Machine Learning-Based Method for Power System Operation Rules Generation

Abstract

With continuous expansion of the scale of power systems and the ever-increasing penetration of intermittent renewable energy generation, the uncertainty, complexity, and diversity of power system operation have increased significantly. Hence, traditional power system operation rules which are based on the experience of operators are difficult to be applied to large-scale modern power systems. The development of machine learning provides a new perspective for online formulation of operation rules. Transmission section determination is the basis in formulating operation rules, and in this paper, the Clustering by Fast Search and Find of Density Peak (CFSFDP) algorithm and a cut-vertex search method are adopted to identify key transmission sections of the power system concerned. Then, the Monte Carlo simulation method is used to generate massive training samples, and the Total Transfer Capability (TTC) is calculated for each key transmission section. After feature selection, the Radial Basis Function (RBF) neural network is used to train the model to attain the operation rules. Finally, the IEEE 39-bus power system is employed to demonstrate the proposed method.