Modeling And Optimal Control Of Ammonia Desulfurization Based On Process Data

Thermal and coking industry,the implementation of SO2 reduction of environmental pollution is important to reduce,as in recent years,the country needs to build environment-friendly enterprises,to achieve ecological civilization construction country’s overall strategy,energy saving and green cycle plays an important Therefore,it is imperative to reform the ammonia desulfurization system in the thermal power generation and coking industries.However,because the chemical reaction in ammonia desulfurization is relatively complex and carries strong nonlinear characteristics at the same time,it is currently The method of desulfurization system still has the problem of low control level.At present,PID and its improvement methods are often used for control in engineering design,but the operating point can only be set for the general operating conditions.When the system operating conditions deviate from the set point,the system performance will decrease.Aiming at the problem that traditional modeling methods and control strategies are difficult to achieve satisfactory results,this paper uses the process data collected on site to establish a mathematical model of the desulfurization system,and uses the established mathematical model for predictive control,and uses simulation to verify the accuracy of the method.1.Collect process data from industrial sites and preprocess the data,including data standardization,detection and correction of outliers,and determination of steady-state operating conditions.Theoretical analysis of the chemical reaction of the ammonia desulfurization process is carried out,and the independent variables that affect the pH value of the absorption liquid in the desulfurization system are screened out from the collected data,which provides reliable data for system modeling and optimization of control methods.2.By analyzing the non-linear characteristics of the pH value of the absorption liquid,considering the use of the Hammerstein non-linear model structure to describe the characteristics of the pH control problem,the system model is established based on the pre-processed data,and the controlled variables of the actual industrial process only need to be maintained at a certain level.The characteristic of the target interval is achieved by adjusting the control weighting matrix of the performance index function in the predictive control to achieve interval predictive control,which reduces the frequent action of the ammonia valve,and the effectiveness of the method is verified by simulation.3.The ammonia desulfurization process has complex nonlinear characteristics,and only the Hammerstein nonlinear description is not accurate.Therefore,the preprocessed process data is used to establish a multi-scale Gaussian radial basis neural network model for the ammonia desulfurization system to model,the simulation results show that the pH value of the output absorption solution of the established system model can fit the pH value of the absorption solution of the test set well.Using the established multi-scale radial basis function(RBF)neural network model,the neural network optimization controller is designed to achieve precise control of the amount of ammonia injection,ensuring that the desulfurization efficiency meets the requirements and avoiding excessive ammonia injection.This article is verified by simulation.The effectiveness of the proposed modeling method and control method.