污水处理过程的多目标多模型预测控制方法研究  

Research on Multi-objective Multi-model Predictive Control Method in Wastewater Treatment Process

作　　者：谢生钢[1] 周立芳[1] 赵麟菱[1] 

机构地区：[1]浙江大学信息学院工业控制研究所,杭州310027

出　　处：《化工自动化及仪表》2008年第1期24-28,共5页Control and Instruments in Chemical Industry

摘　　要：针对污水处理过程节能降耗问题以及污水处理过程的高度非线性、强耦合、不确定性等特点,以基于活性污泥2号模型ASM2的A2/O污水处理过程为研究对象,提出了污水处理过程的多目标多模型预测控制方法。该方法首先采用聚类-PLS方法建立污水过程的多模型预测模型,然后构建了包含出水水质区间控制和经济性能指标的多目标优化结构的预测控制策略。仿真结果表明,与设定值预测控制方法相比,多目标优化预测控制策略在保证出水水质的前提下,能有效地节约能耗费用。For wastewater treatment process with strong nonlinearity, coupling and uncertainty, considering how to decrease power consumption ,a multi-objective multi-model generalized predictive control （GPC）method for wastewater treatment process based on activated sludge model NO. 2 （ASM2） with A^2/O technology was proposed. The predictive model was obtained by clustering and partial least square（PLS） ,and then a multi-object optimization predictive control method is constructed which includes zone control of the effluent water quality and economic object optimization. Compared with set-point predictive control,the result of simulation demonstrates the proposed multi-object optimization predictive control method can decrease the electrical power consumption while assuring the effluent water quality.

关 键 词：聚类-PLS建模 区间控制 广义预测控制 多目标优化 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]