节能2025，Vol.44Issue(4)：61-65,5.DOI:10.3969/j.issn.1004-7948.2025.04.015
基于人工神经网络的板式换热器热阻和能效预测模型研究
Research on thermal resistance and energy efficiency prediction model of plate heat exchangers based on artificial neural network
宋坤卿 1陈镜如 2孙燕华3
作者信息
- 1. 济南市计量检定测试院,山东 济南 250101
- 2. 山东电力工程咨询院有限公司,山东 济南 250013
- 3. 山东中和热工科技有限公司,山东 济南 250401
- 折叠
摘要
Abstract
In response to the prediction requirements of energy efficiency and thermal resistance of plate heat exchangers,taking the flow rates of cold and hot fluids as input parameters and thermal resistance and energy efficiency as output parameters,an artificial neural network model of plate heat exchangers is constructed.The optimal model parameters are determined through hyperparameter analysis.Considering the inlet and outlet temperatures of the fluids on the cold and hot sides,two sets of control models are established respectively.Combined with SHAP analysis,the marginal contributions of each thermodynamic parameter to the prediction of thermal resistance and energy efficiency are given.The results show that the artificial neural network model with the flow rates of cold and hot fluids as input has better predictive performance,and the maximum relative errors for predicting thermal resistance and energy efficiency are 3.08%and 5.97%respectively.The flow rates of cold and hot fluids are important influencing factors for the output of artificial neural network models and should be given priority when constructing the model and selecting the training set.关键词
换热器/性能预测/神经网络/SHAP分析Key words
heat exchanger/performance prediction/neural network/SHAP analysis分类
计算机与自动化引用本文复制引用
宋坤卿,陈镜如,孙燕华..基于人工神经网络的板式换热器热阻和能效预测模型研究[J].节能,2025,44(4):61-65,5.
