Walter Scott, Jr. College of Engineering

Graduate Exam Abstract

Hongfei Sun
Ph.D. Final
Nov 02, 2021, 10:00 am - 12:00 pm
Power System Data Classification and Prediction by Functional Analysis
Abstract: The last couple of decades have witnessed the development of our electric power grid. The growing population size and increasing consumerism have increased the load demand and brought more pressure on the grid. Meanwhile, new elements are being introduced to the power grid such as various forms of renewable energy resources, electric vehicles, and so on, which need to be monitored constantly and managed properly. In addition, the allocation of the various resources in the power systems is now conducted in a much more dynamic manner than ever. All these new dimensions have driven the development of the traditional grid into the smart grid and call for new methodologies in system design, operation, and control. This dissertation focuses on the modeling of power systems with data-driven approaches, with applications in power system cyber-attack detection and recovery, and large-scale long-term load characterization. Firstly, the modeling of the spatial-temporal relationship among the quantities across the entire power systems is provided with applications to cyber-attack detection and data recovery. Then, the non-conforming load classification approaches based on Functional Principle Component Analysis (FPCA) will be introduced. This work is the first effort towards such loads due to the recently growing penetration of Distributed Energy Resources (DER) users. Lastly, the regional high-resolution medium-term load forecasting approach will be introduced. In order to satisfy the new purpose of load forecasting, serving for real-time applications, our approach can provide higher resolution than existing long-term load forecasting and longer leading time than the existing short-term load forecasting time-series load curve. Based on the presented case studies and simulation results, the corresponding suggestions to the present industrial power system were provided.
Adviser: Dr. Liuqing Yang
Co-Adviser: Dr. Jie Luo
Non-ECE Member: Dr. Haonan Wang
Member 3: Dr. Dongliang Duan
Addional Members: Dr. Hongming Zhang
H. Sun, D. Duan, H. Wang, H. Zhang, J. Luo, and L. Yang, “Cyber-Attack Detection based on Nonlinear Spatial-Temporal Model,” IEEE Transactions on Smart Grid (In submission, Oct. 2021)

H. Sun, D. Duan, H. Zhang, J. Luo, H. Wang, and L. Yang, “Classification for Non-conforming Loads Based on Functional Principle Component Analysis (FPCA),” IEEE Transactions on Smart Grid (In submission, Oct. 2021)

H. Sun, D. Duan, H. Zhang, J. Luo, H. Wang, and L. Yang, “Regional Medium-Term Hourly Electricity Demand Forecasting Based on LSTM,” IET Generation, Transmission & Distribution (In submission, Oct. 2021)

C. Qin, H. Sun, X. Liu, J. Chen, "Adaptive learning solution of the nonzero-sum differential game with unknown dynamics using adaptive dynamic programming." 2016 Chinese Control and Decision Conference (CCDC). IEEE, 2016.
Program of Study:
ECE 614
ECE 652
ECD 666
ECE 520
ECE 555
ECE 509
CS 545
MATH 561