Research On Modeling Method Of PH Value Of Wet Desulfurization System Of Coal-fired Unit

The development of society is inseparable from electricity.Today,my country’s power industry is still dominated by thermal power generation.However,under the theme of environmental protection,it is very important to strictly control the emission of pollutants such as SO2.Limestone-gypsum wet flue gas desulfurization(Wet Flue Gas Desulfurization WFGD)technology is currently the most common desulfurization method in thermal power plants.The pH value of the limestone slurry inside the absorption tower has a vital impact on the desulfurization efficiency,so precise pH control is required value.The pH value of limestone slurry presents the characteristics of large inertia,large hysteresis,and many influencing factors.Therefore,the establishment of a pH value model of limestone slurry is an important basis for improving the accuracy of pH control.Based on the historical data of a600 MW unit in a pithead power plant,a multi-model switching slurry pH model was established after data preprocessing and correlation analysis were used to select auxiliary variables.The main content is as follows:(1)Explains the research background and significance of slurry pH value modeling in many aspects.The development status of flue gas desulfurization technology and pH value modeling research is discussed.The process flow and technical characteristics of WFGD are introduced.Explains the important role of pH in WFGD.(2)Use Mutual Information(MI)theory to select auxiliary variables,and use mechanism analysis results and MI to determine auxiliary variables.In the experimental research,the data was preprocessed by coarse value elimination,data filtering and normalization.(3)Use MI to adjust the timing of auxiliary variables.Based on the unit load(high load,medium load,low load,load increase,load reduction)and the working status of the three slurry circulating pumps(at least two are in working status),13 working conditions are divided,and the working conditions are taken as Multi-model switching strategy.(4)Establish a single pH model using the Long Short-Term Memory(LSTM)based on MI timing adjustment.However,through experiments,it is found that the accuracy of a single model is lower under the nine operating conditions.In order to improve the accuracy,a sub-condition model of the LSTM network model based on MI timing adjustment is established for the nine operating conditions.The results show that a single model can maintain accuracy in four operating conditions,and the accuracy of the sub-operating model in the nine operating conditions is higher than that of the single model.In summary,the problem of modeling the pH value of the slurry is addressed.A single model and nine sub-condition models were established,and the validity of the model was verified.