基于机器学习的双循环除湿机的除湿性能预测研究  

Research on dehumidification performance prediction of dual cycle dehumidifiers based on machine learning

作　　者：陈禹宋 蔡慧 蔡文剑 郭倩 CHEN Yusong;CAI Hui;CAI Wenjian;GUO Qian(College of Mechanical and Electrical Engineering,China Jiliang University,Hangzhou 310018,China;Ningbotech University,Ningbo 315100,China)

机构地区：[1]中国计量大学机电工程学院,浙江杭州310018 [2]浙大宁波理工学院,浙江宁波315100

出　　处：《中国计量大学学报》2023年第1期20-25,共6页Journal of China University of Metrology

基　　金：浙江省属高校基本科研业务费专项资金项目(No.2021YW42)。

摘　　要：目的:提出一种基于机器学习的双循环空气除湿机除湿性能预测模型,用于判别系统运行工况。方法:采集系统在相对稳定状态中的运行数据,进行BP神经网络和遗传BP神经网络的模型训练与验证,并采用相对误差、相对误差均方根(RMSRE)和均方根误差(RMSE)作为评判指标验证模型的可靠性,对比得出预测精度较高的神经网络模型。结果:遗传BP神经网络模型对系统出口空气含湿量的预测和实测值吻合良好,相对误差处于4%范围内,RMSE、RMSRE分别为0.0891、1.3542%,均优于BP神经网络模型。结论:本文中模型满足预测精度要求,能够作为系统运行状态判别的依据。Aims:A machine learning-based model for dual-cycle air dehumidification system performance prediction was proposed to identify the operating conditions of the system.Methods:The operation data of the system during the relatively stable state were collected;and the BP neural network model and the genetic BP neural network model were trained and verified.The relative error,the root mean square of the relative error(RMSRE)and the root mean square error(RMSE)were used as the evaluation indexes to verify the reliability of the model.Results:The prediction of the air moisture content at the outlet of the system by the genetic BP neural network model was in good agreement with the measured value;and the relative errors were within the range of 4%.The RMSE and RMSRE were 0.0891 and 1.3542%,respectively,which were better than the BP neural network model.Conclusions:The model meets the requirement of prediction accuracy and can be used as the basis for judging the system running state.

关 键 词：双循环空气除湿系统 预测模型 除湿机稳态 遗传BP神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]