Few industries are as asset-intensive as utilities. And even fewer undergo the public scrutiny utilities face when an asset fails and a service interruption ensues. Utilities work hard to provide highly reliable service, and their success at doing so has painted them into a corner of sorts: customers have come to expect service reliability, and they have little tolerance for the inconvenience of doing without for more than very short periods of time.
Even when major storms destroy significant amounts of utility infrastructure, health, safety and high public expectations put significant pressure on utilities to quickly return service to normal.
The recent California wildfires were the worst in the history of a state that has seen more than its share of fire disasters. Utility asset problems, this time in the form of sparks from frayed transmission lines, are suspected of contributing to the fire’s onset. California Public Utilities Commission President Michael Picker recently compelled the state’s utilities to operate under enhanced rules to reduce risk, introducing a new policy that “includes significant new fire prevention rules for utility poles and wires.”
Across the country, utility-owned assets are strained by storms that are noticeably more intense as a result of climate change. Hurricanes, tornadoes, snowstorms, windstorms, and firestorms are all getting stronger and more frequent. Still, the public’s expectation that critical utility services will persist through a storm event, or if not, be restored quickly, has not changed.
Utilities have traditionally monitored asset health by conducting inspections and performing routine maintenance at fixed time intervals as recommended by equipment manufacturers. Resource limitations, tight budgets, and competing priorities often mean that some number of these preventive maintenance jobs (“PMs”) are done late or not at all due to higher priority issues. Missing an occasional PM—especially an inspection—isn’t seen as a huge problem because very few routine inspections actually identify a problem that requires fixing anyway.
Data analytics can be used to both improve asset maintenance and reduce unproductive inspections. Gathering, integrating, and analyzing multiple types of data can lead to better identification of potential hazards and can alert crews about the need to inspect assets when the data reveals the existence of a potential problem. Analytics minimize both the false positives, i.e. inspecting an asset that is performing normally, and the false negatives, i.e. not inspecting an asset that is about to fail. This way, equipment problems may well be prevented while simultaneously minimizing the cost of inspections. Analytics can also highlight assets that, while performing well during normal conditions, are prone to fail during a storm. This knowledge alerts the utility about opportunities to preposition crews and spare parts that are necessary to best prepare for potential failures during storms.
Similarly, analytics can provide more accurate predictions of which assets are likely to fail or cause problems in fire-prone areas. We have identified two particularly fruitful areas to explore: locating weak poles and identifying frayed wires.
A Data–Driven Approach to Fire Prevention
As we have seen with the recent tragic fires in California, a single spark in a remote area can result in tremendous damage and devastation, which raises the importance of preventative steps. Analytics can help identify weak, failure-prone poles, and frayed lines, especially in the remote areas that are often most susceptible to fire from a small ignition source and which are the hardest to reach for maintenance and inspection. Crews can be dispatched to these remote areas with confidence to proactively inspect and fix problems during times of heightened fire danger.
A variety of data can be used to help pinpoint where preventive maintenance activities are likely to be the most productive. Facts about pole type, age, size, load, ‘lean,’ and condition can be merged with data about soil type and condition, rainfall amount, pole rating from last inspection, and other factors that impact lifespan and durability. This information can be used to predict when potential issues will arise with individual poles and wires. That information can be further refined by merging it with data related to local fire danger, vegetation type, and combustibility factor. In this way, poles and wires in remote, fire-prone areas determined to be at the highest risk for failure can be identified and repaired before tragedy strikes.
Look for Frayed Wires and Tree Problems
Another approach to finding potential problems is to apply a machine learning model. A deep neural network (DNN) can be trained to process images captured from drone video of wires, poles, and trees in remote areas. The model can learn to identify frayed or loose wires, damaged poles, and tree branches that are in danger of falling or blowing into wires. Footage of the items identified as potential hazards is then pushed to humans for review and appropriate action can be taken.
Preventive asset management and the types of analysis discussed here both require a unified view of data that comes in multiple formats from a variety of sources. Many of our customers use a mix of vendor technologies for their utility systems, usually SAP and Oracle. HEXstream can simplify the problem of integrating these and other disparate data sources. Since SAP data lives on a remote island called HANA, it is not easy to combine with other enterprise data needed to conduct a broad analysis. One alternative is to bring the non-SAP data into HANA; however, customers with SAP platforms who have used HANA as the centralized data hub have learned the hard way that it is a very expensive option with limited scalability. To address this problem, HEXstream developed a solution called Harmonia™ that pulls data from SAP and non-SAP sources in real-time and batch and persists it into any platform, including Oracle, either on premise or in the cloud.
Unlike fires, which can change directions quickly, large storms have fairly predictable paths. HEXstream also has a solution called Argos™ that provides real-time asset health data and predicts which assets are likely to fail by matching asset conditions with the upcoming storm’s path, intensity and other data. It understands crew placement and spare parts locations, and it provides suggestions on when and where to move the crew with the right skills and parts to provide the most timely repairs in case of asset failure during a storm.
As the old saying goes, an ounce of prevention is worth a pound of cure. Awareness, preparation, and prevention are keys to optimal asset health. Helping companies access, integrate, and derive value from their data—including helping them see and solve problems before they happen—is our business. Please contact us with questions and for more information.
About the Author
Bill Silkett is the Director of Utility Sales at HEXstream. He has over 30 years of experience in the Utility industry spanning sales consulting and CIS project management. A Southern California native, Bill received his BS in Business and Computer Science from San Francisco State University. In his free time, he enjoys skiing and biking, as well as reading novels by Michael Connelly and John Sandford.