Data-Driven Situational Awareness for Distribution System Resiliency by Chuan Qin – Dissertation Defense
About the event
Student: Chuan Qin
Advisor: Dr. Anurag Srivastava
Degree: Electrical Engineering Ph.D.
Dissertation Title: Data-Driven Situational Awareness for Distribution System Resiliency
Abstract: With the continuous deployment of distributed energy resources (DER), behind-the-meter resources, advanced sensors, and grid-edge intelligence in recent years, the distribution power grid monitoring and operation is increasingly becoming complex. Effectively leveraging the massive amount of data available with digital automation to achieve situational awareness (SA) for the right operational decisions is critical to enhancing system resiliency, stability, and sustainability.
Improving SA can be achieved through sensor data acquisition, real-time data analytics, physics-aware machine learning for the estimation of unmeasured data, efficient data management, and information extraction. Existing state-of-the-art grid data sources and monitoring approaches are typically stovepiped into repositories of operational data, planning data, and third-party data. Data from distribution phasor measurement units (D-PMUs), supervisory control, and data acquisition (SCADA) are examples of operational data. Besides, data from advanced metering infrastructure (AMI) meters is an example of enterprise data. In addition, historical data and asset data are examples of planning data. Furthermore, local news and weather forecasting are examples of third-party data. This segmentation limits realizing the unabridged value of available data. In addition, this set of data varies in terms of volume, velocity, variety, veracity, and value. The objective of this work is to coordinate, estimate and process the segmented data, and enhance the situational awareness of the distribution power grid with high penetration of DER. Specific goals of his work include a) PV forecasting, nowcasting, estimation, and aggregation, and b) utilizing estimated data and aggregation for efficient system monitoring and resilient operation. Results demonstrate the superiority of the developed approach for PV estimation and behavior of the reduced aggregated system being close to a fully modeled distribution network while working with industry collaborators.