空气源热泵结霜量预测技术及试验研究  

Experimental Study of Frost Prediction and Recognition Technology of Household ASHP

作　　者：王现林 徐耿彬 连彩云 杨旭东[1] Wang Xianlin;Xu Gengbin;Lian Caiyun;Yang Xudong(Tsinghua University,Beijing,100084;Gree Electric Appliances,Inc.of Zhuhai,Zhuhai,519070)

机构地区：[1]清华大学,北京100084 [2]珠海格力电器股份有限公司,珠海519070

出　　处：《制冷与空调（四川）》2023年第3期444-450,共7页Refrigeration and Air Conditioning

摘　　要：空气源热泵“监测外管温变化及运行时间”的化霜判断方法,利用了结霜现象对热泵系统运行参数的时空累积影响效应,需要累积到一定结霜量时才能准确识别,无法实时判断当前换热器的结霜情况。基于此痛点问题,提出一种结霜量预测技术,通过分析结霜过程相关原理,利用神经网络和空气源热泵系统特性,搭建结霜量预测模型,从而对结霜量进行准确识别。试验结果表明,搭建的结霜量预测模型能够实时计算室外换热器的结霜量,且与实际结霜量相比误差在15%以内,对探索更精细更灵敏更智能的化霜判断方式具有参考意义。Since the defrost judgment method"monitoring the temperature change of the outdoor heat exchangers and running time"of the air source heat pump uses the temporal and spatial cumulative effect of frost phenomenon on the operating parameters of the heat pump system,it can only be accurate after the frost accumulate to a certain amount,and cannot judge the current frost condition of the heat exchanger in real time.Based on this problem,this paper proposes a frost amount prediction technology,which analyzes the relevant principles of frost process and uses neural network and air source heat pump system characteristics to build a frost amount prediction model,so as to accurately identify the frost amount.The experimental results show that the frost prediction model built in this paper can calculate the frost amount of outdoor heat exchanger in real time,and the error is within 15%compared with the actual frost amount,there is reference significance for exploring a finer,more sensitive and intelligent way of defrosting judgment.

关 键 词：热泵 霜层预测 神经网络 结霜量 

分 类 号：TB615[一般工业技术—制冷工程]