基于预测的复合地源热泵系统控制方法实现

在冬冷夏热且夏季冷负荷远大于冬季热负荷的地区常采用带有冷却塔的复合式地源热泵系统,其控制策略存在极大的优化空间。文章提出了直接比较冷却塔和与土壤换热器相连的板式换热器的出口温度的控制方法,并通过人工神经网络预测板式换热器机组侧的出口水温来实现此控制方法。通过FLUENT软件建立复合式地源热泵系统动态数值模型,获取建立神经网络的数据,采用3层BP网络,建立了多个预测板式换热器机组侧出口温度的模型。研究结果表明,采用神经网络可以准确实现此预测,绝对误差不超过0.4℃。

Study on predictive control of hybrid ground source heat pump system

Hybrid ground source heat pump systems are applied in the cold winter and warm summer areas where the cooling load in summer is much more than heating load in winter,and a huge freedom exists in the control strategy.A new control strategy is proposed,by which the outlet water temperatures of the cooling tower and plate heat exchanger(PHE) coupled with ground heat exchanger(GHE)could be compared directly.The dynamic numerical simulation model of hybrid ground source heat pump system was constructed through sofrware FLUENT to getting the ANN datas,multi-models was construced by adopting three BP layer networks.Results show that ANN could realize the prediction exactly with the absolute error less than 0.4 ℃.