Application Of ANN In Assessing And Predicting Ground Water Quality Of Haizhou Open-cut Mine In Fuxin

This article mainly applies RBF and Elman artificial neural network to assessing and predicting the ground water quality of Haizhou open-cut mining in Fuxin.In the assessment: the network models are built, the normalized standard matrix was taken as training sample, the best hidden nodes and the goal error are determined by matlab test, which can make the network achieve a relatively best performance; Then, training the established network and using the monitoring data as test samples to obtain the evaluation result. By result, the ground water of this mining was seriously polluted, class of its pollution isⅣ-Ⅴ,and the main factor is inorganic salt.In the prediction: Take the last ten years' monitoring data of ninth point as samples, the first six groups of datas were selected for training samples, and the last four for testing samples. By interpolation, training, verifying and testing the model, a feasible and objective prediction model is obtained. The prediction result shows that with the growing years, ground water pollution is becoming more and more heavier. After 15 years, except concentration of Fe3+ did not overweight the standard, the remaining six factors all exceeded it in different degrees. In order to prevent pollution becoming further intensifies, the effective protective measuresmust must be implemented.