Clustering multi-model generalized predictive control and its application in wastewater biological treatment plant

The model mismatch problems can be avoided by using multi-modeling method based on cluster analysis, against using a single model for describing nonlinear process. The sub-models set output error is high precision and it can accurate fitting of non-linear characteristics of the system. Firstly, the generalized predictive controllers are designed based on the clustering multi-model of the wastewater treatment process, then, the overall control increment is synthesized based on the weight of each controller output. The clustering multi-model generalized predictive control strategy of the ammonia concentration is proposed in this study. The simulation results and the application results of the strategy used in the actual wastewater treatment plant verified its higher control accuracy. The effluent ammonia concentration can be controlled effectively. And also, it improved processing efficiency and the effluent quality, reduced the processing costs. © 2013 Asian Network for Scientific Information.