Syed Muhammad Hur Rizvi – Doctoral Final Exam

About the event

Student: Syed Muhammad Hur Rizvi

Degree: Electrical & Computer Engineering Ph.D.

Advisor:  Dr. Anurag Srivastava

Dissertation:  Data-driven Algorithms for Load Parameter Estimation and Voltage Stability Assessment

Abstract:

Representing true behavior of the load is critical in power grid analysis tools for efficient operation and planning. The existing load parameter estimation approaches lacks minutes-by-minutes load behavior tracking in real time, inaccurate based on the changing operating scenarios and often not validated. This work develops several algorithms for accurate steady-state load parameter estimation using synchrophasor data and analyzing the impact on the voltage stability. First, the data-driven least square error based robust load parameter estimation algorithms are developed. Secondly, a framework is developed to use, machine-learning tools for steady-state load parameter estimation. Also, a composite load parameter estimation algorithm is developed to consider electronic load behavior for aggregated load parameter estimation. Estimated load models have been utilized in a developed hybrid voltage stability assessment tool that appropriately considers generators’ reactive power reserve status while computing the voltage stability index. Next, a continuation power flow tool is developed to consider the impact of the distribution system on transmission system voltage stability. Such kind of a tool is needed to accommodate the active behavior of the distribution system for voltage stability assessment. Lastly, a short-term voltage stability assessment framework is developed using a 1D-convolutional neural network for early voltage collapse detection and severity quantification of short-term voltage stability events. All the developed algorithms for load modeling and voltage stability have been validated using multiple test systems to demonstrate the performance superiority compared to the existing state-of-the-art.