When it comes to water utilities managing data from AMI, “everything is evolving around analytics within the industry,” points out John Fillinger, director of utility marketing for Badger Meter.
“We’re at the beginning of smart water and we’re even getting into the artificial intelligence of water,” he says, adding that water utilities are starting to compile information for more extensive use.
Badger Meter’s Beacon AMA system holds incoming data in a database accessible to utilities and to utility customers. The Beacon system enables data to be shared among all of the different legacy software systems available within a utility today, Fillinger adds.
“We want to be able to share the information or bring other pieces of information into Beacon AMA so utilities get a broader perspective of what’s going on not only within the metered portion of the system, but other monitoring they’re doing, such as pressure or tank levels,” says Fillinger.
“From a data standpoint, everything is tied together into one package rather than having separate standalone systems that each serve a function, but are doing it in siloed ways.”
Michael Kanellos, OSIsoft spokesperson, says that while AMI systems gather data which then goes via wireless or Ethernet to a collector, managing it is somewhat different. He points out that there are a number of vendors whose techniques vary slightly, but mostly all concentrate on organizing data with an eye to the fact that it is time-series data.
Meter data is essentially a historical trail, with notices issued every 15 minutes. By preserving that chronology, one can later identify performance problems or figure out long-term fixes for factors such as water theft or energy consumption, notes Kanellos.
The challenge, he adds, is keeping all of the metadata such as time and location so it doesn’t need to be reconstructed later.
OSIsoft’s PI System gathers data primarily from SCADA systems and then organizes it so that people can take advantage of it, Kanellos says.
“We don’t harvest directly from meters, but meter data invariably goes through the system,” he says, adding that the company serves 15 utilities which, in turn, provide service to 250 million customers. The company also serves industrial customers.
Chuck Brunson, director of marketing at Neptune, points out that the company’s N_Sight software is a data collection engine and analytics tool that works with the company’s systems to enable utilities to use the data as needed.
“It helps bring in data from the AMI system relative to consumption profiling information for meters, as well as acoustic leak monitoring data and pressure data gathered by the AMI system,” says Brunson.
“The utility can use it simply to get those reads over to a billing system, or they can dive as deep as they want to perform analyses on the health of the AMI and distribution systems relative to leak, backflow, and all types of non-revenue water applications.”
The system offers the ability to implement custom reports. It also has mapping capabilities providing a GIS view to what’s going on in a given distribution network.
Joe Ball, Itron’s director of solutions marketing for water in North America, notes that in a traditional utility environment, data typically flows from end-devices through a network to be processed, stored, and analyzed.
“Big data refers to the tsunami of data that can result from using only this approach and solely pulling data into the back office to be analyzed,” he adds. “The way to address this big data dilemma is with the right data in the right place.”
In some instances, collection and analysis should take place within the endpoint, such as the meter or sensor, says Ball.
“Very granular data can be collected and analyzed, allowing the endpoint to transmit alerts or alarms, the right data, to the back office,” he adds. “In some instances, the endpoint can even take action based upon the data analysis, decreasing the reaction time greatly while other types of analyses—such as less time-sensitive analyses—are better conducted in the back office.”
Itron’s OpenWay Riva IoT solution is designed to deliver true interoperability and distributed intelligence, says Ball.
“The solution features a secure, multi-purpose network for smart water applications and allows an increasingly diverse ecosystem of meters, sensors, and distributed assets to communicate and interact intelligently throughout the network,” he says.
“This allows utilities and cities to use a single network for AMI, distribution leak detection, pressure monitoring, remote disconnect, and smart city applications, including smart lighting, smart parking, and smart traffic solutions.”
Itron Water Analytics’ cloud-based software manages, analyzes, and stores data collected using any of Itron’s data collection systems, providing users across the enterprise with the tools to streamline their effectiveness and significantly increase operational efficiency, says Ball.
Itron Total Outcomes is designed to take Itron’s existing capabilities in metering, communications, data management, analysis, and managed services and package them together in a subscription-based pricing structure.
As a part of Itron Total Outcomes, the company offers a variety of cloud-based software-as-a-service, managed-service, and onsite managed-service models, all of which are tailored to the utility’s specific operational and business requirements.
Factors such as the Internet of Things (IoT), cloud-based data storage, and artificial intelligence (AI) offer water utilities expanded options in data storage.
IoT starts with data collection and management, notes Kanellos. “It’s the garbage in, garbage out problem,” he points out. “If you don’t start with good data, your results will always be suspect. Ergo, the first step in IoT is setting up sound practices for data management and then data prep. Increasingly, a good portion of data collection and management will be automated.”
Case in point: OSIsoft end-user White House Utility District (WHUD) in Tennessee is a medium-sized water utility whose management sought to switch from detecting leaks with acoustic sensors to an IoT-based approach.
The utility divided its service territory into district metered areas and installed 79 new digital meters to study flow and pressure.
By comparing data from the new meters with data from the existing SCADA system, it was able to pinpoint possible trouble spots, notes Kanellos.
“But there was a problem. Taking data from one system and combining it with another took approximately six hours a day,” he says. “It was a full-time job. By automating data collection and merging it at the data storage level, the time required went down to 10 minutes a day.”
WHUD also projects the data onto a map so it’s easier to see where the problems are geographically, Kanellos says.
“In two years, WHUD has recovered $900,000 worth of water, discovered a 1.4 million-gallon-a-year leak that people thought was a stream, and saved more than $30,000 in employee time,” he adds.
IoT can help data collection and management through error correction, says Kanellos.
“If you compare historical IoT data with current data being collected, you can see if your data collection systems are working properly. IoT becomes a feedback loop,” he adds.
“With the Internet of Things, processing data at the edge becomes even more important as millions—or even billions—of devices become interconnected,” notes Ball. “Called distributed analytics, analysis occurs continuously at the edge of the network, sending insights rather than droves of data to the back office. Intelligent devices communicate and collaborate directly with each other and make decisions in real time.”
While the idea of leveraging public networks in order to be able to connect remotely all the way to individual machines distributed throughout the world and perform a myriad of functions has not become widely adopted yet, it is starting to emerge, notes Brunson.
“There are a number of different folks who have cellular IoT solutions, including Neptune,” he says, adding the company recently launched a new cellular IoT AMI product that works over public networks and sends the data back over to its N_Sight software solution.
“One that is getting a lot of press right now is LoRa, the alternative to cellular that’s maybe a little bit better suited for low-data-requirement, long-life, battery-powered devices,” says Brunson. “A number of different folks—including us—are looking at that as an opportunity to use as an IoT solution to collect data over public networks. We’re working with a couple of different partners to bring a LoRa solution to market. It looks like IoT is going to be a big part of the future for AMI.”
Fillinger says the technology Badger Meter has provided for the past three years dovetails into the Internet of Things in that it’s sending up data at one point in time, but once it comes into the utility, “you need to have that ability to merge it together with other pieces to truly have that picture.
“At that point in time, Beacon is taking data that’s coming in from the end-point and converting it into proactive information. From there, it’s shared across other platforms, or other information is being brought into Beacon to complete that picture.”
Fillinger says he’s hearing the term “Internet of Everything” “because it won’t be long before everything has a sensor of some sort and will be communicating bits and pieces of information so that everybody can make informed decisions.”
Graham Symmonds, chief knowledge officer and senior vice president for Fathom, says meter manufacturer systems tend to focus on the communication network and making sure the meter read gets to the head end as opposed to providing a lot of ancillary benefits.
“The system that comes with your AMI system is really focused on the utility side—on getting the read as opposed to analyzing the read in a lot of unique ways,” he says.
From a data management perspective, a utility can either have that as an on-premise data storage system or it could be a cloud-based data storage system, Symmonds notes.
“It depends on how the utility wants to use it, but in many cases the cloud-based storage facility is starting to win over the on-premise stuff,” he adds.
Fillinger notes that a few years ago, the cloud was the exception and not the rule.
“Today, it is the rule, so as customers are putting in fixed-network systems, the manufacturers have done a really good job of adding different features and functions into the applications,” he says.
Fillinger points out that water utilities are more focused on the tasks of delivery, cleaning, and the management of water and don’t have all of the expertise to manage a system, be it the infrastructure, gateways, servers, or other pieces that come with a fixed-network system.
“It’s almost become a natural fit to take that burden from the utility and move it to the manufacturer who has the experience and expertise to manage it for them,” notes Fillinger. “That really and truly has driven the growth of cloud-based applications on the water side in that it’s allowing water utilities to focus on what they do best, which is managing their water utilities and being able to get the most out of their software platform because they’re relying now on a partnership with their manufacturer to make sure they’re getting the most benefit from the software once it’s deployed.”
As Beacon AMA takes in the utility-owned customer data, once it gets into the cloud, then it’s a tool that can be used to analyze, export, and import information to be able to manage their system accordingly, notes Fillinger.
Cloud services make it simple for utilities to run their operations without having to devote resources to building and maintaining IT hardware and software, notes Ball.
“With more granular data from more devices, the cloud offers storage at a lower cost and the ability to scale as needed,” he adds. “Even with distributed analytics, cloud services are still necessary to store data for reporting and future analytics. The cloud is key to utility efficiency. It accelerates time to innovation, increases reliability, and lowers costs.”
While the cloud can play a role in data storage for utilities, it hasn’t become widespread yet, says Kanellos.
“While security is becoming less of an issue, bandwidth and latency are becoming more of an issue,” he says. “If it’s a large amount of data, you may want to keep most of it local and store only critical snapshots in the cloud. Bandwidth isn’t cheap.”
A single smart meter can generate 400MB of data or more a year, Kanellos points out.
“Sending all of the data from each customer to the cloud can inflate your connectivity costs,” he adds. “Then you have to pay to store it in the cloud. Meanwhile, local storage hardware continues to get cheaper all the time.”
Kanellos says he believes that a dichotomy will evolve over time.
“You’ll see operational data stay local,” he predicts. “Engineers and technicians need to access it and the volumes can be astounding. Billing information and summaries will go to the cloud. Companies also will use the cloud to conduct long-term or large analytics projects.”
Greater utilization of the cloud is definitely in the future, but like IoT, not yet widely adopted, notes Brunson.
“It provides so many advantages to everyone, but in particular our water utility customers because it allows them to not have to have IT infrastructure at the utility,” he says. “We can put all of the data and software solutions on the cloud, which means that instead of expensive servers that they have to manage themselves, they can get their software and their data management as a service.” That creates greater reliability, Brunson points out.
“Contrary to what a lot of folks think intuitively, the data is much more secure in the cloud because of the tools that cloud providers like Amazon and Microsoft and others have,” he says. “It’s actually a lot more secure than having the data on premises yourself on your own server.”
As for artificial intelligence (AI), Fillinger notes that while data has been captured for a while, the industry is in the infancy of it being shared across platforms “as we get to the point of having applications that are willing to share data,” he says, adding that not everybody is willing to do so.
“The business case they’re trying to build is that the data coming into the software is theirs and they want to try to drive everything through that application,” he says. “There are going to be a number of other applications that exist at utilities where that’s not going to be possible or where sharing is not going to occur.”
Once there is a more universal platform of data sharing, “then a look across the entire utility to be able to understand what is causing event A or event B or two independent isolated events that occurred—that is truly what will start driving some of the artificial intelligence so that utilities can then proactively know two events occurred,” says Fillinger.
“They need to be mindful of something else happening in the system, whether it be a temperature shift or a pressure spike—that something else may happen based on the information they’re collecting and the learning that is taking place.”
Artificial Intelligence is in its early stage of use in the water utility sector and mainly overseas, points out Kanellos.
United Utilities in northern England, in collaboration with the University of Exeter, has developed a program that predicts water demand 24 hours in advance with 95% accuracy, he says, adding that by knowing future demand, the utility can save energy by curtailing pumping at certain times. The utility serves nearly seven million customers through 42,000 kilometers of water mains.
MPWiK, a water utility in Poland, uses social media and TV schedules to predict big surge events such as people using their bathrooms after a big soccer match, Kanellos says, adding that utility managers use the information to ramp up and drop services as needed.
“Vitens in the Netherlands can now identify a large pipe break in two minutes,” he says. “You’re also seeing companies replacing scheduled maintenance with maintenance schedules based on hours used or factors such as unusual vibration signals. San Francisco Public Utilities Commission believes it could save 9,000 hours of employee time by smarter maintenance practices.”
While some of those practices are “true artificial intelligence,” many are examples of people using information they have in a more clever way, says Kanellos, adding that “either way, information and data is living up to its promise of being able to give people
the means to cut maintenance or capital costs.”
In rudimentary ways in the water industry, artificial intelligence is about pattern recognition and being able to use that pattern to infer some conclusions from data, Symmonds points out.
“It’s only just now from a metering perspective that you’re actually starting to collect enough data to have the right granularity to be able to do something with it,” he adds.
AI needs a lot of data in order to be able to distill the patterns, Symmonds says, adding that AMI “finally allows you to collect enough data to pull some of that information out of it from a big data perspective. Utilities have SCADA systems where you collect enough data to make some operational assessments possible, but from a metering perspective, this is New World stuff.”
Utilities can now look for patterns associated with leakage and the patterns associated with normal consumption versus abnormal consumption as well as patterns associated with a normal functioning meter as opposed to one that’s about to fail, Symmonds says.
AI is being implemented by electric utilities, such as Itron customer FortisBC, which worked with Itron Idea Labs on a project that uses machine learning techniques for grid connectivity, says Ball. While water and gas utilities have expressed an interest in AI, they have not yet implemented systems capable of this type of trending and analysis, he adds.
“We do see this trend on the horizon, though, as sophisticated networks are deployed by water and gas utilities,” says Ball.
Brunson points out that “a lot of the data management analytics tools we have are doing things to make data actionable for our utility customers in a way that already is artificial intelligence in that the software presents the answer to the question without the individual operator having to go in and crunch a bunch of data and do the analysis themselves based on some raw data. As technology gets more advanced, it’s going to get more sophisticated.”
Water utilities are implementing newer technologies and are reporting immediate positive results. Case in point: Cedar Hill is a 35.5-square mile community of fewer than 50,000 people 13 miles south of Dallas, Texas. The city offers water and sewer utility service.
Some of the challenges facing Cedar Hill have included service consistency issues, city-centric reactive operations, outdated infrastructure, old meters (most between 15 to 20 years old), reliability and accuracy issues, needed recapitalization, lost revenue, and billed water at approximately 60% of purchased water.
In response to its challenges, Cedar Hill transitioned from manual reading to advanced metering infrastructure (AMI), choosing Fathom Water Management, the utility-to-utility Smart Grid for water technology, to implement its complete suite of software-as-a-service (SaaS) and managed services solutions, including upgrading its automatic meter reading (AMR) meters to a fixed network advanced metering infrastructure (AMI) platform, modernizing the meter-to-customer vertical with SaaS-based utility CIS billing operations and customer engagement portal.
In doing so, Cedar Hill reduced the costs of providing water utility services by $394,000 annually and increased the number of readings per month from a single data point to 720, to enable clearer, near-real-time assessment of water demand and quicker response times for water leaks and losses.
Additionally, through active data management and near-real-time data access, Cedar Hill is collecting an additional $426,000 of revenue annually.
The Cedar Hill project stands out as a move toward a greater understanding of the value of smart water grids, the use of data to offset revenue destruction from voluntary and regulated conservation mandates, and extending the life of existing assets through analytics, as well as the intricacies of data and analytics to maximize collectible revenue and improve customer satisfaction, according to Jason Bethke, president of Fathom.
Bethke points out that resource volatility, infrastructure survivability, and financial viability through structural, voluntary, and mandated conservation, as well as customer centricity, are four major challenges faced by the water utility industry.
The installation was also designed to improve sustainability in the face of an increasingly volatile water cycle of extended periods of drought interspersed with extreme wet weather events.
Cedar Hill signed an agreement with Fathom in November 2013 that began a two-component process to provide enhanced water services.
Through the customer service component, the city outsourced billing and added what Joe Komisarz, manager of the city’s utilities services department, calls “some very powerful customer tools”—a customer service portal and a phone app, which went live in October 2014.
The other major component was the AMI hardware piece, replacing all 16,200 meters in the city.
“We added an electronic register and a radio transmitter unit into the meter box, which transmits hourly consumption data to our billing system—a Meter Data Management system,” notes Komisarz.
That, in turn, feeds the portal, the phone app, and the billing system.
Customers can manage their bills every hour if they choose to, and be aware of their water consumption. The data is stored in the cloud.
“The acquisition and implementation of the AMI system has been a major leap forward in service and capability,” says Komisarz. “We’re clearly and quickly transitioning into a customer-centric service model in which the customer has hour-by-hour control of their water bill.”
Through the project, Cedar Hill officials sought to enhance customer service, recover lost revenue, reduce expenditures, be fiscally prudent, and facilitate conservation and green initiatives.
“Cedar Hill generally prides itself on sustainability, conservation, and green guidance of initiatives,” notes Komisarz. “We’re known as Tree City. Our AMI project helps with a lot of the goals that the city has related to those subjects.”
The project added new tools for personal, real-time management of monthly consumption and utility bills with the ability to set consumption alerts through email or text notifications. It also offers immediate awareness of leaks or other unintended consumption, and has helped reduce careless water use.
Enhanced customer service also comes through monthly and daily detailed consumption data and analyses provided to customers with billing questions.
Water utilities can lose a significant amount of revenue through water theft from bypassed meters, unauthorized connections, meter degradation, meter programming errors, missing meters from billing inventory, meter installation errors, improperly sized or specified meters, data transcription errors, incorrect billing codes, and human errors. The newly installed system mitigates those challenges.
Cedar Hill has been able to reduce expenditures through reduced staff and tasks, eliminating six full- and part-time positions since 2014, saving billing and printing costs by contracting out billing functions, eliminating meter reading as a primary task, and deferring meter purchases up to 10–15 years.
The project also helped to meet Cedar Hill’s green initiatives by eliminating meter reading routes covering 210 miles monthly, saving emissions and fuel costs. Additionally, customers can pay bills without leaving home and, for the most part, without paper.
Additional project benefits include the ability for utility managers to make real-time, data-driven decisions as well as improved decisions in budget and planning, rate structure, and resource allocation. Improvements are being derived in training, service, responsiveness, performance, and conservation.
Future opportunities for leveraging the new system include the ability to analyze rate structure to consider shaping customer behavior in response to water consumption data and incorporating consumption and system data into operations models to reduce cost and increase efficiency.
The transition was not devoid of challenges. Among them:
- Large amounts of new data available
- Internal and external resistance to change
- A new public outreach and education workload
- A repurposed workforce and modified duties that previously had included routine, simple, repetitive, and reactive tasks and transitioned to diagnostic, investigative, complex, and proactive tasks
- New maintenance priorities
Still, Komisarz says he doesn’t envision any water utility not doing something similar in the future.
“The only issue now is cost and how a utility can afford to do it,” he says. “I don’t see any way not to do it as far as the utilities are concerned because customers—once they know they can monitor and manage their consumption and bills in near-real-time—are going to want it.”
Komisarz points out that it is a “very complex initiative with a lot of moving parts” involving software, hardware, and services.
“Not only do you have to buy, know, and acquire these various components of the project; these components have to talk to each other as well,” he says. “The reason we went with Fathom as the prime contractor is that we have one phone number for any problem—the hardware and the service.”
Komisarz says he is aware of some municipalities that have bought registers, radio transmitters, and software, all as independent actions, “and they’re having nightmares trying to orchestrate the moving parts to coordinate and synchronize.”
Lake County Public Works in Illinois—located 30 miles north of Chicago along the shores of Lake Michigan—has recently begun using analytics in its water service after completing implementation of an automated meter program in 2015.
Lake County started an in-house walk-by AMR program in 2010 just as its older remote equipment was beginning to fail and was no longer available from its manufacturer. Staff installed 4,000 AMR units.
“We knew we were going to have to increase the number of installs and look to an outside vendor just to get the entire service area done,” says billing supervisor Kathleen Dalessandro. “We hired a consultant who looked at all of the options—the walk-by, drive-by, or a fixed network—and we decided that the fixed network would work out best for Lake County because of our service area.”
A short list of vendors was populated, from which Lake County solicited RFQs. The final choice: a Badger Meter system with an Itron radio transmitter.
In addition to saving approximately $250,000 a year in operating costs through elimination of the meter reading contractor vendor and decreased staff, the data collected allows the department to provide enhanced customer service and revenue protection.
Lake County Public Works provides water and sewer services to 27,000 customers; of those, 21,000 are metered.
“We have an extensive area of unincorporated towns throughout Lake County so it’s not a contiguous area,” notes Dalessandro. “It made it a little unique in planning the AMR program.”
Lake County utilizes Itron’s analytics and fixed network system.
In less than two years, the department has moved from walk-by manual meter reading to a fully automated, hosted fixed-based system with 20,600 meters, 11 collectors, and 66 repeaters located on water towers and street lights throughout the customer area.
“Since beginning to use analytics for leak detection and customer service, utility personnel continue to realize new ways to leverage system data to improve operations and service,” says Dalessandro. “This has included training to spot usage patterns that may be used to help customers identify the type of usage.”
Using the Itron Analytics Metered Leak Detail Report, the department is able to identify ongoing leaks of which customers may not be aware.
“Once a property has been identified as having a potential leak, a member of our meter staff will notify the customer by phone, email, or a postcard left at the residence,” says Dalessandro.
“Customers are able to utilize the online portal to view personal usage data and take action without even needing to contact the utility,” she adds. “We began a customer outreach campaign to inform customers about ways to conserve and see more customers interested in conservation every day.”
Improved revenue protection has been achieved through various event reports now available, including no usage on active, usage on inactive, reverse flow, and tamper, says Dalessandro.
In an operational analysis, the utility saw that it was able to achieve significant annual savings. The meter, billing, and customer service staff has been reduced by 40%, with most cuts occurring through attrition and the positions permanently eliminated.
Cuts in staffing have not resulted in service reductions, with most customers receiving their bills sooner than in the past.
Meter reads for cycle bills are completed and reported in minutes rather than the typical three- to five-day span that it took to manually collect the reads and go back and check any suspect readings. In most cases, cycle bills are ready to print in the same day.
Final reads are collected in-house, eliminating the need to drive to each property.
“We are also able to go back in time and get a reading from a previous date when the notification of a sale or property transfer is made days or even weeks afterward,” says Dalessandro. “Billing accuracy has improved, reducing the number of cancel rebills due to meter misreads and estimated reads.”
The department has become more resourceful and proactive since automating meter reading activities, allowing staff members numerous opportunities to use AMI and analytics to enhance customer service and water conservation in Lake County, she adds.
“It’s really changed our customer service role from the reactive end of sending someone out to the house when the customer has a high bill,” says Dalessandro.
Now, the system enables Lake County and its customers to be proactive.
A customer service portal from Itron analytics, where customers can look at a screen similar to that being viewed by the customer service reps, offers the ability to see hourly use and to sign up for alerts, emails, and texts when they are using a large amount of water or an amount of water over a goal that they can set for themselves.
The system is enabling customers to identify leaks on their side of the meter.
“It’s a proactive customer service because we can find it before you get a 60-day bill where you’ve had this leak going on that you may not have been aware of,” points out Dalessandro.
“We reach out when we notice throughout the cycle that they have high use or continuous use through letters or calls,” she adds. “We still get customers who contact us when their bill is higher than they expected. We can then look at daily and hourly use, help them figure out where this water is being used, and conserve if it’s appropriate or if it’s something they can work on.”
It’s important to have support from the chosen vendor, she adds.
“It was a lot of work,” says Dalessandro. “The installation was a year-long process with a contractor and getting all of the software to mesh, but now that it’s in and functioning, our reading and billing has gone from being a days-long process to a couple of hours.
“You can pull the reads for the cycle in just minutes and bill it out the same morning. If there are any missed reads, we have to go out, but those are a very small number. Our maintenance staff now will try to get that meter reading while they are out there so that it doesn’t come up again.”
Dalessandro advises water utilities considering the same route to be aware “that whatever work you put into it on the front end really pays off on the back end.”