The utility sector is heavily reliant on equipment, mobile devices, and the Internet of Things (IoT). When properly mined and harnessed, these data sources can streamline operations and optimize output, enabling companies to operate more effectively and increase profitability. 

In today’s post, we’ll examine three major ways that AI can help utility-based organizations achieve these benefits. 

Forecasting Energy Needs with Predictive Analytics

The Electric Power Research Institute, or EPRI, is working with utility companies and the AI community to create data sets for developing and training models. These data sets can then be used to increase efficiency, enhance predictive modeling, and lead to better, more efficient identification of damaged equipment that may need to be repaired or replaced. Utility companies can also use predictive analytics to plan for future customer demand, enabling them to shift from reactive decision-making to more proactive and preventive strategies. 

Improving Resource Management Within the Energy Sector 

EPRI is also developing models and tools that will allow operators to be more responsive and flexible to utility grid signals, ultimately supporting greater affordability, resiliency, environmental performance, and reliability. These and other AI solutions will address challenges utilities face at every stage of resource management, from inventory optimization and energy management to predictive maintenance and process and equipment reliability. 

Increasing Efficiency with AI-Powered Energy Storage

Current new-energy storage systems typically have a four-hour or less duration, corresponding to peaking capacity and ancillary service needs. In the future, there will be a potential need for longer duration storage as systems will be required to absorb longer periods of renewable overgeneration and support resilience during severe weather, among other factors. 

To address these requirements, the Department of Energy is researching how AI and machine learning can accelerate the availability of long-duration energy storage. The proposed initiative would use the technologies to model the performance of various long-duration storage systems, using industry data and digital twins of the storage systems. From there, the DOE would predict how the different technologies may either lose performance or hold up physically over time.

To read more about this research program and how it could benefit the utility sector, head over to Utility Dive.