A Rapid Monte Carlo Reliability Evaluation Method for Integrated Energy Systems Based on Transformer

Abstract

In this paper, a Transformer-based fast Monte Carlo reliability evaluation method for integrated energy systems is proposed. First, the faults of the system components are sampled and the corresponding minimum cut load amounts are calculated to obtain the sample data for training the machine learning model. Then, Transformer is used as a machine learning algorithm for mining the nonlinear mapping relationship between system component faults and the minimum cut load, and the estimation model of the minimum cut load under different faults is trained. Finally, the model is combined with Monte Carlo simulation method to randomly sample component states, and for each state, the minimum cut load amount is directly given by the trained estimation model, thus realizing fast evaluation of integrated energy system fast reliability. The proposed method is applied to the integrated energy system test case, which verifies its effectiveness.