Fault Diagnosis of Photovoltaic Array Based on XGBoost Method

Abstract

At present, photovoltaic faults often occur in photovoltaic power plants. In particular, short circuit, open circuit, fragmentation and other faults are particularly common and serious. The fault diagnosis of photovoltaic array has become the focus of the current research on photovoltaic. However, the uncertainty of the parameters of solar panels and the complexity of the physical model make it difficult to identify faults by traditional mathematical model. Therefore, an artificial intelligence diagnosis method called XGBoost based on feature parameters (irradiance, temperature, current, power) to realize fault detection and diagnosis is proposed in this paper. This method adopts a novel sparsity-aware algorithm for sparse data and weighted quantile sketch for approximate tree learning which not only guarantees accuracy, but also saves computing resources and improves computational speed. Compared with the existing technology, the experiment shows that this method can better realize the fault diagnosis of photovoltaic array and be better than other machine learning techniques.