【摘要】 污水处理为高电耗行业,有效预测电耗对污水处理厂节能优化有深远影响。以广东省某中小型污水处理厂18个月的数据为例,采用BP神经网络建立了电耗预测模型,将提升泵、回流泵、鼓风机的运行时间和频率及OPR、MLSS、好氧池DO、缺氧池DO、进水流量Q、COD、氨氮等33个指标作为BP神经网络模型的输入变量进行初步拟合,并用遗传算法对BP神经网络进行优化。结果表明,BP神经网络适用于电耗的拟合计算,通过与遗传算法组合,平均误差低至0.001 6,预测精度较高。更多还原

【Abstract】 The sewage treatment is the industry of high power consumption.Predicting power consumption effectively will have a far-reaching impact on the energy-saving optimization of sewage treatment plants.This article studies the data during 18 months of a small and medium-sized sewage treatment plant in Guangdong Province though the establishment of BP neural network for the prediction model of power consumption,which use the run time and the frequency of lift pump,reflux pump,air blower as well as OPR,MLSS,aerobic pool DO,inlet flow rate Q,COD,ammonia nitrogen and other33 indicators as the input variables of BP neural network model for preliminary fitting.And then genetic algorithm was used to optimize the BP neural network.The results show that neural network is fit to calculate power consumption.It can improve the precision and the average error reduces to 0.001 6 by combining with genetic algorithm.更多还原