文献类型：期刊文献

中文题名：基于支持向量回归的露点间接蒸发冷却模型应用研究

英文题名：Application of dew point indirect evaporative cooling model based on support vector regression

作者：许家翔[1];陈瑞[1];曹军[1]

机构：[1]华东理工大学机械与动力工程学院,上海200237

年份：2025

卷号：53

期号：3

起止页码：77

中文期刊名：化学工程

外文期刊名：Chemical Engineering(China)

收录：;北大核心:【北大核心2023】；

基金：国家自然科学基金重大资助项目(22393954)。

语种：中文

中文关键词：露点间接蒸发冷却;数值模拟;支持向量回归;联合仿真

外文关键词：dew point indirect evaporative cooling;numerical simulation;support vector regression;joint simulation

摘要：露点间接蒸发冷却在制冷循环系统运行中受环境因素影响较大,冷却性能不稳定,为保证制冷系统冷负荷满足设计要求,同时降低系统运行能耗。文中利用数值模拟和SVR(支持向量回归)模型,构建冷却器影响因素与冷却器出口温度的响应关系,样本数据通过实验和数值模型训练获得,同时通过MATLAB和TRNSYS联合仿真,构建蒸发冷却+机械补冷模式的机房制冷系统模式,分析回归预测模型对该系统运行能耗的影响。结果表明:SVR模型准确地构建了冷却器影响因素和出口温度的响应关系,R^(2)和E_(MSE)分别为0.9889、0.0671,平均绝对误差为0.1699℃。SVR模型能够更好地根据响应关系控制系统运行策略,以达到更为节能的运行效果。
Dew point indirect evaporative cooling is greatly affected by environmental factors in the operation of refrigeration cycle system,and the cooling performance is unstable.In order to ensure that the cooling load of the refrigeration system meets the design requirements,the energy consumption of the system operation is reduced.Numerical simulation and SVR(support vector regression)model were used to construct the response relationship between the influencing factors of the cooler and the outlet temperature of the cooler.The sample data were obtained through experiments and numerical model training.At the same time,through the joint simulation of MATLAB and TRNSYS,the refrigeration system mode of evaporative cooling+mechanical cooling mode was constructed,and the influence of regression prediction model on the energy consumption of the system was analyzed.The results show that the SVR model accurately constructs the response relationship between the influencing factors of the cooler and the outlet temperature.The R^(2) and E_(MSE) are 0.9889 and 0.0671,respectively,and the average absolute error is 0.1699℃.The SVR model can better control the operation strategy of the system according to the response relationship to achieve a more energy-saving operation effect.