板式换热器冷却水污垢热阻预测的偏最小二乘回归法

搭建板式换热器冷却水污垢热阻实验台，测得不同时间、流速和温度下天然循环冷却水（松花江水）中铁离子、氯离子、细菌总数、pH值、溶解氧、浊度、电导率等水质参数，随机取一组实验的水质参数作为输入变量，建立换热器冷却水污垢热阻预测的偏最小二乘回归模型，对板式换热器的污垢热阻进行预测。整个实验过程中，热水进口温度为43.5～44.5℃，冷却水进口温度为21.5～22.5℃，流速为0.104 m·s-1，当温度和流速发生变化时，则重新采取数据。经过计算，确立本模型应提取4个潜变量，由此建立了板式换热器冷却水污垢热阻预测模型。预测结果和实验结果最大相对误差在5.11%以内。结果表明偏最小二乘回归算法的污垢模型预测精度高，所建预测模型是合理可行的。

Cooling water fouling resistance prediction of plate heat exchanger based on partial least-squares regression

A fouling test system was set up to measure water quality parameters of natural cooling water circulation from Songhua river, such as iron ions, chloride ions, total bacteria, pH, dissolved oxygen, turbidity and electric conductivity at different time, flow rates and water temperatures.A partial least squares regression(PLS)model was built, with a set of data measured as the input variables, to predict the fouling characteristics of the plate heat exchanger, which are iron ions, chloride ions, total bacteria, pH, dissolved oxygen, turbidity, conductivity, etc.The input water quality parameters were selected randomly.In the experiment, the hot water inlet temperature was between 43.5℃ and 44.5℃, cooling water inlet temperature was between 21.5℃ and 22.5℃, and water velocity was 0.104 m·s-1.The calculation shows that the PLS model contains four latent variables.Compared with the experimental results, the maximal relative error of the model was in 5.11%.The partial least square regression algorithm of fouling model is reasonable and feasible.