Autonomous Grid using Machine Learning
Submitted by Charan Puvvala (@charanpuvvala) on Tuesday, 25 April 2017
Section: Full talk for data engineering track Technical level: Intermediate
In this talk we deep dive into how we are assisting Energy Utilities using IOT and Machine Learning to build the next generation of Autonomous grid. The potential impact of applying Machine Learning, IOT, IIOT is estimated at 2-4% of annual revenue, 3-5% of annual accounts receivable, cost improvement of 4-8% per campaign against their consumers. The topics include application of Machine Learning on 10 specific problems in Utilies ranging from Power generation to consumption and post service.
The principles from this talk can be applied to not just Energy but other consumables within utilities like Natural Gas, Water, Sewage
The problems and techniques which will be covered in the talk are:
1. Calculation and verification of demand response.
2. Identification of fraud/theft of power via bottom-up consumption analysis
3. Meter-to-Cash Analysis
4. Gamification of energy usage among users at locality level
5. Customer Usage Pattern Analysis
6. Reports to regulatory bodies on Performance and Service reliability
7. Transformer overload detection and circuit analysis
8. Extension of the life of assets via predictive and pro-active maintenance
9. Load profile characterization and definition for segmented customers – residential and commercial
10. Capacity Offset Forecasting for load reduction programs (like demand response and interruptible load control programs).
Charan is a Chief Data Architect at Proarch. In this role he is responsible for architecting and building systems in domains of Utilities and BFSI. His special interests lie on solving problems using ANNs, Semantic and Linked Data. He also asists the Standard Chartered Bank in building their Data Science Stack, teams.
Charan was previously associated with Amazon & World Bank, respectively. During his stint with World Bank, Charan worked on the prestigious Open Data project as an Integration Specialist in 2011.
He was a Research Scholar at The Harvard University, Center for Italian Renaissance, Florence, Italy on Linked Data & Graph theory.
Charan is also the organizer and a regular Speaker at the Hyderabad Machine Learning Community.