Implementing Automatic Meter Reading for : A Reflective Approach to Weighing Options

Roy McKenzie and Brian Houston

he water industry is both a business workshops and individual interviews. and a public service, relying on rev - Through SAIC’s expertise in meter reading Roy McKenzie, MSCE, is an engineer and enues to provide both. Because all systems and the city staff’s understanding project manager for the Tampa Water De - Tthings within the water utility typically are of city-specific procedures, policies, cus - partment. Brian Houston, P.E., LEED AP, is paid for with revenue income from water tomer concerns, and organizational priori - a senior project manager for water, envi - meters, revenue recovery, protection, and ties, a customized, comprehensive business ronment, and transportation with SAIC En - maintenance are integral to its successful case was developed that addresses the im - ergy, Environment & Infrastructure LLC. operation. Though revenues are not gener - plications to the city of transitioning to an ated from the service side of the industry, automated system. customer service is critical to the success of Properly constructed, a utility’s busi - the utility’s business. ness case is an extensive, customized evalu - Modern Automatic Meter Reading ation of all types of costs and benefits. AMI A well-run AMI system would have enabled (AMR) technology and Advanced Metering has the potential to offer far-reaching ben - Tampa to notify these customers of leaks Infrastructure (AMI) enable water utilities efits across many areas of a utility organi - within days after they began, rather than to achieve the following objectives more ef - zation—some which are financially waiting until the leaks were discovered fectively: quantifiable and others which may carry a weeks or months later as a result of a very  Maximize the use of available resources. very subjective value. Whether or not these high bill.  Lower costs. benefits are attainable greatly depends on a In AMR and AMI systems, water me -  Increase efficiencies. utility’s ability to implement and manage ters are read by electronic devices installed  Improve customer service. the organizational change that successful on each meter (meter interface units, or  Improve conservation. AMI transitions can create. MIUs) that collect readings from the meter Given the rapid technological develop - Very recently, a number of high-profile and transmit them via signals to data ments in automatic metering reading tech - complaints have been made against the city collection units (DCUs). In AMR systems, nology over the last decade, selecting from because of high water bills that may have DCUs are mobile, requiring meter readers the technologies available today and pre - been the result of customer-side leaks. This to walk or drive routes with a DCU in order serving the ability to build on today’s in - situation has prompted the city staff and to collect readings. With AMI systems, vestment tomorrow are daunting tasks. leadership to give special attention to AMI. Continued on page 26 In 2002, the Tampa Water Department staff conducted cost-benefit studies and small test pilot programs on AMR systems available at that time. The results indicated that cost savings and improved operational efficiency were achievable, but because the investigations focused mainly on direct cost savings that could be created from elimi - nating meter reading personnel, the pay - back period was unacceptable. Recognizing that many other benefits and implications are available today than in 2002, last year Tampa engaged the manage - ment consulting and engineering firm SAIC Energy, Environment & Infrastructure LLC to provide a thorough, unbiased assessment of the costs and benefits associated with transitioning from a manual water meter reading system to the advanced meter sys - tem known as AMI. The goal was to help control costs by improving the efficiency of meter reading and related customer service operations while enhancing the level of the services provided to customers. A cost-benefit analysis was prepared in collaboration with city staff through group FLORIDA WATER RESOURCES JOURNAL • AUGUST 2011 • 25 Continued from page 25 For a utility looking to collect once- of AMI (i.e., for all metered accounts), con - DCUs are permanently located strategically monthly meter readings for billing pur - sidering a range of equipment costs and a across the service area and relay the col - poses only, converting from manually read range of potential benefit levels for Tampa. lected data to a single central location, meters to AMI is rarely cost effective. To Many of the potential benefits can be quan - where it is organized in a meter data man - make AMI a sensible investment, a utility tified financially, based on reduced work - agement system (MDMS) database. must extend its benefits across many areas loads and associated cost reductions. Some A key distinction between AMI and of the organization, including billing and are well documented in other business cases other meter reading techniques is how often customer service, operations and mainte - and may create eventual financial savings for meter data can be collected. AMI-enabled nance, engineering and planning, and con - the city, but are difficult to quantify in meters can be read as often as desired, typ - servation and public relations. Also, the Tampa’s current situation. Still others can be ically every hour, or instantaneously upon impacts on human resources, information characterized as “soft” benefits that can only demand. Such detail offers tremendous po - technology, and finance that inevitably will be weighed subjectively against the net costs tential for analysis-based decision making, occur during implementation must be of AMI in order to support the city’s decision and timely information about individual planned and managed carefully to minimize regarding implementation of AMI. customer usage can be used to transform costs and transition complications. Recent customer leak issues and related customer service drastically. SAIC evaluated a full-scale deployment high bill complaints have caused Tampa to move from the prior practice of bi-monthly meter reading to monthly reading. The es - timated additional cost of this decision is Table 1: Summary of Costs and Financial Benefits $826,000 annually. Since monthly reads have become the policy, AMI is compared (Full Deployment for All Metered Accounts) financially against the city’s costs of opera - tion, including this recent addition. Recen t Con version to Con version to AMI f rom Manua l Monthly The financial evaluation of AMI iden -

Mon thly Read ing Con servative Le ss Con servative tifies capital costs as well as achievable an - nual operational savings. Capital costs are Cap ital Cost n/a $32 .3 million $30 .1 million amortized over 15 years to correspond with the expected life of the electronic field com - Cap ital Cost on a Add ition al Add ition al n/a pe r-meter pe r-mon th ba sis $1 .72 $1 .60 ponents, which comprise the majority of the system infrastructure. The amortized Add ition al Savings of Savings of Expe cted Annua l Operating Cost $826 ,00 0 $1 .7 million $2 .1 million capital cost and operational savings then can be represented as a net savings or net Add ition al Ope rating Cost on a Savings of Savings of cost per meter per month. pe r-meter pe r-mon th ba sis $0 .47 $0 .98 $1 .22 If the operational savings outweigh the Net Presen t Value Cost of AMI $16 .7 million $8 .7 million capital costs, AMI represents an investment (cost less bene fits) that provides a financial return, even if only Add ition al Add ition al Add ition al Net Per-Meter Per-Mon th Cost in the readily identifiable financial benefits. $0 .47 $0 .74 $0 .38 If the capital costs outweigh the operational savings, the city must weigh other potential benefits—soft benefits, as well as benefits that may be realized when the city’s cir - cumstances change—against the net cost of AMI to decide if implementation is war - ranted. Utilities often decide that the cus - tomer service benefits tip the scale toward AMI. Table 1 identifies the costs, savings, and net per meter per month amount for full- scale implementation of AMI in Tampa, based on a conservative set of assumptions. As shown, transitioning to an AMI system from the recently adopted practice of monthly manual meter reading would re - sult in a net cost per meter per month (over and above the $0.47 recently incurred as a result of the transition to monthly meter reading) of between $0.38 and $0.74. Subject to market fluctuations, the com - modity nature of AMI components implies that capital costs are relatively certain. While the precise cost depends on factors such as contract risk assignment, feature selection, and delivery and financing method, the val -

26 • AUGUST 2011 • FLORIDA WATER RESOURCES JOURNAL ues shown in Table 1 should fairly approxi -  Reducing labor and vehicle costs, along customers with new, accurate meters subsi - mate the contracted cost Tampa would incur with human error associated with rou - dize customers with older meters that do for the initial implementation. The opera - tine manual meter reading. not register all the water going through tional savings associated with AMI, however,  Downloading all the meter readings in them. depend greatly on the city’s internal ability every billing cycle quickly so bills can be Any new meter reading system that the to implement the identified organizational generated rapidly, enabling the city to city acquires should: 1) provide benefits (fi - changes. The city staff’s contributions to the notify customers proactively about ab - nancial and/or qualitative) equal to or ex - business case in this regard were crucial in normally high consumption (such as ceeding the costs of the system, 2) remain developing these estimates. that caused by leaks), before several reliable over its entire service life (15 years The transition to AMI must be planned weeks of consumption have accumulated or more for the electronics), and 3) avoid and managed carefully to ensure that costs into a large bill. technological obsolescence during its serv - are contained and savings are realized. Plan -  Eliminating the difficulties and risks as - ice life. ning will entail fully defining the desired sociated with reading difficult-to-access outcome and identifying the needed depart - meters. Alternative Strategies mental involvement and organizational im -  Reducing theft of service by enabling the pacts of the transition. A well-managed AMI city to observe evidence of possible theft The city of Tampa supplies water to implementation will include a unique pro - (tamper flags, sudden decrease in con - more than 146,000 customer meters. All of curement, changes in staff roles and respon - sumption between regular billing dates, its customers are metered, and meters are sibilities in many areas of the organization, etc.). installed outdoors in meter boxes and and extensive internal communication and  Continuously monitoring accounts that vaults. The city has approximately 1,000 public relations. have been shut off to ensure they stay off. meters that are very difficult to access be -  Creating greater customer awareness of cause they are located in locked back yards. Objectives water consumption habits, leading to These customers have been directed to read conservation improvements. their own meters. Primary objectives for considering a When implementing an AMI system, Several options are available to im - new metering system include improving the Tampa would replace older meters through prove meter reading and customer service. effectiveness of the city’s meter reading and the meter replacement program. This prac - Understanding the limitations and benefits customer service operations and enhancing tice provides added benefits of increasing of each option helps to focus the discussion the city’s service to its customers. AMI can revenues, reducing the city’s non-revenue of the city’s alternative strategies. help accomplish these objectives by: water and also reducing the extent to which Continued on page 28

FLORIDA WATER RESOURCES JOURNAL • AUGUST 2011 • 27 Continued from page 27 cost over prior bi-monthly meter reading line for evaluating an investment in alter - operations of $826,000 per year. When av - natives. Continue Current Operations eraged across the city’s entire meter popu - During the course of this business case lation, including those read by the city’s Mobile AMR evaluation, Tampa transitioned from bi- own staff, this additional cost translates to Tampa could invest in mobile AMR, monthly meter reading to monthly meter $0.47 per meter for each monthly reading which would mean that over 99 percent of reading, using an outside contractor to col - and billing. the meters would be read by meter readers lect the additional readings. The contracted The “no-change” option is that the city driving throughout the service area with cost per additional read is $0.56, and staff continues to physically read meters every portable radio-based meter-reading de - estimates that managing the additional data month in accordance with this recent deci - vices. will require a total of six additional staff sion. No additional investment in staff or Meter data collected by this method members in the call center, field service, and technology would be required, so monthly could include not only readings, but also billing area. The net result is an additional manual meter reading is used as the base - data flags indicating leaks or tampering that is detected between readings; however, the flags are not known to the utility until the reader drives by (i.e., monthly). Also, gath - ering more detailed information about con - sumption (such as very frequent readings to describe consumption patterns for meter sizing, to detect specific wasteful use pat - terns, or to describe leaks) would require a separate trip to each subject meter. In recent years, many larger water util - ities—including Washington, D.C.; New York City; Toronto; Cleveland; and San Francisco—have rejected mobile AMR, hav - ing determined that the cost differential be - tween mobile AMR and AMI systems is relatively small compared to the additional benefits AMI offers when deployed across the entire service territory. Mobile AMR is perceived to be an inferior choice for utili - ties anticipating that future opportunities and requirements will be placed on their customer service operations.

Fixed Network AMI The city could elect to install AMI equip - ment and the fixed network communications system that allows meters to be read without routinely deploying meter readers. Fixed net - works are somewhat more expensive than mobile systems, but in exchange for higher cost they offer greater operational savings and a richer customer service experience. These networks can enable the utility to gather a great deal of water-use informa - tion quickly, such as detailed descriptions of changes in a customer’s consumption patterns that could indicate leaks, and to share that information proactively with the customer, thereby reducing customer in - quiries. They also can provide the commu - nications network for other water operations data functions, such as water main leak detection, backflow monitoring, or pressure monitoring. The economic models presented in this article are based on cost estimates associ - ated with a fixed network solution because it is the only alternative that can provide the operational and customer service benefits sought by the city.

28 • AUGUST 2011 • FLORIDA WATER RESOURCES JOURNAL Comparison of Existing tomer service improvements and 2) engi - a timeframe much shorter than three years & Alternate Scenarios neering and operations improvements that would be overly aggressive. eventually may become financially quan - On the other hand, completing the This article presents the business case tifiable, but for which current city circum - transition more quickly would enable the comparing manual meter reading processes stances do not provide support to quantify. utility to enjoy the benefits of the new tech - with the costs and benefits of a prototypi - The following sections present a summary nology more quickly. Although not recom - cal fixed-base network system, or AMI. The of the quantitative economic analysis and a mended for planning purposes, the business case was developed from two per - review of the other potential benefits AMI potential savings associated with a faster spectives: 1) using a more conservative set provides the city. implementation could be investigated dur - of assumptions to illustrate the effect of ing the procurement process, should instal - current uncertainties on the costs of imple - Quantitative Economic Model lation contractors express interest. mentation and operational savings and, 2) There are disadvantages to making the using less conservative assumptions. The re - The quantitative comparisons are installation process longer than three years. sult is ranges of potential costs and savings, based on installing the proposed AMI sys - First, installation contractors prefer to with greater costs and lower savings being tems over a three-year term and phasing in move faster, making the most of the con - associated with the more conservative per - the benefits of the systems over the instal - tract management staff they need on-site, spective. lation period. The three-year timeframe is lowering project overhead costs and thereby The business case includes the princi - deemed the optimal installation period for creating economies of scale and generally pal components of a quantitative economic the following reasons: lower per-unit installation costs. Second, analysis of costs associated with AMI de - An aggressive installation schedule installation contractors proposing on a ployment and the resultant benefits that (shorter than three years) requires compre - longer-term implementation would likely readily lend themselves to quantitative hensive management, including a reliable submit higher unit prices to reflect the im - measurements, along with consideration of mechanism for expediting the resolution of pact of inflation on their prices for material additional benefits realized from AMI uti - contract management issues, adequate and labor. lization that are not as readily quantified staffing for field installation support and Third, the project management in the with a dollar value. oversight, and intense change management city itself may not benefit from a longer The second component can be divided within the utility and with the public. For a timeframe. Longer timeframes increase the into two groups: 1) “soft” benefits, which utility the size of Tampa, and with the likelihood of turnover of critical staff work - typically result in subjectively valued cus - changes anticipated by transitioning to AMI, Continued on page 30

FLORIDA WATER RESOURCES JOURNAL • AUGUST 2011 • 29 Continued from page 29 tured by Neptune Technology Group. The The less conservative version of the ing directly on the project, and of loss of economic service life of a meter is a function model assumes a price of a meter register critical support for the program at other lev - of its accuracy decline (which is related to and MIU installation is $25. This figure re - els of the organization. Finally, taking longer water quality, total usage and usage pattern), flects recent contracting experience in would delay realizing the savings in operat - the cost of water (which influences the value which Tampa paid $17 to $18 per meter in ing expenses and the increases in revenues. of under-registration), and the cost of the replacement costs, which elsewhere may meter and the labor to replace it. Typical eco - cost double that amount. During AMI in - Meters nomic lives for small meters among U.S. water stallation, the city could avoid some of the More than 97 percent of the city’s more utilities range from 12 to 25 years. current costs of its meter change-out pro - than 146,000 meters are one inch or smaller, Tampa’s research into the accuracy of gram. and 99 percent of the meters are manufac - its own meters reveals an unusually high Tampa would need to purchase meter- rate of wear and degradation among the reading equipment, which would include small meters. In 2002, the city initiated a data collection units (DCUs) installed test program to assess the accuracy of rep - throughout the service territory. Some resentative samples of its 5/8-inch and one- companies design systems to use fewer inch meters at American Water Works DCUs by making more powerful MIUs that Association standard low, medium, and are capable of transmitting data over high flow rates (0.25 gpm, 2 gpm, and 15 greater distances and by making more sen - gpm, respectively). From the weighted av - sitive DCUs that can collect MIU data from erage rate of registration for meters in each a greater distance. Other companies have age group, the city determined that meters developed what they believe is a more cost- over eight years old can be replaced cost ef - effective solution by providing less sensitive fectively. and less powerful DCUs. While more of the Based on these findings, the model pre - less sensitive devices are needed to cover the sumes that three-eighths of Tampa’s meters service territory, the unit price of the DCUs would age beyond the eight-year mark dur - and their installation requirements are less ing the AMI implementation project, so expensive, yielding a lower cost system as a these would be replaced as part of the AMI whole. system installation. The balance of the small The best specific solution for Tampa meters must be retrofitted with new regis - would be determined after actual proposals ters as part of an AMI installation because and prices were submitted by bidding con - the current registers are designed to be read tractors. The economic analysis presumes visually and don’t include the electronic in - one (of the less costly variety) DCU for each terface required for AMI. square mile of the service territory. While fewer than 3 percent of Tampa’s The meter reading system would also water meters are over one inch in size, they include a computer and software for run - generate 60 percent to 70 percent of water ning the network; portable computers to revenues. The city has established regular, maintain and program meter interface frequent schedules to test large meters for units; and a meter data management system accuracy, so for this AMI project evaluation to store, control, and use the very large it is assumed that no meters larger than one amount of data an AMI system generates. inch would be replaced, and accordingly the financial model assumes no additional rev - Staffing enue generation resulting from replacing AMI deployment requires a project the older large meters. management team for the duration of the The increased revenues associated with installation period. The team would include replacing the oldest and least accurate me - six to seven people, including three to four ters during an AMI installation would be field inspectors checking installations and $152,000 annually. resolving problems. These staff members Each meter would require a meter in - would control the quality of the data gen - terface unit (MIU). Since Tampa’s meter pit erated by the AMI installer and resolve lids are nearly exclusively cast iron (which problems at customers’ premises. Associ - is sub-optimal for radio signal transmission ated costs include office and field staff, as and also costly to modify), all lids would well as vehicles. The project could be man - also be replaced with plastic or fibrous con - aged by a specialized third-party contractor crete lids. In the conservative model, the or by in-house staff. cost of replacing a meter register and in - Many utilities are constrained from stalling an MIU is $42 for a 5/8-inch meter, hiring additional staff, they may be con - by far the predominant size in Tampa’s strained competitively when attempting to meter population. While this figure is rea - hire specialized new personnel, or their ex - sonable based on national experience, local isting staff is already fully committed to circumstances and the city’s recent experi - current assignments. Using a third party to ence indicate that this estimate may be high. perform contract management would still

30 • AUGUST 2011 • FLORIDA WATER RESOURCES JOURNAL require some in-house staff to handle cus -  Calls related to changes in occupancy a customer service complaint. tomer issues associated with the project, would be unchanged, although the calls The workshop also produced the fol - such as consumption true-ups. for turn-offs and turn-ons would require lowing conclusions about the field staff The Tampa call center has undergone fewer field trips because the city would workload: significant changes within the last year. A monitor the accounts electronically in -  Currently 1 percent of cycle bills, or ap - total of 20 call center staff from different stead of turning them off. proximately 50 cases per day, are referred centers were consolidated into one center,  General topic calls would be modestly re - to the field staff for investigation. These with the ultimate goal of cross training staff duced. cases generally involve verifying a meter to handle a wider variety of calls. Eight staff In summary, participants at the work - reading or gathering a reading if none members are associated with water cus - shop concluded that the call center work - was gathered for billing. In the discus - tomer service inquiries, and two more are load would be reduced by 15 percent sion about the implications of AMI, being added to support the recent conver - compared to the workload experienced workshop participants felt that 95 per - sion to monthly meter reading. The call under the prior bi-monthly meter reading cent of these orders would be eliminated, center receives 1,400 customer service calls operations. This lower workload would re - saving the equivalent of one full-time per day. sult in a staff reduction to seven people position. SAIC led a workshop with Tampa staff (conservatively). The less conservative  Field staff workers are assigned approxi - members to review the flows of work han - model reflects a reduction of staff to five mately 250-300 orders per day to visit dled by the call center and to explore how people. properties to turn off water service, turn that work would change if AMI were in - The customer service representatives it back on, or read the meter. These work stalled system-wide and integrated into the would be able to access detailed consump - orders are created in response to changes city’s business processes. The staff reached tion data that would enable them to explain of ownership or occupancy, although the following conclusions for the four cate - a customer’s water use patterns precisely some of the orders to turn water service gories of calls typically received: during an initial telephone call, possibly re - on are in response to a payment made on  Payment-related calls, currently using 25 solving the inquiry without needing to per - an account that had been turned off for percent of the call center staff resources, form (and wait for) field investigations. non-payment. Workshop participants would be reduced by 5 percent. Also, AMI systems would alert the city to felt that these orders would be reduced  High-bill complaints, currently using 25 suspected customer leaks, enabling the city by 50 percent if Tampa used AMI sys - percent of the call center staff time, to notify customers of a potential problem tems to monitor accounts for unautho - would be reduced by 45 percent. before it becomes a high bill and generates Continued on page 32

FLORIDA WATER RESOURCES JOURNAL • AUGUST 2011 • 31 Continued from page 31 and financial evaluation described previ - ture, but conservation earns significant rized water use. Each field worker com - ously. These factors, though not possible to goodwill with many populations. Conser - pletes up to 50 orders per day, so the sav - quantify precisely, are significant both in vation benefits of AMI could include: ings would be the equivalent of three fashioning the relationship with Tampa’s  Monitoring and enforcing alternate-day positions. customers and in improving the manage - sprinkling bans and time-of-day outdoor  Field staff workers in the collections unit ment of the city’s water distribution system, use limitations. shut off services to delinquent accounts. which eventually may yield additional  Establishing pricing and Workshop participants concluded that quantifiable financial benefits. using AMI to measure the water used AMI-based monthly bills and faster leak Managers of water utilities that have im - during the various rate timeframes. detection would reduce some payment plemented AMI systems are quick to point  Analyzing customer consumption data problems because surprising large bills out that their systems have improved cus - in order to educate customers about op - would be less frequent; they estimated tomer service and customer satisfaction. De - timal irrigation practices. that the workload would be reduced by spite the difficulty of measurement,  Enabling customer-specific water budg - one person. competitive companies recognize the value eting. These field work reductions result in a of customer service, and in competitive in - Several potential benefits which may be total staff workload equivalent of five full- dustries, good customer service encourages quantifiable financially in the future are re - time positions. The less conservative model loyalty and increases revenue. For water util - lated to non-revenue water (NRW): the dif - estimates these reductions at a total of seven ities, whose customers are “captive,” it is ference between the amount of water full-time positions. The total savings in harder to correlate customer service with introduced into the distribution system and staffing was projected to be between 25 and revenues or costs. the total amount of water billed as con - 32 full-time positions. Water utility customers, however, sumption. It reflects leaks, theft, flushing clearly appreciate good service. AMI enables and other hydrant use, and meter inaccura - Additional Benefits of AMI the city to improve customer service levels cies. Tampa’s current NRW is approximately and levels of satisfaction substantially by 10 percent of production, which is not out - In addition to the quantifiable reduc - providing: side the normal range for utilities of its size, tions in workloads that would be created,  Frequent (monthly) bills based on error- but that percentage corresponds to more the fixed network and the detailed con - free actual meter readings. than $15,000 per day of lost water on a rev - sumption data would benefit the city in  Near-real time notification of leaks and enue basis. AMI potentially could reduce areas that are not included in the economic tampering. that loss.  Online access to detailed descriptions of AMI data could enhance meter right- water use to help identify wasteful water sizing programs and meter wear detection. consumption and help educate cus - These improvements will maximize registra - tomers about their consumption. tion and revenue, as well as optimize the in -  Customer choices about billing fre - vestment in water meters to ensure that the quency, and even the time of the month meters are not too large (thereby under-reg - that bills are created. istering flow) or too small (thereby wearing  New services such as submetering (for out prematurely from overuse). stores or buildings in a complex, or for A fixed network system could offer sup - deduct meters), which the city might port for acoustic leak detection devices that provide reliably and at low marginal can provide the utility with the earliest possi - costs with AMI in place. ble notification of a perceptible change in sys - Customers’ attitudes come into play tem acoustics, indicating a possible leak near through the political or regulatory process installed monitors. Finding leaks when they when the utility has to raise rates, undertake are small or non-surfacing could provide huge needed infrastructure improvements, annex savings where service lines or mains are prone new service areas, or inconvenience cus - to joint and small corrosion leaks. Systems tomers for street openings for repairs. Utili - where the surface and subsurface conditions ties with poor customer service have keep leaks hidden (such as sandy soil) could experienced difficulty in obtaining needed benefit substantially. If a leak is identified approvals, while utilities with high levels of promptly, it typically costs less to repair, hav - customer service are seen as valuable assets to ing done less damage during a short-duration the community. While difficult to quantify in flow. monetary terms, good customer service does Proactive leak repair could also be per - lower the cost per customer of providing formed during regular working hours, re - service and improves employee morale and ducing costs. Installing AMI equipment on turnover. fire meters or bypass meters on detector AMI data could be used to enable or checks could allow close monitoring of fire significantly enhance conservation efforts lines to detect leaks or theft. targeted at high demand periods, and the AMI enables mass balancing with dis - city could lower the maximum day water trict metering. The American Water Works demand as a result. This is not considered a Association’s Manual on Water Audits and quantifiable benefit because it reduces rev - Loss Control (M36) promotes district me - enue from a consumption-based rate struc - tering as a best-practice method to reduce

32 • AUGUST 2011 • FLORIDA WATER RESOURCES JOURNAL leakage. By observing flow during the off- costs; however, during periods of high de - ability to detect flow reversal in the distribu - peak period and contrasting that with ex - mand or if distribution system storage is tion system quickly, supporting distribution pected minimal consumption by customers, constrained, the higher energy costs still system analysis and public safety; and the abil - the net background flow could be calcu - could be incurred. The challenge is to avoid ity to analyze and influence peak demands to lated. compromising the ability to deal with fire - more appropriately size and schedule capital With AMI, the customer demand for fighting, unusual customer demands, and projects. any geographic area could be measured for main breaks or other interruptions of serv - Should Tampa choose to proceed with specific periods of time and analyzed rela - ice. Better information about the patterns of AMI, a number of the above benefits could tive to district meters. If the net flow, as - water use could allow the city to optimize the provide significant financial benefits in the sumed to be leakage and other forms of use of storage reserves and pumping without future, which may even result in a positive non-revenue water (e.g., theft) were high, compromising operational reliability. return on the capital investment.  field staff could focus on that area to reduce AMI could provide potentially signifi - non-revenue water. Like acoustic leak de - cant information about reverse flow at me - tection, district metering coupled with de - ters that could alert the city to water tailed customer meter data could be helpful distribution system issues, meter errors or in finding non-surfacing leaks by concen - inappropriate customer activity. Reverse flow trating investigations where there is meter detected by a meter or an AMI system may data evidence for them. be caused by four possible conditions, all of AMI provides a number of benefits which should be of interest to the city: back- through its ability to support monitoring of siphoning caused by pressure changes in the daily peak and off-peak consumption. Some distribution system; backflow caused by an utilities are considering imposing peak de - intentional or unintentional increase in pres - mand charges on some customers to en - sure on the customer side; meter tampering courage them to reduce water demands and or improper installation (where the meter is related pumping costs during periods of reversed in its setting); or failure of the meter high demand or high energy cost. Adminis - or the MIU to transmit correct readings, tering such rates and providing monitoring usually caused by programming errors in and feedback to customers in real time re - these devices. quires AMI. If Tampa could lower the maximum AMI data could be used to improve de - daily water demand (and/or the peak-to- mand forecasting because it allows access to average ratio), it might be able to design detailed consumption data, enhancing the new capital facilities slightly smaller, or ability to analyze demand from various classes postpone certain capital projects and ex - of customers. Detailed consumption data aids tend the life of existing infrastructure. in determining not only how to meet future Every year the city could postpone the con - peak demand, but potentially how to reshape struction, annual financing costs would be it through conservation efforts like time-of- saved. use billing. By doing so, the city could more Conclusion accurately size and schedule replacement or expansion of mains, tanks, pumps, and addi - The presently quantifiable financial tional water production facilities—and per - benefits of an AMI implementation in the haps wastewater collection and treatment city of Tampa do not, in and of themselves, facilities, as well. outweigh the capital costs when amortized AMI data could be used as an energy over the useful life of the needed AMI in - management tool. Energy for pumping is frastructure. The anticipated net cost of the largest significant component of the AMI, ranges between $0.38 and $0.74 per cost of water service, so better managing meter per monthly bill. The city must de - energy use may provide significant savings. termine if this cost is worth the additional Many electric utilities impose peak kilowatt benefits that are not yet readily quantifiable. demand charges as well as peak and off- These additional benefits are significant, peak kilowatt-hour charges on their large and other water utilities have judged them to customers, which include water utilities. offset sufficiently the net cost of AMI imple - When subject to such electric utility rate mentation. They include: improved customer schedules water utilities are motivated to re - service, including additional service offerings; duce peak energy requirements, especially substantial reduction of high bills and a re - if they coincide with the electric utility’s sultant overall increase in customer satisfac - peak rate periods. tion; opportunities to implement a number of Water storage management and pump innovative conservation tools, which would scheduling commonly are used to reduce increase goodwill among regulators and cer - peak energy use. To maximize capacity and tain customer populations; collection of de - minimize pumping costs, distribution sys - tailed data that would enable more targeted tem storage is replenished during off-peak reduction of non-revenue water, an issue that hours that coincide with lower costs the city more than $15,000 per day; the FLORIDA WATER RESOURCES JOURNAL • AUGUST 2011 • 33