摘要： 针对活性污泥法污水处理过程的特点，提出了一种基于ANFIS 的建模方法。以污水处理厂水质数据为样本，首先利用主元分析(PCA)实现输入变量的降维和去相关，随后采用减法聚类算法生成初始模糊推理系统，进而采用ANFIS 建立活性污泥法污水处理过程模型。仿真实验表明：基于ANFIS 的建模方法是有效的；通过仿真结果比较，PCA-减法聚类-ANFIS 网络的拟合能力及收敛速度均优于PCA-BP，更适合作为活性污泥法污水处理过程模型。

关键词: 活性污泥法；主元分析；减法聚类；ANFIS

Abstract: Aiming at activated sludge process’ characteristic, a modeling algorithm based on ANFIS is presented. Using water quality data of wastewater treatment plant as sample, first, principal component analysis (PCA) is used to reduce the dimensionality of input variables and to remove colinearity; Second, the subtractive clustering is used to build initial fuzzy inference system; Then ANFIS is adopted to build activated sludge process model. Simulation experiment results show: the modeling algorithm based on ANFIS is effective; compared with the simulation results, the fitting ability and convergence speed of the PCA-subtractive clustering-ANFIS network are all better than the PCA-BP’s, and much more suitable as activated sludge process model.

Key words: Activated sludge method; Principal component analysis (PCA); Subtractive clustering; ANFIS