Design Of Intelligent Monitored Control System For Wastewater Treatment Based On Prediction Of Improved Neural Network

This project was researched in the wastewater treatment plant of New District in Qingdao.First,we made a detailed introduction and analysis of the design process,equipment and facilities in this plant,and then we made the overall design of this intelligent wastewater treatment monitoring system according to Industry 4.0 and Cyber-Physical System.We divided the system into four parts,the first part is perception and communication,the second part is fusion processing,the third part is human-computer interaction,the fourth part is actuator.The part of fusion processing can use neural network algorithms to predict DO concentration in Aeration Tank,it will adjust oxygen supply to achieve accurate aeration according to the predicted value of DO,it will optimize the effluent quality.The part of human-computer interaction and actuator use multi parameters monitoring to realize the stable operation of the equipment and ensures the system's security.Because treatment process uses A2/O method,its main mechanism is that microorganism in sludge decompose pollutants through metabolism in the aeration tank,it can only happen in appropriate oxygen condition.The control of DO is crucial,so a new optimal method based on neural network of DO control was used in this project.First,through the test and the historical data,when the effluent quality was good,we got different indexes of wastewater and value of DO in aeration tank.We used these data to train BP neural network which was optimized by PSO algorithm and we successfully realized the accurate prediction of value of DO.The smooth operation of the wastewater treatment equipment is accomplished by the slave computer and the host computer.In this project we set up acquisition devices for each process section in accordance with order.These devices can make a comprehensive monitoring for wastewater plant,got the key parameters and status of critical equipments.We designed the communication network for the host computer,slave computer and acquisition devices.The slave computer uses PLC as the core,the program of PLC was develop by STEP7,used PID algorithm to predict value of DO.Host computer was made by C#,it has many functions: login,display of real-time data,trouble alarm,data query and user management.And we integrated the prediction system of DO made by Matlab into the host computer,it can give the right value to slave computer by communication to set value of DO.The system we designed works well in wastewater treatment plant.Using detector to test water proves that it optimizes the effluent quality than before.it proves that this system is very effective.