The client is a global energy and sustainability advisory firm that helps large corporate, industrial, and institutional clients better navigate the choices, opportunities, and risks emerging from the transition to a net-zero future.
The Challenge
The client is one of the US’ largest electric utilities and a long-time leader in renewable energy and energy efficiency. The client launched a grid resilience project that monitored and predicted energy demand from customer utility assets, such as electric cars, home utility meters, etc., to anticipate the surge in demand and to prioritize supply to areas with a surge. This was essential to remediate asset failures and pre-empt hazards related to the electric system in high-fire-risk areas.
The client collected swathes of unstructured and structured data as images which had to be analyzed in real-time to glean actionable insights.
The client turned to Persistent to create an image analytics framework that could ingest a large volume of structured and unstructured data in real time to intervene timely around electricity supply.
The Solution
The Google Cloud experts at Persistent developed and deployed an image processing workflow using Cloud Composer, Kubernetes Engine, Cloud Storage, and Dataflow.
With the powerful capabilities of VertexAI, we built an analytics engine for near-real-time insights.
We switched to an event-based approach to improve the processing model and utilized BigQuery and ElasticSearch as dashboards for a comprehensive view of electricity consumption.
The Outcome
The client witnessed:
- 10X improvement in data processing capabilities, enabling more vendors to upload images.
- Real-time image processing for object detection and immediate feedback to image vendors.
