An Intelligent Stability Prediction Method of Grid-Connected Inverter Considering Time-Varying Parameters

This paper presents an intelligent stability prediction method for high-frequency oscillation of grid-connected inverter considering time-varying parameters of power grid and inverter. A data-based analysis method based on radial basis function neural network (RBFNN) is first developed to identify and predict time-varying parameters of grid and inverter. Then, the oscillation characteristic represented by physical model is combined to predict real-time stability of grid-connected inverter. Furthermore, the stability prediction criterion is developed according to real-time parameter identification and physical model. Simulation and experimental results are given to validate the proposed intelligent stability prediction method. The proposed method is able to predict time-varying stability region and stability margin of grid-connected inverter considering parameters variation, which thus improves the self-learning capability and adaptivity of grid-connected inverter system.