Research On Combustion Optimization And Low NO_x Emission Control For Utility Boilers

Nitrogen oxide?NO_x?is one of the major air pollutants released from coal-fired plants.Nowadays,low NO_x combustion combined with selective catalytic reduction?SCR?has become a general DeNOx strategy in utility boilers.However,using low NO_x combustion method will reduce the economy of boilers under some operational conditions;thus,the investigation of combustion optimization is necessary to simultaneously decrease the emissions of pollutants and promote operational safety and economy.This paper takes a 660 MW tangential firing boiler and the downstream SCR system as objects.Based on the test,numerical simulation,and data mining methods,several problems of combustion optimization and low NO_x emission are studied to finalize the optimization strategy and method for boiler operation and flue design.This study introduces a newly coupled simulation of boiler combustion model and thermal hydraulic model.The distributed parameter models of the gas and steam sides of the boiler were first established based on the CFD and MATLAB platforms.In addition,the direct mapping method and the auxiliary cells method were proposed to match different grid systems,and an interface program was further developed to exchange the information from different models.Therefore,the tube wall temperature could be calculated by considering the coupled heat transfer on the gas and steam sides.According to this model,the positions inside the boiler at which overheating problem is more likely to occur was evaluated,and this capability is essential to properly selecting a steel alloy and deciding the location of measuring points.Moreover,the variety of air-staging effects,swirling momentum,turbulence intensity,flame center,and gas temperature deviation were analyzed at different vertical SOFA tilt angle,and further attention was paid on the combustion products and tube wall temperature.The simulation results reveal that an optimal value for the boiler efficiency and NO_x concentration will be achieved at the maximum SOFA angle of+30°;however,when SOFA angle tilts upward from+20°to+30°,0.16%of the outer surfaces of the final re-heater would exceed the allowable temperature of the material SA-213 TP347H.In this work,a novel approach is presented to predict the erosion condition of SCR catalysts in actual environment.First,a gas solid flow model was established for SCR reactor to obtain the distributions of flue gas and ash particles at the entrance of the catalyst.Then,these distributions were coupled with the developed erosion model for single catalyst channel thus achieving the erosion calculation of the catalyst layer.Based on this model,the influence of triangular beams,support steels,and turning vanes on the gas flow and particle trajectories were studied,and the erosion caused by different ash particles was discussed.Finally,a series of retrofit schemes were successfully implemented to optimize the flow and ash filed,and the local severe erosion of catalyst had been eliminated.Aiming at the difference and non-uniformity of velocity and NO_x concentration at the entrances of two SCR reactors,an ammonia injection grid?AIG?tuning method considering the combustion effect is proposed.First,the flow coefficient and concentration coefficient was calculated according to the results of boiler combustion modeling.Based on the principle that the flow rate of NO is equal to the supply rate of ammonia,the ammonia coefficients for different reactors were deduced.Then,coupled with the flow fields obtained from the combustion modeling,the SCR flow model and component transport model were solved to calculate the ammonia coefficients for each branch pipe.Finally,different ammonia control valves were adjusted to achieve the target flow rate distribution given by ammonia coefficients.As is revealed by the results,the presented method can make the relative standard deviation?RSD?of NH₃/NO molar ratio meet the design requirement,and has a good effect on the problem of non-uniform NH₃/NO molar ratio at the entrances of the catalysts.This work incorporates the monotone knowledge into the data-driven modeling of combustion system.The virtual samples were used to add derivative information as inequality constraints of support vector regression?SVR?,and additive kernel was introduced to reduce the calculation cost in the model training.Finally,a fusion monotony support vector regression?FM-SVR?was achieved.The results of the test cases show that,under the circumstance of sample incompleteness and data with large noise,FM-SVR presents a higher prediction accuracy and can better describe the monotony relationship of the function.In addition,FM-SVR was applied in the modeling of the formed NO_x concentration,the boiler efficiency,and the gas temperature deviation.The results show that,compared with LS-SVR,the incremental analysis results based on FM-SVR were more consistent with the operation and emission characteristics of the boiler.In this paper,an adaptive NO_x emission model and ammonia escape model are proposed.First,the steady state samples were acquired from the massive historical data using the R-statistic method,and the system inlet parameters were clustered according to the condensed nearest neighbor?CNN?rule.Thus,the classified steady state samples formed a database.On this basis,a sliding window approach was used to deal with the continuous data stream.As the newest steady state sample was introduced into the database,the most similar old sample in the same class was replaced.The crowding distance?CD?operator was also used to eliminate the redundant samples.Finally,the RCNN-CD method was obtained.A total of 1084 days operational data from SCR system were analyzed.As a result,RCNN-CD can improve the representatives of the samples,and the trend shown in these samples agrees with SCR mechanism.Based on the selected samples,FM-SVR was used to establish the NO_x emission model,which was applied to further predict the ammonia escape.The results show that,these models can reasonably reflect the variation of SCR DeNOx performance.This work presents an integrated optimization approach for boiler combustion and SCR combined system.Considering gas temperature deviation,the combustion optimization problem was first solved using the inequality constraint NSGA-II algorithm.Taking this result as the boundary condition and ensuring the NO_x emission within the limit,the total operation cost and ammonia escape were further optimized.The results show that,for the solutions in combustion optimal Pareto front,the total cost drops as the boiler efficiency increases.However,when the boiler efficiency rises to a certain extent,its small increase will lead to a dramatic growth of NO_x formation,and the solutions also have a relative large O₂content and gas flow rate in this region.To meet the NO_x emission standard,the ammonia flow will also increase substantially,resulting in an increasing total operation cost.This inflection point of the total operation cost also has a relative low ammonia escape,which is the optimal solution for the combined system.