Research And Application Of SCR Control System For Coal-fired Power Plants

In recent years,China's economy has developed rapidly,and the demand for electric energy has been increasing.However,the environmental pollution caused by coal-fired power generation is becoming more and more serious.Among them,nitrogen oxides(NOx)is one of the most concerned pollutants,although currently the SCR control system can achieve certain effects,but it is difficult to achieve actual operation requirements in terms of stability and economy.In order to comply with the national environmental protection policy and improve the control effect of the SCR system of coal-fired power stations,it is necessary to carry out refined control of the ammonia injection amount based on the high-precision SCR system model.In this paper,the actual operation data in the thermal power station is used,and the SCR system model is established based on the deep learning.Based on this,the ammonia injection model predictive control strategy is proposed to improve the stability of the system while ensuring the NOx emission.The problem of uneven distribution of NOx concentration in the gas has developed a valve control system,which effectively solves the problem of uneven distribution of NOx concentration at the outlet and improves the operating efficiency of the SCR system.The main research contents of this paper are as follows:Based on the premise mechanism of SCR system denitration and actual operation characteristics,the data model of SCR system is established,including improved BP neural network model and long-term short-term memory(LSTM)neural network model.The parameters and optimization functions in the modeling process are selected.The method was introduced,and the actual operation data of the coal-fired power station was used as a test set to compare the prediction accuracy of the two models.Aiming at the shortcomings of the current SCR denitrification control strategy and control method,the model prediction controller of SCR system is built with LSTM neural network model.The particle optimization algorithm is used to optimize the control quantity,in order to verify its control effect,the actual operation data is used to simulate and verify the control quality.After analyzing the simulation results,we found that the predictive control has better control effect than the traditional control mode,the ammonia injection amount is adjusted more rapidly,and the outlet NOx concentration fluctuation is significantly reduced.From the perspective of the actual operation of the SCR denitration system of coal-fired power stations,the factors affecting the denitration effect are analyzed,and the SCR refined ammonia control system of coal-fired power station is developed,including the ammonia-based main valve control system based on model predictive control,and the ammonia injection valve control system for exporting NOx concentration is applied to a coal-fired power plant to verify its actual control effect,and compared with the original control system of the power plant.The results show that the control effect of the system is much better than the original system.The effect of NOx concentration fluctuation at the outlet under variable load is obviously improved.The standard deviation of NOx concentration at the outlet of SCR system is reduced from21.7mg/Nm~3 to less than 5mg/Nm~3,and the system stability and denitration efficiency are also greatly improved.