一种基于PSO-RBF网络算法的热舒适性指标预测新方法  

A new method for prediction of thermal comfort index based on PSO-RBF network algorithm

作　　者：陆烨 朱其新 周敬松 朱永红 LU Ye;ZHU Qixin;ZHOU Jingsong;ZHU Yonghong(School of Environmental Science and Engineering,SUST,Suzhou 215009,China;School of Mechanical Engineering,SUST,Suzhou 215009,China;School of Mechanical and Electronic Engineering,Jingdezhen Ceramic Institute,Jingdezhen 333001,China)

机构地区：[1]苏州科技大学环境科学与工程学院,江苏苏州215009 [2]苏州科技大学机械工程学院,江苏苏州215009 [3]景德镇陶瓷大学机电工程学院,江西景德镇333001

出　　处：《苏州科技大学学报（自然科学版）》2020年第1期73-78,共6页Journal of Suzhou University of Science and Technology（Natural Science Edition）

基　　金：国家自然科学基金资助项目(51875380,51375323,61563022);江苏省“六大人才高峰”高层次人才资助项目(DZXX-046);江苏省产学研前瞻性联合研究资金项目(BY2016044-01);江西省自然科学重大基金资助项目(20152ACB20009)。

摘　　要：研究了一种新的网络算法,该算法采用粒子群优化(PSO)算法对径向基(RBF)神经网络的权值进行优化,并将这种算法运用于热舒适性指标的预测中,通过建立PMV指标预测模型,实现对PMV指标的智能预测。通过Matlab仿真计算,结果表明,基于粒子群优化的径向基网络(PSO-RBF)的预测方法误差精度更小,较之未优化的RBF网络,误差精度提高了79.5%。采用粒子群优化的径向基网络(PSO-RBF)对PMV指标进行预测是完全可行的,将PSO-RBF算法应用于基于PMV指标的空调系统控制中,可为PMV指标的实时控制打下基础。In this paper,we have discussed a new network algorithm,which is applied to the prediction of PMV index so as to solve the difficulty of traditional method in predicting PMV index.The weight of radial basis function(RBF)neural network was optimized by particle swarm optimization(PSO)algorithm.The algorithm was applied to the prediction of thermal comfort index,and the prediction model of PMV index was established to realize the intelligent prediction of PMV index.The results of MATLAB simulation show that the prediction method based on PSO-RBF neural network has less error precision.Compared with the unoptimized RBF neural network algorithm,the error accuracy is improved by 79.5%.It is feasible to predict PMV index by using PSO-RBF network.Applying PSO-RBF algorithm to air conditioning system control based on PMV index can lay a foundation for real-time control of PMV index.

关 键 词：PMV指标 RBF神经网络 粒子群算法 HVAC系统 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]