Join the Asset Health Analytics Community on Wednesday, September 27, 2023 at 2:00 PM CT for a presentation with Q&A and open discussion on "SDG&E’s AI-Powered Asset Management & Inspections Reducing Wildfire Risk" led by Matan Gandelman, Product Owner, Image Analytics Capability at SDG&E.
Session Description:
The Digital Innovation team scaled the Intelligent Image Processing (IIP) initiative, designed to help address one of the region’s greatest challenges: the evolving and expanding threat of devastating wildfires.
Intelligent Image Processing is an innovative human + machine partnership and IT solution that leverages Artificial Intelligence (AI) and Machine Learning (ML) to rapidly detect assets and damages in millions of images captured via drones and mobile devices in the field. Over 3.4+ million images of equipment in high-risk fire areas have been ingested and analyzed by in-house-built ML models. Once images are uploaded to a central cloud repository, the visualization platform supports the review process for a qualified electrical worker. They validate the algorithm by annotating the images detecting SDG&E assets and visible damages, which further trains the model. Previously, the start-to-end process was manual and time-consuming; on average taking a team of 10 a day to get through 2,000 images. Today, 27,000 images are examined in the same amount of time. This capability shows the true value of a human + machine relationship and how the future of work can be more efficient and effective in the long term, moving with speed and agility.
This program is a "first-of-its-kind" cloud-based solution for the San Diego-based utility and demonstrates how the cloud can fuel innovation. SDG&E received the 2021 CIO 100 award for its use of innovative technology to deliver greater business value and enhanced customer experience. Since our initial proof of concept, we have deployed 94+ asset, damage, and third-party attachment detection capabilities that can identify fire risks in images in CPUC-defined high-fire threat areas. As new imagery is captured across the service territory, IIP will continuously flag damaged assets for human validation. The program has expanded to leverage LiDAR for vegetation encroachment identification, passive and active mobile phone image capture for asset and damage detection, and future-state integration of satellite imagery to provide context for the entire service territory.
Key Takeaways:
• Learn how SDG&E’s innovation framework enables the organization to rapidly test and scale programs that drive informed technology investments and add value to the business
• Gain an understanding of how SDG&E is using cloud technology and AI to predict asset and damage predictions in areas across their High Fire Threat Districts
• Experience firsthand the capabilities of SDG&E’s Intelligent Image Processing through a live demonstration
• Recognize the capacity to scale Intelligent Image Processing and learn about SDG&E’s plans for future asset management and inspection initiatives
Members of the Asset Health Analytics Community will receive a Microsoft Teams link prior to this meeting. Not a member of this community? UAI Utility Members can register for any of our communities by using our Request To Join link.