结合神经网络的热泵冷凝器模型优化  

Heat Pump Condenser Optimization Model Integrated with Artificial Neural Network

作　　者：范波[1] 于进洋[2] 李伟鸿[2] 李英[1] 

机构地区：[1]青岛理工大学环境与市政工程学院 [2]91286部队后勤部

出　　处：《建筑热能通风空调》2007年第6期46-50,共5页Building Energy & Environment

摘　　要：本文采用了按照制冷剂状态来划分分布参数模型微元的方法,建立了冷凝器的稳态分布参数模型,并将其计算结果与实验数据做了验证。在此基础上建立了和稳态模型相结合的BP神经网络和RBF神经网络模型,通过神经网络对稳态模型进行修正,有效地提高了冷凝器仿真模型在热泵系统中的仿真精度,并比较了用两种神经网络模型进行校正的差异。The steady-state distributed-parameter of the double-pipe condenser model is constructed by controlling volumetric method. Two simulation methods integrating the condenser model with RBF artificial neural network and BP artificial neural network were presented, as well as difference between calculation methods of two models is analyzed. It shows that this method is an effective way to improve the generalization and precision of the heat pump condensers general condenser model.

关 键 词：套管式冷凝器 稳态模型 径向基神经网络 BP神经网络 

分 类 号：TU833[建筑科学—供热、供燃气、通风及空调工程]