Study on hybrid modeling of urban wastewater treatment process

In view of the complexity of the actual wastewater treatment process, it is very difficult to establish an accurate mechanism model. The simple mechanism model cannot accurately describe all the characteristics of the actual process, and moreover the complex relationship between variables is difficult to be completely included in the mechanism model. In general, the prediction accuracy of the mechanism model is not high. Therefore, the data-driven modeling method is used to model the complex relationship between the actural process data. In this paper, the hybrid model of the activated sludge system in the wastewater treatment process is established, which combines the simplified BSMI mechanism model with the BP neural network. In the hybrid model, the BP neural network is used to compensate the predictive en'or between the mechanism model output and the actual output. The simulation results show that the hybrid model can overcome the shortcomings of the traditional mechanism model and improve the prediction accuracy. The predictive performance using the hybrid model is improved considerably, which validates the effectiveness and applicability of the proposed hybrid model into the urban wastewater treatment process.