基于小波神经网络的污水厂总磷预测模型

污水厂进水污染物与出水总磷的变化规律呈高度非线性，而传统机理预测模型需要依据经验设定大量参数，预测精度较低，预测相对误差处于15%~25%之间。为提高预测精度，以进水化学需氧量、总氮、氨氮、总磷、进水量5个进水指标与出水总磷浓度的映射关系建立小波神经网络预测模型。结果表明，小波神经网络模型模拟相对误差为9.87%，相较于机理模型，预测误差降低了5%~15%；同时模型收敛速度快，具有强大的非线性拟合能力，运行稳定性强，对污水厂实际运行中出水总磷预测有一定参考作用。

Research of TP Prediction Model in Wastewater Treatment Plant Based on Wavelet Neural Network

The variation of influent pollutants and effluent TP in wastewater treatment plant is highly non-linear. The traditional mechanism prediction model needs to set a large number of parameters based on experience， so that its prediction accuracy is low，the predicted relative error is between 15%-25%. In order to solve this problem，this paper establishes a wavelet neural network prediction model based on the mapping relationship between COD， TN， NH3-N， TP， influent volume and effluent TP. The results show that the relative error of the wavelet neural network model simulation is 9.87%， and the prediction error is reduced by 5%-15% compared with the mechanism model. In summary，the wavelet neural network model has fast convergence speed， strong non-linear fitting ability， strong operational stability， and higher prediction accuracy than the mechanism model， which can provide reference for predicting effluent TP in actual operation of wastewater treatment plant.