Method for predicting wind power based on mechanism and data hybrid drive neural network

The invention relates to the technical field of industrial process soft measurement modeling, and particularly provides a method for predicting wind power based on a mechanism and data hybrid drive neural network. The method comprises the following steps: constructing a mechanism model of a wind power curve; an activation function Leaky ReLU is added to an output layer of the GRU model, output is constrained, and an improved GRU model is obtained; combining the mechanism model with the improved GRU model to construct a hybrid drive neural network model; and training the hybrid drive neural network model through the smoothed data to obtain a predicted wind power value. The method overcomes the defect of constructing a soft measurement algorithm based on a single data drive model or a mechanism model; the problems that abnormal values exist in measured data and estimated values which do not conform to physical rules exist in estimation results are solved, accurate and robust estimation is carried out on the win