基于GA-BP神经网络和改进粒子群算法的碰撞射流和冷却顶板复合空调系统优化  

Optimization of impinging jet ventilation combined with chilled ceiling air-conditioning system based on GA-BP neural network coupled with improved particle swarm optimization

作　　者：齐贺闯 叶筱 高延峰 亢燕铭[3] 钟珂[3] QI Hechuang;YE Xiao;GAO Yanfeng;KANG Yanming;ZHONG Ke(School of Mechanical and Automobile Engineering,Shanghai University of Engineering Science,Shanghai,China;Shanghai Collaborative Innovation Center of Intelligent Manufacturing Robot Technology for Large Components,Shanghai,China;College of Environmental Science and Engineering,Donghua University,Shanghai,China)

机构地区：[1]上海工程技术大学机械与汽车工程学院,上海 [2]上海市大型构件智能制造机器人技术协同创新中心,上海 [3]东华大学环境科学与工程学院,上海

出　　处：《东华大学学报（自然科学版）》2024年第1期110-117,共8页Journal of Donghua University(Natural Science)

基　　金：国家自然科学基金资助项目(51478098);上海市青年科技英才扬帆计划资助项目(21YF1415500)。

摘　　要：对碰撞射流和辐射顶板(IJV/RC)复合空调在不同室内负荷条件下运行时的室内热环境进行数值模拟,基于遗传算法-反馈(GA-BP)神经网络建立运行性能(吹风感R_(PD),头足温差Δt,空气交换效率e ACE,工作区平均温度t_(a))与设计变量(送风温度t_(s)、送风速度v_(s)、冷却顶板内表面温度t_(c)、房间负荷Q_(c))之间的预测模型,通过相关性分析确定设计变量对运行性能影响的显著性并排序。结果表明,增大v_(s)可使Δt降低,但R_(PD)增大;增大t_(c)有助于降低Δt和R_(PD),但t_(a)升高;为使t_(a)下降,可通过降低t_(s)来实现,但室内空气质量变差。为确保IJV/RC复合空调能在保证室内热舒适的同时提供良好室内空气品质,利用改进粒子群算法对复合空调的运行性能进行多目标同时优化,建立不同房间负荷条件下的设计参量最优匹配关系。研究结果可为IJV/RC复合空调的优化设计和运行控制提供理论指导。The matching relationship between the air supply and radiant_(s)ystems is the key to ensure the composite air-conditioning system operates efficiently.CFD numerical simulations are conducted to investigate the distribution pattern of The indoor thermal environment for the impinging jet ventilation combined with radiant_(c)eiling(IJV/RC)air-conditioning system with different operating load conditions was simulated.On this basis,t The GA-BP neural network was applied to develop the predictive models relationships between the operating performance(including draught discomfort R_(PD),temperature difference between the head and ankle levelΔt,air change efficiency e ACE,and average temperature of operation t_(a))and design variables(such as supply temperature t_(s),supply velocity v_(s),chilled ceiling temperature t_(c),and room load Q_(c)).The significance of each design variable on the studied operating performance was is then determined and ranked through correlation analysis.The results show thatΔt decreases with the increase of v_(s),but the value of R_(PD) increases accordingly.The increase of t_(c) is helpful for the decrease ofΔt_(a)nd R_(PD),but the value of t_(a) increases.The decrease of In order to reduce the t_(a),it_(c)an be achieved by reducing t_(s),but the indoor air quality becomes worse.To achieve the goals of providing good indoor air quality and indoor thermal comfort,a multi-objective optimization for the IJV/RC was conducted by applying the improved particle swarm optimization(PSO)algorithm,and the optimal combinations of the studied design variables corresponding to the room loads were developed.The current results can provide theoretical guidance for the design and operation control for the IJV/RC system.

关 键 词：碰撞射流通风 冷却顶板 GA-BP神经网络 粒子群优化算法 多目标优化 

分 类 号：TU834.3[建筑科学—供热、供燃气、通风及空调工程]