Intelligent Control Of Power Plant SCR Flue Gas Denitration System Based On Mechanism Modeling

Along with the growth of comprehensive strength,the nation’s consciousness of environmental protection is increased,the limitation of allowed emission of nitrogen oxide from coal-fired power plant becomes more and more rigorous,these new polices bring additional pressure to electric producing industry and make higher requirements to SCR denitration system SCR denitration system belongs to object system that have feature of great inertia and large delay;and the process of SCR denitration involving chemical reactions,the way of denitration can be classified as chemical method;also the reaction of ammonia and nitrogen oxide during denitration process can be affected by complex situation of flue gas flow field,the SCR denitration system can be summarized as complicated system with above conditions,and is difficult to control.Among the old units that have already finished the optimal denitration transformation and the newly built units,the problem of low control quality and poor rate of automated operating is widespread,this paper analysis the reason of these problems from the perspective of control,an optimum proposal that control valve and concentration transmitter are added is presented,and the corresponding control strategy based on process of control is given,the purpose of improving denitrification efficiency and decreasing the amount of ammonia escaping should be achieved by equipment modification and optimization of control scheme,and finally achieve the goal that the operation of SCR denitration system can meet the need of economic request and the environmental protection policies.The main research contents are as follows:1.To meet with the need of high-precision control,it is necessary to add equipment to achieve better control ability for control subsystem and get better measuring ability for measurement system.An optimal dividing proposal is given and the scheme that adding control valve and concentration transmitter is presented based on above.The scheme that use a switch as core node with programmable logical controller and the scheme using linear form with industry computer as controller are also give.The purpose of these measurement is to make enough room for advanced control algorithms and strategies without changing the object characteristics,and it is the basis of further optimization.2.By induced the industrial field data into mass balance equation,the transfer function model of SCR denitration system based on denitration reaction mechanism is give.To proving the reasonable of the model,comparing with the result of step simulation and the field data,including several situation as follows:the concentration of nitrogen oxide and the amount of ammonia slip while the ammonia injection control valve step;the concentration of nitrogen oxide and the amount of smmonia slip while the gas flow step;the concentration of nitrogen oxide while nitrogen oxide concentration at reactor inlet step.The results indicate that the trend of the simulation and the real data is same and the significance is proved.3.On the basis of installation of control valve and concentration transmitter,the neural network-feedback linearization(NNFL)algorithm is proposed to control the main ammonia injection valve,meanwhile the receding horizon-single neuron adaptive-artificial fish swarm(RSNAAFS)algorithm is proposed to control branch pipe controlling valve.The simulation indicate that the optimal control strategy and advanced algorithm give above have enough ability to make the denitration systen operation automatically,and the control quality is better than traditional control scheme in aspect of response speed and control accuracy,and prove its significance.4.By using the design thought of restricted boltzmann machine,the multi-output cascade field control strategy is give,based on this,the single neuron adaptive simulated annealing(SNASA)is proposed and is used as controller to give deep structure control to the denitration system;the deep particle swarm optimization(DPSO)algorithm is proposed and is used to give deep control to the denitration system,for the sake of comparing,the control quality factor is designed.The simulation result prove that the control strategy and deep control algorithm have better robustness,and is better than conventional scheme.5.Aim for higher precision control of SCR denitration system,the optimal sliding mode control which the amplitude and changing rate of control variable is limited is proposed;using differential evolution algorithm to bring the ability of adaptive and high speed to the sliding mode control by the help of designing auxiliary sliding surface.Based on the proposals mentioned above,the adaptive optimized sliding mode control by using cascaded differential evolution algorithm is presented,and meet the need of object that have great inertia and great delay such as SCR denitration system.The simulation result indicate that the optimal proposal can meet with the need of the denitration system and have good robustness,the control quality is better than traditional PID algorithm.