Photovoltaic power station inverter fault diagnosis method based on LightGBM model

The invention discloses a photovoltaic power station inverter fault diagnosis method based on a LightGBM model, and relates to the field of photovoltaic power generation detection. The photovoltaic power station inverter fault diagnosis method based on the LightGBM model comprises the following steps: S1, data acquisition; s2, feature extraction; s3, model selection; and S4, realizing the model. According to the photovoltaic power station inverter fault diagnosis method based on the LightGBM model, the health state of the photovoltaic power station inverter is evaluated by adopting a big data classification method, so that fault diagnosis is carried out, the photovoltaic power station inverter has multiple data acquisition points, the acquired inverter operation data is large in quantity, multiple in variety and long in time span, the data acquisition is simple and convenient, and the fault diagnosis efficiency is improved. Data cleaning and feature extraction are carried out on the collected original data by