神经网络对微砂加重絮凝处理矿井水效果预测

Treatment Effect Prediction of Mine Water with Micro-sand Loading Flocculation by Neural Network

通过混凝沉淀单因素试验确定最佳反应条件根据试验结果,使用交叉验证法训练广义回归神经网络(GRNN)来预测微砂加重絮凝工艺的出水水质。结果显示,浊度和固体悬浮物(SS)去除率的实际值与预测值误差小于2%,说明GRNN具有良好的非线性拟合性,并广泛适用于微砂加重絮凝工艺。

Through the coagula reaction conditions. Using the tion and sedimentation of single factor test to determine the optimum cross validation training method named generalized regression neural network （GRNN） to predict ,,rater quality of micro sand adding flocculation process. The error between the actual value and the predicted value is less than 2%, which shows that GRNN has good nonlinear fitting property and is suitable for micro sand floeculation process.