Reliable power is mission critical for electric utilities. They manage several independent siloed software systems and database to provide reliable power to their customers. General managers and engineers are constantly responding to power reliability issues. As utilities progressed, they have legacy systems and different outage codes for issues based on the system that was in place at the time. Therefore, the standardization and the utilization of data is more difficult than most would expect.
It is difficult to know which systems provide the results that utilities need to survive in the new economy, especially with the effect of Covid and climate change. The type of system that utilities need is one that predicts what is going to happen before it happens. However, there is a lot more that has to happen before that can begin. These 5 steps will get you started in the right direction.
1. Identify the people and process
2. Fix the data first
3. Set up a secure and real-time process to see and extract the data
4. Generate standard reports to check and verify the process
5. Use AI tools to get ahead of future issues
Identify the people and process
It is important for everyone in the company to become data driven. Using data to do your job will save tons of time and money. Organizations must ensure that there is standard operating procedure in reviewing the data. This must be reviewed by the team to make sure that everyone becomes a data advocate.
Utilities are working with companies like Aerinet Solutions to document their processes and ensure that the data flows easily, securely and automatically. More importantly, companies like Aerinet Solutions works closely with the utility team to show them how to make use of the various data coming from independent systems. This way they can see how the data is used to reduce costs and increase revenue.
Fix the Data First
Most staff do not have the time nor the experience to fix their data issues. Therefore, getting to use new technologies like Artificial Intelligence (AI) and Machine Learning Algorithms are usually not going to happen until the data is extracted, cleaned and loaded. If you use poor data, you are prone to garbage-in and garbage-out (GIGO) syndrome. Who wants to have unreliable and untrusted reports that can cause millions of dollars due to incorrect business decisions?
Find a trusted partner who provides an easy, fast and cost-effective way to extract, clean and load your data. Take a look at this easy spreadsheet to see how you can easily get started with loading up your outage data. Cooperatives are uploading their data as part of the NRECA Reliability benchmarking study and are quickly yielding returns on seeing their data compared to other utility systems.
Set up a secure real-time process to see and extract the data
Most data solutions are usually not as easy to manage in real-time. Several of the older systems have delays using their data. Electric cooperatives, who work with the National Rural Electric Cooperative Association (NRECA), use the MultiSpeak data interoperability standard. MultiSpeak is used in the daily operations of more than 800 electric cooperatives, investor-owned utilities, municipals, and public power districts in at least 21 different countries. Through MultiSpeak, PwrMetrix makes it easy to connect your data in real-time, predict future outages, quantify costs and most importantly, pinpoint the “best bang for the buck” or “Return of Investment ROI” when planning for preventive maintenance. Even if you do not use MultiSpeak, your platform should easily accept the unstructured data. The best platforms are open and standard agnostic.
Generate your standard reports to check your process
Most companies do not know how to use specific data to get the reports they want. When starting your project, start small with some standard reports. It is best to use a trusted partner with experience in your industry. Once you become more comfortable with the data, you can identify significant savings from the data you are collecting.
Data lakes allow for unstructured data, but many times there is too much data. Platforms that allow you to import unstructured data in a systematic way yield higher results. For example, data from substations, feeders, smart meters, field devices and different cloud systems can easily be ingested and used within the PwrMetrix system. Overlaying GIS (Geographic Information Systems) maps allow you to see the worst feeder, substations and devices by duration and cost within seconds instead of days or weeks. Most importantly, showing how it relates to the reliability and resiliency of the system is paramount. Most visualizing systems merely give you a chart and do not provide accurate maps or detailed relational data. Other platforms say they report real-time outages but are quick to say that most of the data is spotty and old based on what is publicly available.
Having a platform that can show you real-time electric reliability makes the data more relevant to the users. When starting your project start small with some standard report and mapping. Once you get more comfortable with the data, you can apply more advance tools to show predictive (What will happen?) and prescriptive (How can I make it happen?). Cooperatives are using advanced AI tools to reduce losses of revenues and find issues before they become extreme problems.
Use AI tools to get ahead of future issues
Chris Hamon, CEO of White River Valley Electric Cooperative Missouri, explains how his team maximizes the power of PwrMetrix AI to predict top feeders to have power outages based on the time of the month and certain weather condition. This allows White River Valley Electric Cooperative to plan their future, perform predictive maintenance and limit future outages. Chris says, “when we ran the outage predictions on the AI platform and it gave us the probabilities of outages for the next month, we were a little skeptical until in the first week after all top 5 predicted feeders went out. Literally it picked all five!!!”
White River Valley is taking the AI report a step further by “looking at minimizing outage length or even keeping the outage from happening at all by taking the AI get probabilities and use it to work smarter by changing breaker settings and minimizing the effect of the outage.” White River Valley engineers are taking lightning strike data to look for damaged lightning arrestors in areas where outages happened that had no probable cause.
Chris further explains, “once we get the fault current info, then we lay it over the digitized map and get the location of the nearest lightning arrestor to the lightning strike. We then send personnel to the site to check out the arrestor in question. We are 2 for 2 using this method and not letting time take the arrestor out on its own. We are using science and artificial intelligence to improve reliability and reduce loses of revenue during outages. Again, working smarter and not harder, and reducing incremental costs while reducing revenue losses.”
What an Effective Data Solution Does For Your Business?
So, the best data platform should grow as you go through your data journey and progress to actionable insights or in AI terms called “Prescriptive Analysis”. Buying a system that is too big and daunting is one of the reasons most companies’ data programs fail. It is better and cost-effective to find a partner that is the right fit. Start with a simple spreadsheet to get your team going. Try some basic reporting and once comfortable apply the AI tools to predict next steps.
White River Valley started out slow with PwrMetrix and kept adding the additional features and methods to get amazing results. Having the best data solution shouldn’t be daunting, but rather exciting to quickly see returns of your efforts.
Click here to see how Aerinet can help you today.