Research On Intelligent Prediction Of Ammonia Injection And Coupling With Flow Field In SCR Denitration System Of Boiler

At present,the main form of power production in China is coal-fired power.In 2020,China’s coal-fired power generation will account for 60.8%of the total national power generation.Due to the characteristics of coal itself,the flue gas produced in the production process of coal-fired power plant contains a lot of air pollutants.The state has made strict regulations on the emission standard of flue gas,and the ultra-low emission requires that the emission of flue gas from coal-fired power plant should not exceed 50mg/m3.SCR denitrification technology is mainly used in power plant to remove NOx in flue gas,that is,SCR denitrification device is installed between economizer and air preheater,NH3 is injected into flue gas,and NOx in flue gas reacts with N2 and H2O under the action of catalyst.However,in the process of SCR denitration system,due to the design and operation of denitration system,NH3 injected into flue gas can not completely react with NOx in flue gas.Unreacted ammonia reacts with SO3 to form ammonium bisulfate(NH4HSO4).Ammonium bisulfate will liquefy gradually from gaseous state at about 260℃,causing the resistance of the air preheater to increase,which seriously affects the performance of the air preheater and the operation safety of the unit.How to reduce ammonia escape is the key to reduce the formation of ammonium bisulfate.The main causes of ammonia escape are excessive ammonia injection and partial incomplete reaction caused by uneven mixing with flue gas.Therefore,this paper studies the optimization of ammonia injection and flow field.Firstly,through collecting the real operation data of 660 MW units in a power plant and building an analysis platform,the relevant laws of historical operation data related to denitration system operation are analyzed,and the relationship between ammonia injection amount and load is determined.The results showed that the ammonia injection amount of the SCR denitrification system of the studied unit decreased with the increase of the load in the load range of 200-375MW,and in the load range of 375MW-660MW,the amount of ammonia injection increased with the increase of the load.This is mainly due to the varying rates of flue gas flow and NOx changes with the load under different loads.According to the analysis results of denitration-related factors,this paper divides the working conditions,and builds a platform for optimizing the amount of ammonia injection under different working conditions.The platform is used to find the optimal amount of ammonia injection under the current operating conditions corresponding to the historical operating conditions,and realizes the guidance of optimizing steady-state operating conditions based on the big data analysis method.This paper also trained the ammonia injection prediction model of SCR denitrification system based on neural network method of L-M algorithm,and predicted the ammonia injection demand according to the parameters of SCR inlet flue.According to the data test,the average prediction error between the predicted ammonia injection volume output by the trained model and the actual ammonia injection volume data in the field is within 5%.In addition to the optimization prediction of ammonia injection,it is also an important measure to realize the uniform mixing of NH3 and NOx.The effect of various flow sharing schemes of SCR denitrification system of boiler is studied by numerical simulation.The results of component simulation show that the uniformity of velocity field and NH3 distribution above catalyst do not match completely.By adjusting the arrangement of ammonia spray grid,the uniformity of NH3 distribution is further improved,and the concept of determining ammonia injection quantity by flux is proposed.