电网智能化调度与控制教育部重点实验室(山东大学), 山东省济南市 250061
Key Laboratory of Power System Intelligent Dispatch and Control of Ministry of Education(Shandong University), Jinan 250061, China
With the development of ultra-high-voltage direct current(HVDC)and the changes of load composition and characteristics, the safe and stable operation of power system is seriously threatened by transient voltage problem. Based on convolutional neural network(CNN), a fast assessment method for partitioning transient voltage stability of AC/DC receiving-end power grid is proposed. Considering the influence of fast dynamic response components, the transient voltage sag area matrix is set up based on the time sequence information of transient voltage. This matrix is mapped into two-dimensional plane to divide the receiving-end power grid into several partitions based on the t-stochastic neighbor embedding(t-SNE)algorithm. The steady power flow features of each partition are selected according to relative distance of buses. Fault severity index is constructed and the fault line number is coded based on the order of fault severity index. The coding result and fault line number are taken as fault features. The particle swarm optimization(PSO)algorithm is adopted to determine the optimal size and number of convolution kernel in each partition of CNN to improve the performance of CNN. Simulation results of a real multi-infeed AC/DC power system show the effectiveness of the proposed method. This work is supported by National Key R&D Program of China(No. 2017YFB0902600)and State Grid Corporation of China(No. SGJS0000DKJS1700840).
YANG Weiquan, ZHU Yuanzhen, LIU Yutian. Fast Assessment of Transient Voltage Stability Based on Convolutional Neural Network[J]. Automation of Electric Power Systems, 2019, 43(22):46-51. DOI:10.7500/AEPS20190430037.