应用神经网络对地下水反硝化效果预测与验证

Prediction and Verification of Groundwater Denitification by Artificial Neural networks

应用人工神经网络对生物脱氮工艺进行预测并验证。采用工艺参数如COD、NO3-N、NO2-N、DO等作为输入节点，COD、NO3-N作为输出节点。结果显示神经网络能够较好地预报出水的水质参数。出水硝酸盐和COD预测结果与试验结果符合得较好，相对误差分别在10％和5％范围内。

The artificial neural networks（ ANN ）was applied in denitrification process in ground water and this prediction and verification of this process were investigated. The parameters such as COD, NO3 - N, NO2 - N, DO has been used for input, whereas the nitrate and COD used for output. The experimental results show that the ANN is able to predict the output water quality parameters very well - including nitrate as well as nitrite and COD. Most of relative error of NO3 - N and COD were in the range of ±10% and ±5% respectively.