City of Riverside PUD on quantifying ROI and digital transformation in the water industry

The City of Riverside Public Utilities Department has embraced digital transformation by making better use of sensor-based data. Because they work closely with their board of public utilities on expenditures they track the value they are getting from their efforts. In this episode, we talk to the people who have grown the return on investment (ROI) they get from their work on the PI System from $672,000 per year in 2019 to $1,300,000 per year today. We discuss how they track ROI, how they improved adoption of the use of their Operational Data Management System (ODMS), what they did to improve water operations, how they use it in their market operations system for trading energy, and what value they found in putting their sensor-based data on their ArcGIS map displays.

Guests: Jennifer Tavaglione, Principal Project Manager

Brando Crozier - Administrative Analyst

Marc Smith - Utilities Principal Analyst

Ed Cortez - Principal Electrical Engineer

Robin Glenney - Water Quality Manager

Carey Tilden - Utilities Senior Resource Analyst

Wes Wisniewski - Utilities Senior Resource Analyst

Ivan Velasco - Power Scheduler/Trader

Co-Host: Gary Wong, Water Industry Principal, AVEVA

Subscribe to our biweekly newsletter. Visit our website. Follow us on LinkedIn.

Transcript

Unknown: 0:08

Welcome to radio PI everybody. I'm Nick D'Orazio from OSI soft, which is now part of a Viva. Our guest today are from Riverside public utilities department. And unlike a lot of folks in operations, they are doing digital transformation by making better use of their sensor based data. Now, what they're doing is they have to work closely with their Board of Public Utilities on expenditures. So they track it in a way that I think you're going to find very, very interesting. So let's start by the quick overview of RPU. Its provided water and electric service since 1895, in the Inland Empire region of Southern California, and their current service area population is more than 325,000. Our first guest is Jennifer Tagliani. She's a principal project Manor manager. Good morning, Jennifer. Good morning. And also as a co host today we have Gary Wong from OSI soft. He's an industry principal here. Good morning, Gary. Hi, Nick. Thanks for having me here. Well, thanks for Thanks for joining us, everybody. And we're going to introduce some more folks as we go along. Jennifer, can you describe briefly what you're doing with central bank's data? Sure. So it started for our utility rivers, I public utilities back in 2013. When we rolled out a strategic technology plan, it basically outlines strategic investments for us in new operational technologies. And we identified 22 operational technology projects with the intent of implementing them over a period of 10 years. So one of the most critical projects for us and foundational projects was the operational data management system. So that was really foundational for us to serve as a data hub or central repository, allowing us to collect and analyze and visualize operational data. And for us in effective odms basically was meant to manage large amount large amounts of data across multiple systems and workgroups. And then also help staff turn the data into actionable information to drive critical business decisions. For odms system, we selected the PI System. And so PI is actually that data hub that we were looking for. So we implemented PI in 2016. So July of 2016, we entered into an enterprise agreement, a five year enterprise agreement with OSI soft and we have been using the PI System as our basically our operational data management system to organize that data and make data driven decisions based on what were previously disparate systems now all feeding data into the PI System for analyzing that information. Okay, any ROI you can report. So yes, we actually have tracked very closely all of the dashboards and reports and information that we're now getting out of PI. So, again, we started this project in 2016. And in 2019, we basically had a three year report that we put together that track the ROI. So we basically went out to our end users made a list of all the dashboards and all the reports and all the efficiencies that they had gained from using PI. And we put that into a report and did some detailed analysis on it. And we found that at that time, our return on investment for the PI System was over 600, I'm sorry, $672,000 a year. Since then, we have continued to track our ROI. And we are about to present to our board a 2021, a five year report that now the investment or I'm sorry, the ROI has grown even further an additional $640,000 a year, which makes that total over $1.3 million a year that we have calculated as far as savings for the PI System. Wow, that's that's fantastic. Jennifer, I wish all our customers did that. Great job, too. Can you describe I mean, how what kind of things do you have to track to get that kind of accuracy? I'm curious. So I'll let Brandon go into a little more detail on that in a future response. But basically, we use data factors, primarily Redux action and stuff time. So the amount of times individual staff members were taking to complete tasks and so we captured time savings for each dashboard that we created in PI. And we multiply the hours saved per year by the average hourly rate of that South position responsible for the task and adjusted it by the payroll burden multiplier. So other quantitative benefits that have also been realized that are not calculated in the ROI, such as reduced Chuck roll to investigate problems, reduce operating costs and reduce paper waste. We didn't actually capture those numbers. So we actually think that the ROI is is higher than what we're reporting. We only reported really the staff savings in those hard numbers that we could find. There are also numerous qualitative benefits being derived from the system which are more, again more difficult to calculate as tangible savings and those are basically increased visibility into the system avoided costs for potential potential regulatory fines or lawsuits, reduce links to outages, fewer customer complaints, and and things like that, that really we couldn't calculate. But again, we feel like the ROI is much higher than that 1.3 million. So let's bring in now administrative analyst Brando Crozier, libretto. Hello, how you doing? Good. Thanks for joining us. So you were key player in kind of validating what that ROI was. The numbers that you submitted to this Public Utilities Board? Is my getting a Board of Public Utilities. Is that correct? Yeah. Yeah. Can you describe what kind of things that you were looking for? Um, well, a lot of things that we're looking for is just making sure that we were still improving operational efficiencies. A lot of what Jennifer had pointed out was reducing staff time. And if we could recognize operational costs, sometimes a lot of that was harder to put on paper, but we knew that we could at least share this important language of time save other things that we would look for, and we would still pursue if it was hard to, to quantify was good asset management or system reliability and the impacts those could have, we would record that and present that. But sometimes it just didn't come along with a number, but still has a lot of a lot of value. So do you have any advice you can give people who were, you know, just trying to do this on their own and want to build in kind of the interesting engagement that that you've got at RPO. So some things that we did, for gathering the data, sometimes it takes a lot of times of really working with people, some programs that we did, we created our own little training program within the utility to give people a good idea of what PI could do, what they could even do with PI. Easily things, you know, showed them DataLink, and we showed them PI Vision, kind of tried to show them these ways of empowering themselves. And some people really were able to take off with that others were at least able to see that it would get them excited about it. So then as they would bring up possible solutions. We could start asking them, you know, how long they spent doing that? From the beginning? I've definitely found through the personal experience, kind of asking those questions in the beginning has been able to get answers better than waiting until it was over. Because once it's over and they they're starting to have the new solution, then you can follow up and ask them, well, how long you spending doing that thing now. And usually, they become very excited in sharing that information. So you know, really empowering people. And then making sure that you keep open communication with them all the time, can really make this easier. Brando just wanted to know, was that way of thinking already there? Or did that type of culture was that fostered over time? The empowering people, you know, having them actively participate? That's not always easy to find, and utilities sometimes. So it sounds like that, the way that they're thinking the way you know, the whole culture was just kind of curious, was that already there? Or was that something that changed and grew over time? No, I feel that Jennifer really had come up with the idea that we should pass along this training and get people engaged, and I think, but as soon as that happened, our team, you know, Carrie was me and our consultants, we had really worked together to make sure to keep doing that. And as people started to ask questions, we would sometimes find an excuse to do it again. And I think that that has really, really helped build that culture. So we've watched it develop from this and been able to push it into other technologies that we've been trying to work with, which we'll talk about a little bit, I think later with Power BI and stuff like that. Gary, in addition to that, to answer your question, I think when we started this five years ago, we really were looking for groups or managers that were the most interested and had the biggest problems to solve. Right. So we started in Robin, we'll talk more about this. But we started in water operations where they had a huge problem to solve. They had a ton of data, they were manually calculating it, and they were the most interested in moving forward. And we also could see that if we were able to accomplish, you know, streamlining their workflows that they would have a story to tell others in the utility to get them excited about using PI as well. So we started with that, and that really trickled down to other managers seeing the benefit that that water operations Obtained and then really being interested in doing that for their own group. So it really started a trickle down effect. And then from that point on, we really prioritized our work based on value. So if we saw a problem that, for example, Ivan's group was spending 15 minutes every hour running a report and manually calculating in Excel, we made that high priority, because we saw the potential for that ROI, and that big savings for that group. So we really, since then have prioritized that, you know, the work that we do, because we're a limited number of people, we can only do so much. And especially in the beginning, there were, there was a ton of interest. Once they started seeing the value of PI, we prioritize the work based on where do we see the most value and the most time savings. So that helped us prioritize the work and also obtain that ROI that we we've seen any operational savings for the from the utility that that we attain, that gets passed on to our customers. So if we can reduce costs, we can, you know, operate more efficiently, we can reduce truck rolls, any of those savings get passed on to our customers, you know, in the form of, you know, not increasing rates in the future. So let me ask now about the savings. What is it about the savings that has allowed you to do, maybe stuff that you weren't able to do, and especially documenting the savings that way you did. So I want to bring in Ed Cortez, he's a principal electrical engineer, hello, ed. Hi, Nick, about the savings? Can you tell me a little bit about the odms system in general? How did it get started the operational data management system? Can you explain that a bit? Sure, sure. what occurred was that one of our departments or our departments had already purchased PI, electric electric group, which I belong to start investigating whether we should join in that venture. But as we moved along to see, you know, what the capabilities of with the PI software, we started brainstorming more into and on what other things can we do with this. So to put your perspective, RPU, like other utilities of our size, has electric water generation and other departments that use information systems, applications and databases to support our daily business processes? Well, what occurs is these applications and databases consist of commercial off the shelf applications that were sometimes outdated, and it was hard to get access to. We also had internal develop applications and spreadsheets and databases at remote locations. So by having all these different formats, they were not interchangeable, which made reporting, analyzing and just reviewing data very cumbersome. So the lack of a central data management system, delayed our decision making processes. So once we got PI, and were able to each group, look at their system and see what they needed. It just made the decision making process simpler and quicker. And we could, instead of waiting days or weeks to identify an issue, we can do it within an hour or quicker. Now I know some of our customers tell us they have a capital expenditure projects that they actually can defer. And I think last time we talked, you were telling me that the extra granularity allows you to kind of just kind of prioritize things better. Is that is that the case? Can you explain kind of how this aligns to what you're doing with capital expenditures. My group, their system planning group looks at the whole service site, electric system over a five year period to determine capital improvement projects. Well, the data we were looking at was just as I mentioned earlier, was just so cumbersome and hard to access that it made the analytics and decision making process very difficult. So we would only do it once a year, because of the time consuming effort to do it. Well, by introducing PI data sets, we're able to access that granular data quickly. And with very little effort by our engineers. So what we're doing now is our engineers are able to do their analytics and studies and be engineers instead of you know, spending a lot of time looking at data. So as far as time savings, it is more efficient, and we're able to make our decisions quicker and and better decision making instead The old way that we are dealing it, which is a lot of manual intensive efforts. Okay. And is there anything new that you can tell us about? I guess just the latest maybe what kind of return on investment you're getting? You don't have to talk about dollars, but just how is the value of real-time data to you now, as compared to say 2019? Last time we talked about this? Yes. So we're looking at it. Now, I just mentioned real-time, to where we're evaluating our system, as things conditions occur out in the field. So you know, during the summer months, we see a lot of solar PV activity on the system. And we were able to determine where the locations are, even though we don't have you know, the data, which homes are which businesses have solar. But with the data from PI, we can identify regions where we have high PV penetration and make improvements to the system accordingly. Okay, well, let's check on that same issue in water. We're going to talk now to Robin Glenny. She's a water quality management. Hello, Robin. Hello. So what were you able to what were you able to do with the sensors data to help help water quality. In water operations, we use PI in our blend model dashboard, we have about 50 groundwater wells. Some of them are treated through single stage or multiple stage processes of GAC ion exchange, or membrane filtration, all which blend in our three transmission mains, and then again at our 32 million gallon reservoir complex. And so the water quality blend coming in and going out to the distribution system is very important for compliance purposes, especially with psps events and power outages. And the blend model dashboard has really helped us in operations, manage what wells are turned on and off efficiently to make sure the water quality meets that final blend compliance. Okay, great. Now, I'm just kind of curious, did you find the system that you were working with easy to use, did you find people were able to self self serve when it came to finding things that they needed with regard to water quality? We did have an in house Water Quality Management database, but not very many people have access to it. And that's just the water quality data. And this dashboard, that's up real time for everybody to access. It takes that database through an API and integrates it with our SCADA flow meter data, and displays that for everybody to see. So there's, I think it's over 500 calculations running real time in the background for this dashboard that's available for all levels of management and whoever's interested. When you say dashboard, I assume it's on the web. It's on mobile devices, pads, stuff like that. Correct. Okay, yeah. Well, it's just, you know, nowadays, it's so ubiquitous, but you have to ask, because it's me, that's, that's the way people want it. Yeah. I've got a question for Robin, just to follow up there. So Robin, I was wondering, do you report those figures to the EPA or to a governing body automatically? Or is that something in the future. So we have several different reports that we have to send to the State Water Resources Control Board on a monthly basis for samples collected the previous month. And it was a very manual process. Previously, we've been able to automate, I would say, 98% of our reports with PI DataLink. So just updating the date range, everything populates automatically into it. And then we just do a quick QC check, add some signatures, and they're ready to be emailed out to the state. Thank you so much, Robin. So I'd like to introduce now Kerry Tilden, he's a senior resource analyst. And I wanted to ask you, there's some fairly low hanging fruit that people typically get when they have better access to their data. Can you describe some of the simple things people have been able to do to improve what they're doing? Yeah, that's right. So you know, we have this training program that we've been running where we've, you know, identified users from different departments and brought them in and gave them training on sort of basics, you know, hoping that they could build on that, or at the very least understand more about the technology. And so, you know, we've had some really great successes out of that we've had people building complicated Excel reports, or, you know, even visual displays, we had one user who, in order to kind of practice from the training, he put together a display with call center data, which was the group that he was in, and it turned out to be such a good display that we used it to replace the one we had built and put it up on the screens for all the other customer service for reps to see. So, you know, we were extremely excited about that. And we love when we get a power user like that, right? Who will kind of take it and run with it, build tools that, you know, we don't have to build for them, right? A brand new and carry the enterprise agreement licensing model is unique. Can you let us know how that's made a difference? Yeah, we're able to do the self paced training, which is a lot of training that comes from OSI soft, so we'll get things we'll send people that direction. And, and because of our enterprise agreement we have that is basically free now. And as we send people to things like that, and they start learning and, and we start learning about other technologies or tools that are that are part of PI, because of the Enterprise Agreement. A lot of these things are readily available. And as soon as they start asking about it, then it's just a matter of applying it to their question in there and their concern. And so I think that's been a huge benefit it, it reduces a lot of this downtime, like they showed interest, and we get back to them a month later, no, we can get back to them immediately. And that's really awesome. Yeah, so you know, having the EA is really helpful, not just for getting access to the software, but also, you know, the whole suite of training that comes along with it, we can provide that to our people, we can also use it for us, right, you know, we are fairly knowledgeable in the system, but we're not experts necessarily. And so when, especially, you know, someone will bring us a solution or a possible solution, and we won't necessarily know how to solve it right away, but we can access the training material for our own uses. And additionally, the support is really great. And so when we reach, you know, a dark corner of the system that we don't fully understand, we can kind of call in the experts and get, you know, get some really great support and helping us solve problems and build new solutions and, you know, move our own knowledge forward. That's great. So one of the systems you're working with, is putting data on ArcGIS maps. And to talk about that we're gonna bring in, let's introduce Mark Smith, he's a utilities principal analyst. And also, Wes was new ski senior analyst. So, either of you, I guess, if I could ask you how has putting sensor based data on the maps actually helped. I'll go ahead and start. So about 10 years ago, we began the effort to really industrialize our data, to try and organize it to get it out from this unstructured environment. It was within a SQL database. But we just we weren't doing a great job of organizing it. And so we spent a lot of time getting our data organized. And that stepped us into custom reporting. And we eliminated our paper based process off of that custom reporting. And then, you know, with every solution, you start to recognize that yeah, this is great, this is a great step forward. But you know, when things got busy, right, if I have 6070 locations that are leaking and need to be repaired, or it's a Monday morning, and we have a lot of phone calls coming in, list based data starts to, you know, become inefficient, right, you're starting to pick and choose and henpecked stuff out. And so we've kicked around this idea for a long time about, hey, this data is nice and structured, right? I got addresses on everything, you know, what would happen if we, you know, plot these on a map? geospatially, right, we, we feel like we could really use the benefits, especially when things get busy. And it's, it's been a game changer for us. We've recognized a lot of efficiencies, just purely representing the data and in a different function to be able to see our work. As it comes in streamlined. My field staff owns and manages 100% of that data from the start of the process to the finish of it. And to see their efforts represented in a clean, concise, and effective manner, I think speaks well for our entire division. And like I said, just has produced all kinds of efficiencies from, you know, where crews are located in relation to issues coming up to customer complaints, to, you know, being able to pick out your day and not spend your time driving from one side of the city to the other and realize like, hey, these three are right here, they may not be of the highest priority, but as far as our time efficiency is concerned, let's focus our efforts here today and get these knocked out. And, you know, that is really extrapolated into a bunch of different avenues for us, we were able to build pre emptive replacement program, just being able to view our data spatially. we've extended it out into Other mapping for our Public Works departments. And you know, we're really working off this synergy associated with geospatial data and allowing our customer service center to see it. And so as calls are coming in, instead of having to spend the time to get ahold of somebody, or try and send an email, they can pull this data see it just like we do, they can see the notes, they can see the progress, they can see when things are scheduled. And it's really eliminate a lot of this back and forth. And it's really allowed my group to focus on the most important thing, which is our customer service response. You know, it's is difficult, right? Your water goes out, it's 108, outside, right? It's, you know, it's something that needs to be fixed quickly. And so we recognize that it's created huge efficiencies for us. And really, you know, it's allowed us to better serve our customers. Okay, so Wes, tell me what the what technology do you have in place to make all this work? Yeah, we use the PI Integrator. For Esri to stream live data to the maps, we use either ArcGIS Online or we have our in house portal server, historically, they would use paper based maps that works. Static, you know, and now they're having, you know, web maps that are live and automated. And then near real time, you know, just the time efficiencies of staff being out in the field and physically sitting on a map, you know, where your new workers. So they used, they used to see this in a tabular data now that they can just see it on the map. That's what you're talking about. ACC correct. They use a laptop or tablet. Thank you, gentlemen. So let's talk to Ivan valeska. Now, he's a power scheduler slash trader. And I wanted to talk about your market operations system for trading energy, can you just describe what that is? Absolute, it's basically just an accounting system that is run every hour, we look into the load for our system, we look in to the inputs into our system. And we look into any trades with other entities, whether they're bilateral or third party, and and we run an analysis. Where do we stand? Are we short on energy? Are we long in energy? And what do we do on that following now? Or do we let the market take care of it? Or can we shorten the position to be in a better financial outcome? What improvements have you been able to make and how you do this over the years? Yeah, so when I started working with market operations, everything was done manually. So at the beginning of the hour, we had to go to different software's and grab load data, we had to grab import data, we had to grab weather data and see what else if we had any generation online, we had to grab generation from the different generators that we we own. And that could take quite a significant amount. When we did the analysis, it was taking about 15 minutes in on average, to get all that data. And for a scheduler on a afternoon, noon, and a summer day, we don't have 15 minutes, especially when when you have an outage at one of the intertype or with one of your generators, you do not have 15 minutes, because you should be looking for energy to replace that that outage. And so it was very stressful, very time consuming. And once we were ever to implement data link, and all of our different data that we use into the spreadsheet that cut the time of have are inputted by vital 15 minutes, either there's still some some data that has to be manually inputted, but it's a very minimum. And it's helped us greatly, especially with load. Are you looking at predictive analytics in terms of what the demands going to be to help with the market operations and trading? Yeah, so that's something that that's been discussed by utilizing Power BI in the future, and also trying to get away from from the spreadsheets, and just doing all of our operations and accounting in Power BI. So it's something in the long run. Well, great, well, ours almost up thank you so much. Before we move on, what is what's next? What are what are your future plans anything? So our biggest priority right now is rolling out ami. And looking at ways we can use PI to do that efficiently. So right now we're we're just getting finished with testing and ready to go live. And one of the ways we're using PI is by gathering data from the ami headend system and the meter data management system in PI to validate calculations and just kind of do that cross check during testing. And then once we go live, we're going to be looking at PI and relying on PI to utilize operational efficiencies and our outage management and how we can pull data from the AMI system. And now there's, you know, a lot more data than there used to be, and how we can provide that through PI and useful dashboards to both engineering for system planning. And then also the operations for responding to outages and you know, work that's going on, or issues going on in the field. As this helped in just your better use of the data, has it helped in any larger organizational goals? We hear people talk about sustainability goals, or digital transformation, your specific digital transformation projects branded your own way, has it been helpful in those types of things? Oh, definitely. I mean, we're always looking for ways to be more efficient to be more resilient for that sustainability. So you know, PI is a huge, like we said, In the beginning, operational data management system, it's foundational for us to, you know, move in the right direction in all those areas. Right. Also, one more thing we love to talk with people about nowadays, what have you been doing there may be different or helpful during the pandemic that you were able to accomplish with better access to your data? Yeah, so for us in market operations, one of the things that took a huge impact is that before we were able to have two traders, and on the trading floor, after COVID, it was down to one, and we had to spread the other ones around. And so by having PI, we didn't have to physically exchange any, any schedule for any particular trading log. It all could be done through through PI, and very easily accessed from multiple devices at the same time. Nice. And for us and water operations, we get a lot of data requests that come from our SCADA information. And to create space in the office, we were telecommuting, or are just trying to stay out of the office as much as possible to allow space for some of our operators and other staff that couldn't telecommute. So I was able to utilize PI DataLink to fulfill a lot of those requests. Whereas if we didn't have PI would have to physically go in. Well, I appreciate it. Thank you so much. Thanks for filling us in. And thanks for taking the time to answer all these questions. Now, I want to spring something on you. I haven't talked to any of you about this yet. But we love doing a lightning round at the end of these interviews, where we talk about the things that people only an operations can appreciate the cultural things that go with working in operations. Sure. Is that okay with everybody? Okay, great. Now, I gotta ask you, you're all a lot of you are in operations been there forever. Like a lot of folks in operations do you have on your desk or somewhere a memento, which is like a broken valve or a burned out motherboard or something like that, that you keep for sentimental reasons, or just as a remembrance? Looks like Mark has an answer. I mean, my entire office is like, broken computers and messed up pieces of copper and broken valves. I think my foot is on a hydrant check valve down here. So yeah. Thanks. Great paperweight. Oh, that's great. Well, we didn't we didn't see your thanks for Mark, thank you for sharing. We didn't see your face during that. But thank you for sharing about this. In operations, this is something that I don't think anybody except folks in operations now is you get to see sites or you just don't see anywhere else. And I imagine, you know, some of you have done some nice work and some pretty disparate, you know, just bizarre places, all around your physical facility. What's the most interesting scene you've ever seen? Or most interesting view you've ever seen? I definitely have one for that. Because that was that speakings that, Ivan? smart. Oh, go ahead mark. About three and a half years ago, our underground water vaults. The majority of what we put in the ground are traffic rated, but we have some older ones that we've been working to replace, that aren't really close to traffic lanes, but are were put in there. They're fiberglass and they're not traffic rated. In an area of our city, which is not agricultural whatsoever. We had a horse fall into our vault that required the fire department to come out with a hoist to lift the horse out of the vault. And it has accelerated our replacement of fiberglass vaults. Quite a scene made the paper seven CHANNEL SEVEN news was an interesting day. I have a nice what's that something that's not nearly as exciting as what Mark just shared. But I know in our electric meter shop, we have a meter on display that was in a in a house fire and it was it's there because the components all held up, just the plastic cover melted, but it melted in the shape of the screen mask. And so it's kind of exciting. Okay, well speaking of horses, then maybe you know, as a public utility, maybe this isn't the best of questions, but I haven't seen an operating area yet. That didn't have some strange wildlife that wanders on site all the time, and just bedevils the people who are working here. What's the most common animal that you see? Just kind of getting into the into the works of what you all do for a living? I don't know. I've never seen this, but I do know that we often have poor little squirrels who give their lives upon. Goodness, it's even a cause in our outage management system. Okay. Okay, great. Well, again, I can't thank everybody enough. We've been talking to the city of Riverside public utilities department, folks, I really appreciate you sharing this information with everybody. And thanks again for joining us. Thank you for having Thank you. Thank you. Thank you. And thank you all for joining us. We'll see you in another two weeks. Bye bye.