How do utilities manage risk using Machine Learning | Technology

How do utilities manage risk using Machine Learning | Technology
How do utilities manage risk using Machine Learning 

How Machine Learning is being used by utilities to manage risk

Using machine learning to manage risk is what utilities utilities use to increase security, improve regulatory compliance, manage costs, and improve service for their customers.

Driven by increasing compliance requirements, increasing customer expectations, and growing concerns over safety and grid reliability, utility companies are seeking innovative solutions to more effectively manage risk. They know the answers lie in their data—the challenge lies in finding cost-effective ways to streamline the data for them. This is where Machine Learning (ML) comes in.

Rapid adoption of Artificial Intelligence (AI)-powered solutions, especially machine learning, enables utilities to harness the power of their data in mission-critical decision making. When utility companies apply machine learning to manage risk, they are better equipped to meet the challenges of the current environment and be prepared for the challenges ahead.

New Disruptions, New Challenges

While some utilities are known to be early adopters of the latest innovations, some recent trends have forced even the most reluctant providers to rethink their approach to AI-powered automation.

Climate Change Events

As climate change continues to cause extreme weather events, utilities must adapt to the resulting demand spikes, outage risks, fire hazards and other challenges. The western United States is enduring a multi-year drought – at least half of the past 20 years have been longer-than-average droughts – with increased demand for water utilities and increased fire risks. In California alone, 8,835 fires were recorded in 2021, burning 2.5 million acres across the state. 2021 also saw a record-setting heat wave on the West Coast and Hurricane Ida, which left more than 1 million residents without electricity. In February 2022, Winter Storm Landon swept through much of the country, causing power outages in more than 330,000 homes and businesses in several states.

While the need to predict and prepare for severe weather is nothing new, the increasing frequency and intensity of climate crises is calling for the search for more effective solutions.

Renewable Energy Driver

To combat climate change, the Biden administration has called for an end to carbon emissions from the power sector by 2035 and to reach economy-wide net-zero emissions by 2050. On the consumer side, 2021 saw sales of electric vehicles (EVs). ) jump 83 percent and hybrid sales climb 76 percent, leading to both an increase in demand for public charging stations and pressure on the electric grid.

Increasing Compliance Requirements

The already highly regulated utility industry is facing a steady stream of new regulations, and the cost of keeping pace—in terms of both time and dollars—keeps rising. In the area of ​​safety compliance, for example, California electric utilities are required to submit an annual wildfire mitigation plan (WMP) detailing how they are building, maintaining and operating the equipment. to reduce the risk of forest fires.

This reporting requirement is much more than a simple rubber-stamp exercise—as one of our clients discovered when the California Public Utilities Commission (CPUC) requested additional data to support claims in its WMP. Fortunately, the utility was backed by the Logic20/20 machine learning teams and was able to respond to requests quickly and accurately (see "Utilities Analytics Case Studies" below).

Why Machine Learning is important for Utilities 

Most utilities have more than enough data to make informed, timely decisions to support risk management. Drones, stationary cameras, smart sensors, other IoT resources, and human reporting build up to massive amounts of real-time data almost every hour of every day. However, this bounty of data is only as valuable as the utility's ability to do something with it – to analyze the information and identify irregularities that warrant investigation. This is an impossible task for even the most knowledgeable, experienced human workers. But for machine learning technology, it's all in a day's work.

Historical data enables ML solutions to "learn" acceptable parameters, such as the amount of wear that can appear on a piece of equipment before it becomes a failure risk or the amount of vegetation that is exposed to electricity prior to the risk of fire. May trespass on the line. The system can then evaluate the data as it streams and flag irregularities that fall outside these parameters, which can be analyzed and detected by human response teams. As more irregularities are defined and addressed, this information is returned to the system's historical data, making the platform more efficient over time.

Use cases for Utility Risk Management 

Plant Management 

Using machine learning, utilities can gather data from drone footage, climatic conditions, observation records, and tree species profiles into training machine learning models to assign risk scores to each individual tree. These models can be used to prioritize trees for inspection, recommend optimal trimming clearances, and estimate the number of risk events prevented by well-planned trimming.

Predictive Asset Management 

Machine learning solutions can take massive amounts of equipment-related data – including brands/suppliers, installation dates, weather conditions, maintenance records and historical records of failures – to predict when a specific piece of equipment will fail. is at risk of.

Rapid Threat Identification

With machine learning, utilities can review thousands of records from scores of drone footage, lidar, GIS mapping, and other sources to identify irregularities that could indicate security threats. Problematic records can be automatically flagged for closer analysis and treatment by human response teams.

Weather based outage and demand forecast

Historical weather data can be combined with usage patterns, outage reports, and other data to help machine learning solutions identify weather patterns that are likely to lead to outages, increased demand, and other situations.

Maintenance Priority

Traditionally, many utilities have maintained their equipment according to a set schedule - or have responded to equipment failure. Machine learning enables them to prioritize assets that are at highest risk of failure according to their age, surrounding weather conditions, and other factors. By giving top priority to the highest-risk assets, utilities can avoid failures that lead to outages and fire hazards—and possibly even extend the lifespan of their equipment.

Guiding Grid Safety Precautions

During extreme weather events, it may be necessary to turn off parts of the electrical grid as a safety precaution. Machine learning provides the quick insights and forecasts utilities need to identify and prioritize areas at highest risk for security issues in rapidly changing conditions and plan shutdowns accordingly.

Benefits of Machine Learning in Utility Risk Management 

Because it combines an enormous potential for reviewing data and the ability to continuously learn from the results, machine learning is ideally suited for utilities that want to improve their risk management practices in a cost-effective manner. Utility companies that are incorporating machine learning into their operations are reaping a range of benefits.

Enhanced Security 

Machine learning enables utilities to quickly identify and address potential security threats, before they can cause harm or put communities at risk.

Better Compliance 

Thanks to machine learning, utility companies can more easily gather the data needed to respond to compliance reporting requirements and ensure consistent alignment with regulatory guidelines.

Better Cost Management

Utilities can take advantage of machine learning to save hundreds of hours on manual inspections, maintenance reports, and other tasks traditionally performed by human employees.

Better Customer Satisfaction

Every time an outage can be avoided, utility companies have an opportunity to increase satisfaction and trust among their customers.

Machine Learning Tools for Effective Risk Management 

Technology products alone won't guarantee optimal results in machine learning—but it's hard to be successful without the right combination of the right solutions.

At Logic20/20, we take a technology-agnostic approach to machine learning for our customers, tailoring each solution to the specific needs of the organization. Over the years, we have used various combinations of Palantir Foundry, AWS solutions (Amazon Athena, AWS Glu, Amazon SageMaker), Google Cloud products (Vertex AI, AutoML, AI Infrastructure), and Azure Machine Learning. Each solution offers its own benefits, so it is important that we get a deep understanding of the current state and end objectives of the organization before exploring specific technologies.

Utilities Analytics Case Studies

Logic20/20 has the opportunity to support multiple utility companies in their mission to harness the power of machine learning for their risk management initiatives.

Vegetation Management Plan Verification

The State Utility Commission required a major California utility to prove that their vegetation management program was effective in reducing fire risk. Our team has developed a machine learning model to analyze program results. After analyzing more than 12.5 million records – incorporating factors such as line clearance distance and the location of genus and trees – we delivered an analysis that provides essential information for a data-driven, actionable wildfire mitigation plan.

Asset Management

Identifying asset (equipment) risks is critical to maintaining grid security; However, the task of continuously monitoring the condition of every instrument over a distance of hundreds of miles is beyond the realm of human capabilities. For a power provider serving over 3 million customers, we have developed Deep Neural Networks to detect asset issues from drone footage, satellite images, lidar and inspector photos, which issue capture rates using image detection. Improves and minimizes inspection costs.

Looking Ahead

As more utilities adopt machine learning, they are able to go beyond traditional risk management (controlling the various things that can go wrong) in order to achieve real flexibility - handling a wide range of situations quickly Be prepared to recognize and respond.

While technology is making this development possible, utilities must recognize the change in mindset required for a successful machine learning strategy. Utility companies have historically operated according to fixed plans, addressing risk management through rules and regulations. In an environment that is rapidly moving and fluid, this approach can only go so far. By using daily data to solve problems before they become problems, utilities can gain the agility they need to protect their customers, control costs, manage compliance and provide more reliable service to their customers in a rapidly changing environment.

Source: T&D World, Direct News 99