Fault diagnosis method for oil-immersed transformer

The invention relates to an oil-immersed transformer fault diagnosis method, and belongs to the technical field of power systems. According to the method, firstly, collected fault data of the oil-immersed transformer are processed and divided, then transformer fault state codes are formulated, a classification regression tree weak classifier model is constructed, and the weak classifier model is trained in an addition learning mode; and optimizing classifier parameters by using a sparrow search algorithm to obtain an XGBoost fault diagnosis model of the oil-immersed transformer, and predicting to-be-tested data to obtain a fault diagnosis result. The method uses the sparrow search algorithm to optimize the XGBoost model, is simple in structure, is easy to realize, has good global exploration and local development capabilities, accelerates XGBoost model training, improves the accuracy of transformer fault diagnosis, enables a maintenance department to timely and accurately maintain the transformer, remarkably