Fault Diagnosis and State Evaluation of Distributed Photovoltaic Systems in Microgrids

Aiming at the crack, short-circuit, and open-circuit faults of distributed photovoltaic (PV) systems in microgrids, a fault diagnosis method of the distributed PV system based on CatBoost is proposed. Based on the Gradient Boosting Decision Tree framework, the CatBoost can efficiently and accurately determine the characteristics of PV fault types. Based on the effect of this method, in order to further evaluate the operating state of the distributed PV system. This paper proposes a state evaluation method based on the technique for order of preference by similarity to ideal solution (TOPSIS). This method sets the feature weight factor by the feature importance of CatBoost to quantitatively characterize the state of the distributed PV system under different conditions. Experiments show the effectiveness and reliability of the proposed methods, and the methods improve the intelligent management level of distributed PV systems in microgrids.