Dissertation Title: Large-Scale Reserve Constrained Alternating Current Optimal Power Flow Methods
When: 12/11/2024 1:00 PM
Where: Simrall 228
Candidate: Yehong Peng
Degree: Doctor of Philosophy in Electrical and Computer Engineering
Committee Members: Dr. Yong Fu, Dr. Seungdeog Choi, Dr. Masoud Karimi, Dr. Xin Fang
Abstract:
The Alternating Current Optimal Power Flow (ACOPF) problem seeks the optimal operational planning for power systems with the lowest cost within the operational and physical constraints, ensuring higher efficiency and reliability for the power system. Meanwhile, the power reserve scheduling, which ensures an adequate backup capacity to compensate for demand and supply fluctuations, contingencies, and various unforeseen disruptions in the power system, plays a crucial role in power system reliability. As the evolution of power systems towards a more stochastic and dynamic feature, the co-optimization of energy and reserve has emerged as a significant trend among independent system operators (ISOs) and regional transmission organizations (RTOs), as it yields substantial improvement in benefits over independent optimizations. Furthermore, expanding ACOPF to a multi-period format facilitates the incorporation of temporal dependencies and operational dynamics, enabling a more comprehensive modeling of realistic power system operations, which enhances the strategic management of the power system. However, the large-scale nature of the power system, the inherent nonlinearity and nonconvexity of the ACOPF problem, the increased model complexity due to the additional reserve scheduling, and the expanded model scale with linked decision variables across multiple time periods, collectively contribute to its substantial computational complexity, making it a challenging optimization problem to solve.
In order to tackle this challenge and achieve a fast, optimal, and reliable solution to power system operational planning, the advanced solution methods are developed in the proposed research, which involve two representative endeavors, including the development of mathematical optimization algorithms with enhanced computational efficiency, and the investigation of the decomposition techniques along with the implementation on the advanced computational resources. Firstly, the single-period reserve constrained AC Optimal Power Flow (RCOPF), which co-optimizes energy and reserve through detailed modeling of various reserve types with considering AC network constraint, is addressed with an accelerated primal-dual interior point optimization algorithm. Secondly, to address the multi-period RCOPF (M-RCOPF), a parallel optimization approach is developed, which reformulates the intertemporal coupling constraints with introduced auxiliary variables to enable decomposability, and fully decomposes the originally temporal coupled problem into two major solution modules. Each module consists of multiple independent smaller subproblems that are solved simultaneously in a fully parallel manner on the advanced High-Performance Computing (HPC) platform. As the results, with the proposed methods, the large-scale reserve-constrained ACOPF problems can be solved with computational efficiency, which ensures the efficient and reliable operation planning for the power system.
