现代电子技术2024，Vol.47Issue(14)：20-29,10.DOI:10.16652/j.issn.1004-373x.2024.14.004
基于改进沙猫群算法优化CNN-BiLSTM的热负荷预测
Optimised CNN-BiLSTM for heat load prediction based on improved sand cat swarm algorithm
王耀辉 1薛贵军 2赵广昊1
作者信息
- 1. 华北理工大学 电气工程学院,河北 唐山 063200
- 2. 华北理工大学 电气工程学院,河北 唐山 063200||华北理工大学智能仪器厂,河北 唐山 063000
- 折叠
摘要
Abstract
In allusion to the problems that the traditional heat load prediction accuracy is not high enough to meet the demand of heat exchange station and heat network optimisation and regulation,a VMD-ISCSO-CNN-BiLSTM heat load prediction model is proposed.The variational mode decomposition(VMD)is used to denoise the original heating load data and reduce its instability.The K-means algorithm is used to improve population initialization,the evolutionary mechanism is used to improve optimization ability,the mutation mechanism is used to improve the ability to jump out of local optima,and the improved sand cat swarm algorithm(ISCSO)is used to optimize the hyperparameters of convolutional neural networks and bidirectional long short-term memory neural networks(CNN-BiLSTM),so as to establish the heat load prediction model.The model is analysed by examples.The results show that the model prediction accuracy is higher after data noise reduction,and the R2 is improved by 1.1%.The model optimised by ISCSO is better than the models optimised by other algorithms,with a fit of 99.4%.In comparison with the single model,the combined prediction model of VMD-ISCSO-CNN-BiLSTM has a lower RMSE by 18.5%,a lower MAE by 13.8%,and a higher R2 by 15.8%.It has better goodness of fit and strong generalization,which meets the actual requirements of the project.关键词
热负荷预测/卷积神经网络/双向长短期记忆神经网络/改进沙猫群算法/变分模态分解(VMD)/K-means算法/演变机制/变异机制Key words
heat load prediction/convolutional neural networks/bidirectional long short-term memory neural network/improved sand cat swarm algorithm/variational mode decomposition/K-means algorithm/evolution mechanism/mutation mechanism分类
电子信息工程引用本文复制引用
王耀辉,薛贵军,赵广昊..基于改进沙猫群算法优化CNN-BiLSTM的热负荷预测[J].现代电子技术,2024,47(14):20-29,10.
