软件导刊2025，Vol.24Issue(7)：46-53,8.DOI:10.11907/rjdk.241457
基于IWOA-BiLSTM-MHSA神经网络的超短期风电功率预测
Ultra-Short-Term Wind Power Prediction Based on IWOA-BiLSTM-MHSA Neural Network
摘要
Abstract
To improve the accuracy of ultra short term wind power prediction,an improved whale optimization algorithm combined with multi head attention mechanism is proposed from the perspectives of hyperparameter optimization and model optimization,and a bidirectional long short-term memory neural network model is proposed.Firstly,based on the bidirectional long short-term memory neural network,a temporal multi head self attention mechanism is introduced to capture longer distance dependencies in the temporal sequence,increasing the model's representation and generalization ability;Secondly,an improved whale optimization algorithm is adopted to optimize the hyperparameters of the model,proposing adaptive parameter and threshold strategies to enhance the algorithm's global search capability and local search rate;Fi-nally,based on the measured data of a wind farm in a certain region of Xinjiang,a case study analysis was conducted.It was found that com-pared with various traditional machine learning and deep learning prediction models,the proposed model has higher prediction accuracy,es-pecially in dataset 1,where MAPE was reduced by 4.892%,7.722%,6.196%,3.864%,and 2.159%compared to LSTM,BP,RNN,GRU,and BiLSTM models,respectively.关键词
双向长短时期记忆网络/多头自注意力机制/风电功率预测/改进鲸鱼优化算法Key words
BiLSTM/multi-head self-attention mechanism/wind power prediction/improved whale optimization algorithm分类
信息技术与安全科学引用本文复制引用
张州潼,郭欢..基于IWOA-BiLSTM-MHSA神经网络的超短期风电功率预测[J].软件导刊,2025,24(7):46-53,8.基金项目
教育部人文社科基金项目(23YJCZH062) （23YJCZH062）
