以实际试验数据为依据,通过网络结构优化建立了一个可用于磷酸铵镁法除磷模拟的BP神经网络模型。二维和三维的数值模拟表明,所建立的BP模型能够获得多因素条件下磷酸铵镁法除磷的变化规律和趋势,可为进一步的MAP法除磷及磷回收试验研究提供重要参考依据。
An BP neural network model which can be used to predict the phosphorus removal by magnesium ammonium phosphate process was established through network structure optimization based on the acquired experiment data. The two-dimensional and three-dimensional numerical simulations show that the BP inodel can acquire the basic rules and trends of magnesium ammonium phosphate method under multi-factor conditions and can offer important references to the further studies on phosphorus removal and recovery by magnesium ammonium phosphate process.