摘要
针对目前短期空调负荷预测方法预测精度低、稳定性差等问题,提出一种基于微生物遗传算法(Microbial genetic algorithm,MGA)和野狗优化算法(Dingo optimization algorithm,DOA)优化极限学习机(Extreme learning machine,ELM)的空调负荷预测模型。首先利用DOA优化ELM的输入权值和隐层阈值,建立DOA-ELM预测模型,利用MGA改进DOA-ELM模型的预测稳定性和预测精度,建立(Microbial genetic algorithm Dingo optimization algorithm-Extreme learning machine)MDOA-ELM预测模型。为降低预测模型的维度,通过灰色关联分析(GRA)筛选影响空调负荷的输入输出因素。为验证算法有效性,以某工厂中央空调系统为例进行实例分析。实验结果表明,所建负荷预测模型相较于对比模型预测精度高,稳定性好,因此可更好地满足工程实际需求。
A novel air conditioning load prediction model based on Microbial genetic algorithm(MGA)and Dingo optimization algorithm(DOA)optimized Extreme learning machine(ELM)is proposed in this paper to address the issues of low prediction accuracy and poor stability in short-term air conditioning load prediction methods.A DOA-ELM prediction model is established by using DOA to optimize the input weights and hidden layer thresholds of ELM.An MDOA-ELM prediction model is established by using MGA to improve the prediction stability and accuracy of the DOA-ELM model.To reduce the dimensionality of the prediction model,Grey relational analysis(GRA)is used to screen the input and output factors that affect air conditioning load.An air conditioning load prediction example on the central air conditioning system of a factory is provided to verify the effectiveness of the proposed algorithm.Comparing with the reported model,the experimental results show that the established load prediction model has higher prediction accuracy and better stability,and therefore is able to better meet the actual needs of the project.
作者
代广超
吴维敏
Dai Guangchao;Wu Weimin(Polytechnic Institute of Zhejiang University,Hangzhou,310015;School of Control Science and Engineering,Zhejiang University,Hangzhou,310027)
出处
《制冷与空调（四川）》
2024年第3期320-329,共10页
Refrigeration and Air Conditioning
关键词
负荷预测
微生物遗传算法
野狗优化算法
极限学习机
灰色关联分析
Load Prediction
Microbial Genetic Algorithm
Dingo Optimization Algorithm
Extreme Learning Machine
Grey Relational Analysis
