Study On The Multi-objective Of Anaerobic Digestion Process In Paper Wastewater Based On Multi-intelligence-algorithm

With the rapid development of our economy, urbanization and industrialization process accelerated, under the new situation of energy conservation and emissions reduction, the problem of paper wastewater has become one of the important issues which affected the economic development and people’s daily life. In order to accelerate the economic transition in our paper industry by transform the development from a rapid development to a sustainable development, a multi-objective optimization in paper wastewater treatment process is necessary especially for the anaerobic process. In this way, the wastewater treatment equipment will be not only highly-efficiency but also kept stable, at the same time, it is a important stage to realize the emission reduction and sustainable development of energy saving. At first, a comprehensive analysis of the intelligent algorithms which have been applied in WWTP is shown in this paper, the basic concepts and structures of cloud model, GA-BPNN and NSGA-II are also introduced. Then a systematic study on building the multi-objective optimization model is given, cloud model, GA-BPNN and NSGA-II were applied for this purpose. Several extended and exploratory researches are also made to provide reference for improving the efficiency of wastewater treatment. The main research contents and results are shown as follows:1. A series work of selecting the controlling equipment and configuration is taken after analyzing the characteristics and treatment requires in paper wastewater anaerobic treatment process. An automatic monitoring system of wastewater treatment is built up successfully in our laboratory.2. After analysis the principle of pH adjustment, traditional PID controller and cloud model theory, a traditional PID controller and a cloud controller are built to control the pH and also they simulate in MATLAB. The simulation results show that the cloud controller performs better in adjusting pH value than the traditional PID controller. With the help of OPC technology, an real-time information communication between MATLAB and MCGS is built to verify the effect of the cloud controller in actual operation, and the result indicates it has a good performance.3. For the relationship between the CODin, pH, temperature, the flow rate, the CODeff and Qgas during the anaerobic process, BPNN predict models and GA-BPNN predict models, which are used to predict CODeff and Qgas,are built after analyzing the structure of BPNN and genetic algorithm. The MAPE of CODeff and Qgas BPNN predict models are 61.7742% and 10.5959%, the RMSE are 447.6696 and 2.4325. However, the MAPE of CODeff and Qgas GA-BPNN predict models are 21.7263% and 7.6443%, and the RMSE are 196.0658 and 2.0094. The results show that compared with BPNN predict models, GA-BPNN predict models perform better and are more suitable to build up the predict models.4. A multi-objective optimization model is built by GA-BPNN and NSGA-II, applied to the anaerobic process of paper wastewater treatment, the result shows this optimization model is a good optimizer which can solve the multi-objective optimization problem effectively. The optimization result provides a reference for design and operation of anaerobic digestion technology, and this method combining neural networks with NSGA-II is also useful to solve other multi-objective optimization problems.