Utility Lowers Maintenance Costs and Improves Service with AI Analytics

Utility Lowers Maintenance Costs and Improves Service with AI Analytics

Torgeir Brovold
Torgeir Brovold,

Monitoring and maintenance of components in spread-out electric distribution networks, is a cumbersome and expensive task for utilities. A self-service AI driven applications helps this company analyze drone-captured images, a solution that saves both time and money.

Fixing damaged or failing components, such as the insulators that connect transmission wires to poles, is essential to maintaining service levels and preventing outages for any utility company. It is also a very time-consuming and complex process.

Download Report: Enhancing the Utility Customer Experience

That’s why this U.S. based electric utility company had invested heavily in drones to captures images of its power network. This helped them identify equipment in need of repair. However, manually scanning the images and then opening a repair ticket was an inefficient process. To automate the image scanning, the utility turned to Cognizant for an artificial intelligence solution. 

Lack of Training Data
The Cognizant team was initially faced with several challenges: the AI solution should ignore objects such as trees, poles and wires, provide real-time notifications, and automatically generate work tickets for maintenance staff. Unfortunately, there was a limited number of images showing the correct classification of equipment problems (AI training data).

To supply the required amount of training data, the Cognizant team developed a deep neural network cognitive model using image augmentation. The utility company now has a fully-managed data and analytics platform, built on the Cognizant AI Data Modernization Platform, that enables data scientists to “visually” build, train and deploy AI models at any scale either on-site or in the cloud. Drone-captured images of distribution equipment can now be processed in real time.

Better Service, Lower Cost
All in all, an expensive, time-consuming manual process of identifying equipment failures and ordering repairs, has been replaced with a self-service AI driven application that analyzes drone-captured images using dynamic visualizations to enhance the user experience. Thanks to the solution, the company has experienced a 60% reduction in image analysis efforts, something that positively affects both labor and maintenance costs as well as increases service levels. 

Related Publication

White Paper: Blockchain For Power Utilities

Can blockchain be an option for power utilities that want to streamline operations, decrease costs and deliver on the promise of the distributed smart grid? This white paper helps you evaluate blockchain-based technologies and provides guidance on a framework approach to move forward.