Predicting disrupting events in power distribution systems for power grid management

This technology is a method of predicting disruptive events in power grid systems to communicate contingency and real-time management decisions to prevent electricity outages, referred to as a Distribution Contingency Management System (DCMS).

Unmet Need: Preventing electricity outages before they happen based on real-time data measurements

Energy distribution relies on a complex power grid that can be prone to errors. Real-time measurement of all recordings in the grid allows utilities to understand which components of the system are defective and need to be repaired in the event of an electricity outage. Current distribution management systems record these disruptive events at their occurrence, but do not provide information in order to prevent outages from happening. Although some researchers are attempting to develop a system to prevent outages from occurring, there is no methodical system that has been implemented to predict future disrupting events and inform utilities to perform future improvements on their power distribution systems.

The Technology: Computational management system to predict disruptive events and implement future improvements in power distribution

This technology describes the DCMS, a computational method that uses real-time measurements from power distribution systems to predict disruptive events in order to prevent electricity outages. Using a reinforcement learning algorithm, data can be processed to predict points of disruption, allowing rapid management decisions to be made to remediate the problemed area. Additionally, DCMS can estimate future needs to provide accurate, controller-based pricing, curtailment, and load pocket management mechanisms.

Applications:

• Predict power outages and prevent them from happening
• Optimized power distribution
• Determine future improvements to the distribution system
• Can be implemented in alternative areas like oil, telephone networks, transit, data communication, and any system where resources rely on a distributed source

Advantages:

• Uses machine learning algorithms to make predictions with high accuracy
• Improves its prediction capabilities as more data is input
• Flexible application to prevent and also predict future improvements
• Saves time and money to utilities by preventing outages
• Enhanced customer experience

Lead Inventor:

Roger N. Anderson

Patent Information:

Patent Status

Related Publications:

• Rudin C, Waltz D, Anderson RN, Boulanger A, Salleb-Aouissi A, Chow M, Dutta H, Gross PN, Huang B, Ierome S, Isaac DF. “Machine learning for the New York City power grid” IEEE transactions on pattern analysis and machine intelligence. 2012 Feb 1;34(2):328.

• Anderson R, Boulanger A, Powell W, Scott W. “Adaptive stochastic control for the smart grid” Proceedings of the IEEE. 2011 Jun; 99(6): 1098-1115

Tech Ventures Reference:

• IR M09-054

• Licensing Contact: Richard Nguyen