摘要：

作为污水处理的核心工艺，生物曝气环节的稳定性受进水水质、水量等因素的影响较大，且电能消耗高。对曝气过程进行优化控制有利于提高污水处理系统的性能。提出一种基于工况感知-自主决策-性能评估方法的污水处理曝气优化控制策略。将K-means聚类算法与注水原理相结合，对入水数据进行入水工况感知；采用最小二乘支持向量机（LS-SVM）与神经网络反向传播算法（BPNN）建立软测量模型，并结合PSO全局寻优算法求解当前入水的溶解氧浓度优化设定值；将所得曝气池溶解氧浓度优化设定值输入仿真模型中进行性能评估，由仿真评估的结果优化更新工况感知与决策控制部分。经仿真验证，优化系统在出水达标且出水水质与原系统相差不到2%的情况下，经济指标下降10% ~15%，节能效果显著。

关键词: 污水处理, 曝气, 工况感知, 决策控制, 性能评估, 算法, 模型预测控制

Abstract:

As core process of wastewater treatment， the stability of biological aeration is greatly influenced by influent quality and quantity， and the power consumption is high. The optimal control of the aeration process is beneficial to improve wastewater treatment performance. A sewage treatment aeration optimization control strategy based on condition perception- autonomous decision-performance evaluation method was proposed. The K-means clustering algorithm and the principle of water injection were used to perceive the water ingress conditions for the water inlet data. The fitness function and constraint function of the global optimization algorithm PSO were established using LS-SVM and BPNN. The obtained optimized dissolved oxygen set value in the aeration tank was input to the simulation model for performance evaluation. The results of the simulation evaluation were used to optimize and update the working condition perception and decision control section. The simulation verification showed that the energy consumption was reduced by 10%-15% with significant energy saving effect， when the effluent composition met the standard， and the difference between optimized system and original system was less than 2%.

Key words: wastewater treatment, aeration, condition perception, decision control, performance evaluation, algorithm, model-predictive control

中图分类号: 

• TP183

• X703.1