基于多目标文化差分进化算法的污水处理厂活性污泥单元多目标操作优化（英文）  

Multi-objective optimizing operational parameters for an activated sludge unit of a wastewater treatment plant using multi-objective cultural differential evolution algorithm

作　　者：于腾[1,2] 曹萃文[1,2] 顾幸生[1,2] 张亚坤[1,2] 

机构地区：[1]华东理工大学"化工过程先进控制与优化技术"教育部重点实验室 [2]华东理工大学信息科学与工程学院

出　　处：《计算机与应用化学》2016年第1期1-9,共9页Computers and Applied Chemistry

基　　金：国家自然科学基金资助项目(61174040,61573144);上海市自然科学基金资助项目(12ZR1408100)

摘　　要：污水处理厂活性污泥单元的操作优化旨在保证出水最佳水质的同时,提高过程的运行效率和经济效益。本文在活性污泥系统机理模型基础上,选取进水流量、出水生物需氧量(BOD)和运行成本作为优化目标,建立基于过程的单目标和多目标非线性规划(NLP)模型,包括一个单目标NLP模型、二个双目标NLP模型和一个三目标NLP模型。通过单目标三层混合文化差分进化算法(3LM-CDE)和基于Pareto准则的多目标三层混合文化差分进化算法(MO-3LM-CDE)对以上模型进行求解,得到单目标问题的唯一最优解和多目标问题的一组非支配解集。计算结果表明,三目标问题模型对实际工况的考量更为周全,并且在过程效率、净化质量和经济效益方面均表现出了良好的优化效果。Optimal operation of the activated sludge unit in a wastewater treatment plant aims to increase the processing efficiency and economic benefits, and ensure the high quality of the effluent water at the same time. Based on the mechanism models of an activated sludge reactor system, influent flow rate, effluent exit biochemical oxygen demand （BOD） concentration and operating cost are chosen as optimization goals, and single-objective and multi-objective nonlinear programming （NLP） models are studied, including one ＂single-objective＂ NLP model, two ＂bi-objective＂ NLP models and one ＂three-objective＂ NLP model. With the 3-layer mixed cultural differential evolution algorithm （3LM-CDE） of and the Pareto-optimal based multi-objective 3-layer mixed cultural differential evolution algorithm （MO-3LM-CDE）, unique solution is obtained for single-objective problem whereas a set of non-dominated solutions are obtained for multi-objective problems. The computing results show that solution of the ＂three-objective problem model＂ is more considerable, and performs overall well in processing efficiency, purification quality and economical benefit in wastewater treatment.

关 键 词：活性污泥单元 污水处理 优化 多目标 MO-3LM-CDE 

分 类 号：TQ015.9[化学工程] TP391.9[自动化与计算机技术—计算机应用技术]