On-line Diagnosis Technology of Distributed Photovoltaic Abnormal Generation Based on Output Prediction

The abnormal situation of photovoltaic power generation, such as the decline of photovoltaic power generation efficiency, panel damage and so on, leads to the decline of photovoltaic panel power generation, which affects the photovoltaic power generation revenue. It is urgent to identify abnormal in time and restore output. The characteristics of generation power curve in abnormal period of time are excavated, and the abnormal diagnosis technology based on output prediction is put forward. Firstly, the data of 96-point generation power curve in the history of power station are collected, and an accurate 96-point output prediction model is constructed. Then, based on 96-point output prediction curve, the actual output curve is compared on-line in real time, which is in line with the actual output curve. When abnormal characteristics occur, the abnormal scenes of power generation efficiency, such as deciduous cover and shadowing, can be identified intelligently by judging abnormalities and alarming. In a 120 kW power plant, simulation of snow cover anomaly scenario, anomaly online diagnosis technology based on output prediction is used to quickly identify the anomaly scenario of power generation efficiency, which shows the feasibility of anomaly diagnosis method.