基于GRNN与遗传算法的除湿机COP预测及其在故障诊断中的应用

COP prediction based on GA and GRNN and its application to faults diagnosis of dehumidifier

COP作为衡量制冷系统性能的一项重要指标,与故障密切相关,因此可作为故障诊断研究的切入点。基于此目的,本文进行基于COP的除湿机故障诊断探索研究,建立针对COP预测的GRNN模型,简化COP的实时测量,当发现模型预测值与正常值差别超出设定的范围时,则认为除湿机出现了故障。利用改进遗传算法对网络中的平滑参数进行优选,提高网络性能。诊断实例表明,基于COP的GRNN故障监测模型用于除湿机的故障诊断是可行的。

As one important index for measuring the performance of refrigeration, COP has close relationship with the faults, so it can be cut-in point for fault diagnosis research. For this target, carries out the fault diagnosis research based on COP for dehumidifier, sets up GRNN model for COP prediction,simplifying its real time measurement,when the differenc between the prediction and the normal value exceeds the set value, thinks that dehumidifier works wrong. With advanced genetic algorithm,optimizes the smoothing parameter,improves the network performance. The diagnosis example shows that GRNN diagnosis model based on COP for dehumidifier is feasible.