基于偏最小二乘法对水库pH值预测与应用

Prediction of reservoir pH value based on partial least square method

根据水库中藻类数量和pH值数据,基于偏最小二乘法对影响pH值变化的5个因素进行建模分析,结果表明,各藻类数量对pH值的影响程度不同,小球藻>小环藻>辐节藻>骨条藻>范式冠盘藻。通过实测数据验证,偏最小二乘法回归模型的拟合能达到满意的效果,预测pH值与实测对比,模型可以很好地表征赫章公鸡寨水库中藻类对pH值的影响趋势,但是不能精确地预测实测值,这主要的原因可能是影响pH值变化的因素还有很多,比如水温、溶解氧、浊度、高锰酸盐指数、年降雨量等。

Based on the data of algae quantity and pH value in the reservoir,modeling and analysis of five factors affecting the pH value were conducted based on the partial least square method.The results showed that the impact of different kinds of algae on the pH value was different(Chlorella>Microcystis>Actinophyta>Skeletonema>Stephanodiscus).The model was verified by actual data,and the fitting effect of the partial least square regression model was satisfactory.The predicted pH value was compared with the measured value,and the model could well characterize the impact trend of algae quantity on the pH value in the Gongjizhai Reservoir in Hezhang,but it could not well predict the measured value.The main possible reason for this is that there are many factors affecting the pH value,including water temperature,dissolved oxygen,turbidity,permanganate index,annual rainfall,etc.