When big problems grow on trees, you need AI for vegetation management
Overgrowth near utility equipment is no idle worry.
Summer heralds the start of warm weather and brings with it opportunities for the public to savor the outdoors. But the season creates problems for public utilities, as the blossoming trees, plants and brushes present concerns about vegetation management.
Overgrowth near utility equipment is no idle worry. Extreme weather events stretch from spring until fall, and any excess of dry vegetation near equipment increases the risk of spreading wildfires. Utilities are on the hot seat for their role in spreading wildfires: Equipment owned by California’s three largest utilities, for example, ignited more than 2,000 fires between 2014 and 2017, the Los Angeles Times reports. Many large utilities are facing lawsuits, regulatory fines and compulsory government oversight, and they need to quickly and drastically improve the way they handle nature.
Utilities should look to artificial intelligence to modernize their vegetation management efforts. AI-based technologies can pinpoint troublesome vegetation areas near utility equipment, and merge historic and real-time weather data to offer clear snapshots of otherwise obscure trouble spots. But AI goes beyond utility management by providing predictive analytics that can reveal impending supply and demands issues and potential equipment failures.
By implementing AI, utilities can have the comprehensive insight they’ll need to satisfy customers and regulators, better manage their equipment and preserve assets, and handle the whims of nature.
Rich insight improves the management of vegetation
Consider the Pacific Gas & Electric Co., California’s largest electric utility, which has been named in a string of lawsuits concerning recent Northern California wildfires. A California Public Utilities Commission report from February 2019 found that nearly half of the fires attributed to PG&E were caused by vegetation coming into contact with power lines. PG&E said that it would remove 375,000 trees and maintain 2,450 circuit miles of vegetation in 2019, but in April, the CPUC said the utility should use more metrics to analyze the effectiveness of its vegetation management and wildfire prevention efforts.
The struggles of utilities companies are a stark reminder that all utilities could follow the CPUC’s suggestion and use analytics to address risk, particularly during the spring and summer, when vegetation growth can most efficiently be maintained before typically stormy fall.
Utilities are relying on human observation and drone imaging to stay ahead of vegetation growth, but neither can effectively reach every square inch of flora. AI-based visual recognition technology can analyze satellite imagery to determine the height and width of vegetation and how wet or dry it is. It can also be used to assess how close vegetation is to equipment, allowing utilities to judge the risk of a fire or if overhanging branches could cause an outage during a severe wind event. By receiving this data in increments as short as 15 minutes, companies can act fast to prevent trouble, enacting operational safeguards and alerting incident management teams, customers and, if necessary, emergency response units.
AI enhances distributed resource management
AI can do more than detect and help prevent the risks of unmanaged vegetation. It can also empower utilities to rely on several energy sources and be in the best possible position to ensure that customers’ diverse demands do not exceed supply.
Many utilities want to move beyond a reliance on fossil-fuel energy and generate energy through renewable sources such as wind and solar power. But distributed energy generation is tricky. When businesses and residences generate their electricity using renewable sources, utilities absorb whatever isn’t used. It can be challenging for utilities to get a clear reading of how much energy is available to offset customer demand, however, due to the lack of analytical power in their current systems.
AI-driven analytics can assess in real time the expected output of renewable energy sources, then combine that data with output from traditional sources to forecast short- and long-term demand. AI can also evaluate detailed weather forecasts to predict voltage fluctuations on a circuit-by-circuit basis and review historical data to make sense of usage patterns. This comprehensive study of supply and demand would enable utilities to automatically and confidently move power through the grid and consistently satisfy customer expectations.
Data reveals ways to optimize equipment and infrastructure
AI can also help clarify another operational aspect that’s difficult to assess: equipment reliability. To keep energy flowing and customers happy, utilities must consistently optimize the performance of their assets. But it’s not easy to detect wear and tear on equipment and infrastructure, nor is it easy to glean the cause of seemingly inexplicable failures of new technologies. By outfitting equipment with AI-driven tools such as IoT sensors, utilities can continuously monitor equipment to stay ahead of deterioration and developing problems. Many utilities are already relying on data algorithms to spot unusual patterns in combustion turbines in energy plants and outdoor wind turbines. The algorithms compare industry failure rates for specific equipment with actual usage and other factors, such as exposure to weather.
Long-term usage analysis also lays out ways that equipment can be more efficiently managed, which reduces the chance of failure and widespread outages. Knowing that a turbine might fail in five years allows a utility to budget and plan for eventual repair and replacement.
The insight gleaned from data generated by AI-supported technologies enables utilities to make the most of workflow, maintenance and portfolio management. By improving vegetation management, distributed resource management and asset optimization, utilities can be more responsive to customer demands and take the necessary steps to avoid the types of systemic failures that have plagued some companies.