污水曝气污染物浓度预测及其成本最优控制

Pollution concentration prediction of sewage aeration treatment and its optimal cost control

在污水处理厂水质检测中,化学需氧量(COD)和氨氮浓度(NH_3-N)等指标很难实现实时、精准、经济的测量,因而难以根据污染物浓度对污水处理过程中曝气需氧量和能耗进行控制。应用支持向量机(SVM)建立了预测模型,再用信息粒化时序回归方法来预测COD和NH_3-N的变化,实现了对COD和NH_3-N的在线软测量。在预测指标符合出水标准的前提下,建立了曝气成本数学模型;并考虑政府COD超量削减补贴,得到综合成本模型。应用NSGA-Ⅱ算法优化曝气处理工艺,可使吨水电耗降低44%~58%。

In the monitoring of water quality of sewage treatment works,it is difficult to get timely,accurate and economic monitoring indexes such as chemical oxygen demand(COD)and ammonia nitrogen concentration(NH 3-N).Therefore it is difficult to control aeration oxygen demand and energy consumption according to pollution concentration.Through establishing prediction model by Support Vector Machine(SVM)and predicting the variation of COD and NH 3-N by information granulation sequential regression process,on-line soft monitoring of COD and NH 3-N is realized.On the premise of the prediction indexes meeting the effluent standard,the mathematical model of aeration cost is established.In addition,with consideration of government subsidies for excessive COD reduction,the comprehensive cost model is established.Optimizing aeration treatment technology with NSGA-Ⅱalgorithm could reduce the power consumption by 44%~58%.