人工神经网络预测尾砂充填对地下水中锌含量影响

Prediction of Effects of Tailing Filling on Zn Content in Groundwater Based on Artificial Neural Network

[目的]研究了人工神经网络预测尾砂充填对地下水中Zn含量影响的可行性。[方法]采用误差反向传播神经网络(BP神经网络),建立了充填体放置时间、不同来源地下水水样、浸泡时间与重金属Zn含量之间的动态模型,预测广西某金属矿山全尾砂胶结充填体试块对地下水Zn含量的影响。[结果]当网络隐含层层数为1、节点数为7时,建立的BP神经网络预测模型能够较好地预测全尾砂胶结充填对地下水重金属Zn含量的影响。[结论]该研究为地下水环境的保护和矿区尾砂的安全处理处置提供了科学依据。

[ Objective ] The study aims to discuss the feasibility of predicting the effects of tailing filling on Zn content in groundwater based on artificial neural network （ANN）. [ Method] With error back propagation （BP） neural network, a dynamic simulation model showing the relationship of Zn content with placed time of tailing filling block, groundwater sources and soaking time was established to predict the effects of tailing filling on Zn content in groundwater from a metal mine of Guangxi. [Result] When the number of ANN hidden layer is 1 and the number of nodes is 7, the established BP neural network can accurately predict the effects of tailing filling on Zn content in groundwater. [Conclusion] The research can provide scientific references for the protection of groundwater environment and safe disposal of tailings in mines.