flow through the turbine is below its accuracy range. If there is a meaningful amount of flow in the cross-over range, an alternative compound meter size or a single meter setting should be considered.

Meter Maintenance Considerations. Another related use of demand profiles is meter maintenance programs, especially large meter maintenance programs. Some utilities consider demand profiles when making meter test, repair, and/or replace decisions because the demand data enables the utility to perform an accurate cost/benefit analysis of these 3 maintenance options on a case by case basis. For example, if a 10" turbine meter tests 100% accurate in a high flow range, 90% accurate in a medium flow range, and 80% accurate in a low flow range, the conventional wisdom would average the 3 accuracies, which would equal 90%, and recommend repair. However, if a demand profile indicates that the flow rate never drops below 1,000 gpm, the in-service meter accuracy for the subject application would equal 100% because all flow is occurring in a high flow range. With the advantage of a demand profile, costly and unnecessary service interruption and repair costs can be avoided and appropriate maintenance programs can be devised. Proper check valve operation in a compound meter setting can also be evaluated by ensuring that the turbine side does not move unless the small side exceeds a specified flow rate.

Water meters, like any piece of machinery, have optimum performance ranges, and projected test requirements can be related to a user's demand profile. If a 4" meter is constantly being driven at a flow rate close to its high-end performance rating, more frequent repair requirements can be anticipated.