Surrogate Modeling-Based Multi-Objective Dynamic VAR Planning Considering Short-Term Voltage Stability and Transient Stability

Tong Han,Yanbo Chen,Jin Ma,Yi Zhao,Yuan-ying Chi

DOI: https://doi.org/10.1109/tpwrs.2017.2696021

IF: 7.326

2018-01-01

IEEE Transactions on Power Systems

Abstract:Transient stability and short-term voltage stability have successively attracted the attention of electric power industry. This paper proposes a novel systematic approach for dynamic VAR planning to improve short-term voltage stability level and transient stability level. The dynamic VAR planning problem is formulated as a multi-objective optimization (MOO) model with objectives including investment cost, short-term voltage stability level, and transient stability level. To reduce the complexity of the proposed MOO model, K-means clustering-based severe contingencies selection and global sensitivity analysis-based potential buses selection are employed, leading to a simplified MOO model. The combination of a surrogate modeling technique called support vector regression and the multi-objective evolutionary algorithm (MOEA) are then used to solve the simplified MOO model, considering both the accuracy of models and the optimization computation cost. This combination makes it feasible to perform multiple runs of MOEAs for weakening the effect of the MOEA's randomness to optimal results and offering more diverse Pareto-optimal solutions for decision makers. Simulations are carried on the IEEE 39-bus system and a real power grid of China, illustrating that our methodology is reliable with high efficiency.