As new energy sources such as wind power and photovoltaic power gradually become the dominant power sources

the challenge of “guaranteeing power supply” of the power system is becoming more and more prominent. In view of the problem that there is still a small probability of extreme positive deviation in the new energy deterministic prediction

which further increases the difficulty of power preservation

a prediction method for wind power generation’s intraday security capacity is proposed

which penalizes the positive deviation and maximizes the proportion of power preservation

and the prediction time scale is 24 hours

with rolling updates every hour. Firstly

a model objective optimization loss calculation function that takes into account the minimization of the positive deviation and the maximization of the predicted output level is created. Then

based on the convolutional neural network-long short-term memory network (CNN-LSTM)

an intraday guaranteed capacity prediction model for wind power generation is constructed. Finally

based on the parameter co-optimization of the guaranteed capacity prediction model of the multi-strategy fusion gray wolf algorithm

a cooperative optimization method of the parameters of the guaranteed capacity prediction model is developed to optimize the super-parameters of the asymmetry penalty loss function. Moreover

a cooperative optimization method based on multi-strategy fusion gray wolf algorithm is used to optimize the hyperparameters of the asymmetric penalty loss function

which forms a complete and implementable technical program. In order to compare the effectiveness of the proposed method

a multi-dimensional evaluation index system is constructed

and the proposed method is validated by using the actual operation data of several wind farms in East China. The results show that the proposed method can be adopted to improve the guaranteed power share index by 15%~25% and reduces the error index by more than 20% compared with the comparison model

which verifies the effectiveness and advancement of the proposed method.