Research On Dynamic Soft Measurement Of No_x Formation In Coal-fired Units

The SCR flue gas denitration technology is a relatively mature means for controlling NO_x emissions in coal-fired units.Timely and accurate acquisition of NO_x production is one of the keys to improving the quality of SCR denitrification control.Soft measurement technology is one of the effective ways to solve the problem of meter measurement lag.In the thermal process,the boiler and other equipment have a large thermal volume inertia,which leads to a large inertia and large delay between the NO_x production amount and its influencing factors.Therefore,the static soft measurement model cannot accurately obtain the NO_x production amount.Therefore,the dynamic soft-measurement model of NO_x production is studied in this paper.The main research contents are as follows:(1)The background significance of the dynamic soft-measurement model of NO_x production and the current research status of dynamic soft-measurement are summarized.Firstly,the background significance of the dynamic soft-measurement model of NO_x production is introduced from the aspects of SCR denitration control system and environmental protection.Then,according to the current methods and application fields of the dynamic soft-measurement model,the problems of the dynamic soft-measurement model of NO_x production are summarized,including the hysteresis of thermal parameter measurement instruments,the inconsistency of time between sample data,and the dynamics of actual production processes.(2)The auxiliary variable selection and data preprocessing of NO_x production are mainly studied.First,the influencing factors are selected according to the NO_x production process.Then,the historical data of each variable is obtained by VeStore,the data is filtered by smoothing,and eliminated the abnormal values by pauta criterion.Finally,the influencing factors of NO_x production are analyzed by KPCA,and the redundant auxiliary variables are eliminated.(3)In order to adapt to the dynamics of the production process and the asynchronousness of historical data,a dynamic soft-measurement model of NO_x production based on LSSVM is proposed.First,the structure of the dynamic soft-measurement model of NO_x production is given.Then,k-nearest neighbor information is used for delay selection of auxiliary variables.After the asynchronous problem between data is eliminated according to the determined delay parameter,the order of each auxiliary variable is obtained by genetic algorithm and k-nearest neighbor mutual information method respectively.Finally,a dynamic soft-sensing model of LSSVM is constructed by using a new set of input variables covering the delay and order information of each variable.The model is corrected by past predictions of NO_x production,the prediction effects of two dynamic soft-measurement models and static soft-measurement models on NO_x production are compared.(4)In order to solve the problem that the SCR denitration control is inaccurate due to the lag of NO_x production measurement,an SCR control system based on dynamic soft measurement of NO_x production amount is proposed.First,the composite control scheme of SCR denitration system is introduced.Then,the NO_x production obtained by the dynamic soft-measurement model is fed forward as an SCR denitration control system.Finally,it is compared with a control system based on actual measurement data.