PG&E Investing in Centralized Data Analytics to Drive Data-Informed Decisions and Reduce Risk

By Paul Doherty

As a part of safely operating the gas and electric infrastructure that delivers clean energy to more than 16 million people across Central and Northern California, PG&E collects between 8 billion and 10 billion data points every day.

This data is coming from myriad sources including smart meters, equipment, sensors, environmental and meteorological systems, and work management and operations activities. It also includes a growing amount of data requiring additional classification or management for processing—known as unstructured or semi-structured data—such as high-resolution photos, video, LiDAR, and sensor logs.

Historically, PG&E’s data source systems and warehouses were deployed to support specific capabilities and business functions and were only integrated with other systems as needed.

However, having no centrally integrated and managed environment for understanding, correlating, and analyzing disparate data sets, it has been challenging to develop enhanced risk analytics to inform PG&E’s wildfire mitigation programs and initiatives, with some efforts requiring months of data gathering and cleansing before new models could be developed and tested.

Until now.

Determined to minimize the impacts of Public Safety Power Shutoff events on its customers, PG&E in late 2019 engaged Palantir, a software company specializing in big data analytics, to improve the end-to-end management of the power shutoff process.

Using Palantir’s Foundry technology, last year PG&E successfully completed multiple initiatives aimed at improving PSPS-related decision making, reporting and communications.

Based on the success of these initiatives, PG&E has begun consolidating key data for assets, customers, and operations into this new centralized data analytics platform, allowing decision-makers and team members to more efficiently combine data and perform analysis to improve data-driven decision making.

“Timely access to high-quality, trusted, data is foundational to improving PG&E’s ability to develop and operationalize new and deeper insights,” said Mike Glass, director, IT Data and Analytics, PG&E. “This comprehensive platform significantly advances PG&E’s ability to take in and organize data from disparate sources, build analytic models, and visualize the results, leading to better decision making and ultimately, enhanced safety and reliability for our customers.”

PG&E is now expanding its deployment of Foundry to serve as the core enabling technology for building out the centralized data system, which will eventually bring together physical, operational, lifecycle and environmental data elements.

This includes leveraging integrated data through this new centralized data system to investigate equipment failure incidents. This enables PG&E to further identify trends in equipment condition, create alerts for asset managers, build predictive risk-scoring metrics, and equip grid operators with a toolset and framework for targeting failing equipment with recognizable electrical patterns. 

Part of a growing trend 

Other innovation-related risk reduction efforts underway at PG&E include testing and deploying nascent sensor technologies to monitor irregular activities and detect emerging conditions on the electric grid that may require action, and using artificial intelligence-enabled software for better identifying risks on the electric transmission system.

Ongoing safety work to enhance grid resilience and address the growing threat of extreme weather and wildfires across PG&E’s service area continues on a risk-based and data-driven basis. This includes expanding the network of enhanced weather technology, installing stronger poles and power lines, doing targeted undergrounding, conducting enhanced vegetation management, and developing microgrids.

Learn more about PG&E’s wildfire safety efforts by visiting pge.com/wildfiresafety.

Email Currents at Currents@pge.com.