Optimized Dissolved Oxygen Control For Greenhouse Gas Reduction In Wastewater Treatment Process Of Paper Mill

Since the beginning of the 20th century,climate change has increased its attentions.As one of the main sources of greenhouse gas?GHG?emissions,wastewater treatment produces a large number of GHG directly to the atmosphere,accelerating the global climate change.The papermaking industry ranks second in the discharge of industrial wastewater in the country,and COD discharges account for 33%of the COD emissions of industrial wastewater in the country.Being the dominate GHG emission industry,the papermaking industry is under huge pressure of GHG emission reduction.In the process of papermaking wastewater treatment,GHGs such as CO₂,CH₄ and N₂O are produced and directly discharged.The energy consumption of wastewater treatment operations also indirectly generates CO₂.Due to the lack of on-line monitoring and reduction of GHG,a reliable approach to reduce emissions based on optimal control of dissolved oxygen?DO?concentration is proposed.Primarily,on-line detection model if established,based on activated sludge model no.1?ASM1?and conservation of mass and dynamic relations between microbial reactions,including on-site source?endogenous decay,biochemical oxygen demand?BOD?removal,nitrification and denitrification process?and off-site source?sludge treatment,aerator and pump power,chemicals consumption?.Secondly,the impacts of DO on papermaking wastewater treatment was determined.Results show that when the DO concentration was less than 2 mg/L,the lower the DO concentration was,the higher the effluent pollutant concentration was,and the lower the total greenhouse gas emissions were.In this study,anoxic/oxic?A/O?processing site of a paper mill in Guangzhou was the object.Firstly,the DO proportional–integral?PI?control with constant set value was designed.The simulation results showed that the effluent pollutants decreased under the control of fixed PI parameters,but with poor control accuracy.Secondly,aiming at improvement of control accuracy,the influent and pollutant concentrations were considered into the back propagation neural network-proportional-integral?BP-PI?control for training,and the dynamic PI parameters were obtained,which can improve the control system accuracy.Finally,based on the mechanism of action of DO,and treating DO as control variable,the idea of hierarchical optimization control that design for papermaking wastewater to reduce greenhouse gas emissions is proposed.This intelligent optimization control scheme is called adapted-kernel-regression back propagation neural network proportional–integral?AKRBP–PI?control.In order to prevent effluent pollutant concentration from exceeding the limit,the adaptive regression kernel function was used to describe the functional relationship between effluent pollutant and DO concentration,and the prediction of effluent pollutant contents were used as the effluent constraint layer.According to the function relation of the constraint,the set value of the optimization is solved by the genetic algorithm?GA?as the optimization setting layer.In the tracking control layer,the optimized set value,inlet flow and pollutant concentration are input into the BP-PI control,and the real-time optimized set value is used for neural network tracking control.The simulation results show that,although the control accuracy of AKRBP-PI control is lower than that of BP-PI control,and the effluent BOD,total nitrogen and ammonia nitrogen are higher than that of PI control and bp-pi control,there is no over-limit phenomenon.Under the AKRBP-PI control,GHG was reduced by 8.55%compared with the open-loop control.Among GHG sources,the reduction of GHG mainly comes from the reduction of power consumption of the aerator?35.9kg/h?.However,under PI control and BP-PI control,compared with open-loop control,the total greenhouse gases increased by 8.09%and 4.13%,respectively.The AKRBP-PI control scheme can effectively reduce the greenhouse gas emissions while ensuring that the effluent pollutant concentration does not exceed the limit.