Study On Slurry PH Control Method In Wet Desulfurization Process Based On Neural Network

The absorber control system is the core subsystem of wet flue gas desulfurization control system,and the PH value of the slurry inside the absorber is an important parameter index of flue gas desulfurization,which directly affects the desulfurization efficiency of the whole system.Due to the hysteresis,nonlinear,multi-variable strong coupling and other characteristics of the grout PH value in the process of control,the PH value of the grout in the absorber often appears in the process of control,which will not only affect the efficiency of wet desulfurization,but also cause unnecessary loss of the grout,and affect the quality of the desulfurization byproduct gypsum.Therefore,accurate control of the slurry PH value in the absorber is a necessary condition to ensure the normal and efficient operation of the wet desulphurization system.First of all,this thesis will introduce the basic structure and working principle of the absorber control system and the mechanism of chemical reaction in the absorber during flue gas desulfurization.It is pointed out that the nonlinear,large hysteresis and multivariable coupling characteristics exist in the control process of PH value of slurry in absorber.Secondly,the Gated Recurrent Unit(GRU)neural network is used to establish a corresponding prediction model for the PH value of the grout in the absorption tower to predict the change of the PH value of the grout.By collecting desulphurization data and pretreating it with noise reduction and screening,the data set is divided,and then the model training,parameter tuning,model testing and other stages are carried out.Finally,the prediction performance of the slurry PH value prediction model established by using different neural networks is compared according to the performance evaluation index of the model.The prediction results show that the GRU neural network model selected in this thesis is the best in the prediction performance of slurry PH value.Thirdly,the PH value of the grout was predicted by the established neural network model,and the PH value of the grout at the next moment was compared with the PH value of the grout under the current desulfurization condition to determine the change trend of the grout PH value.According to the variation trend of PH value of grout,an intermittent pulse method was proposed to accurately adjust PH value of grout.Through the intermittent pulse method unique adjustment of slurry supply gate strategy,to achieve timing and quantitative control of the amount of slurry into the absorption tower,the nonlinear and large lag complex control characteristics of the slurry PH value in the control process can be completely eliminated,and finally make the slurry PH value control precision within the range of ±0.02,fully meet the requirements of wet desulfurization process production.Finally,through the construction of Client-Server(C-S)architecture in the absorption tower slurry PH intelligent control system platform,on the premise of realizing communication with the PLC,the neural network prediction model and intermittent pulse method is combined.The precise control of the grout PH value in the absorber is realized through the interaction of data under the platform,and the feasibility and reliability of this method for the control of grout PH value are verified by practice.