论文部分内容阅读
以某制药废水的升流式厌氧污泥床(UASB)-水解酸化池(HAR)-间歇式循环延时曝气活性污泥法(ICEAS)新型组合处理系统为背景,分析该系统效能,并建立遗传算法优化神经网络(GA-BPNN)模型对系统出水水质进行仿真预测,并利用建立的GA-BPNN模型对系统的运行条件进行优化研究。研究表明,在稳态运行的120 d,系统对废水COD和NH3-N去除率分别为98.6%和86.6%;GA-BPNN模型对出水COD和NH3-N的预测结果和实际监测值之间的平均绝对百分误差为5.55%和6.99%,能很好地应用于组合系统的出水水质预测管理中;GA-BPNN模型还可求解出系统的最优化运行条件,为工程实际操作提供了坚实的理论基础。
Based on the new combination treatment system of UASB - HAR - intermittent cycle aerated activated sludge (ICEAS) system for pharmaceutical wastewater, the system performance, The genetic algorithm optimization neural network (GA-BPNN) model is established to simulate and predict the effluent quality of the system. The GA-BPNN model is established to optimize the system operating conditions. The results showed that the removal rates of COD and NH3-N in the wastewater were 98.6% and 86.6% respectively in the 120 days of steady-state operation. The difference between the prediction results of the COD and NH3-N in the GA-BPNN model and the actual monitoring values The average absolute percentage error of 5.55% and 6.99%, can be well applied to the combined water quality prediction management of the system; GA-BPNN model can solve the system optimal operating conditions for the actual operation of the project provides a solid Theoretical basis.