Multi-objective Optimization Of Diesel Engine SCR System For Virtual Calibration

Environmental pollution and market competition are not only the two major ob-stacles to the development of diesel engine and reprocessing technology,but also the two major boosters of its rapid development.The increasingly strict emission regula-tions on pollutant limits promote the rapid development of the key technologies of die-sel engine to reduce NO_x emissions.With the development of diesel engine,the market demand for diesel engine function increases,which leads to the complexity of the con-trol system,and the control parameters increase exponentially.In addition,the long calibration cycle of diesel engine and the cost of manpower and material resources lead to the economic burden of the early stage of research and development.Therefore,in order to reduce the burden of diesel engine research and development in the early stage,it has become an important research direction to seek an effective method to shorten the diesel engine calibration period and save R&D cost.Virtual calibration can greatly shorten the calibration cycle and reduce the con-sumption of manpower and material resources in calibration work to a certain extent.In the study of virtual calibration of diesel engine abroad,the hardware-in-the-loop simulation method has been used for a long time,and the domestic similar research is mostly based on foreign hardware-in-the-loop simulation equipment.Compared with the relatively mature hardware-in-the-loop simulation technology abroad,model-based simulation research is an easier virtual calibration method.In recent years,diesel engine emission problem has attracted much attention,so it is more urgent to apply virtual calibration technology to diesel engine reprocessing system calibration research.selective catalyst reduction（SCR）technology is the key technology to achieve NO_x purification in diesel engine reprocessing system.The NO_x in diesel exhaust is reduced to N₂ and H₂O by appropriate urea injection.In this process,a series of complex catalytic reduction reactions and the development of control strategies for urea injection volume need to be involved.In the early stage of the development of control strategies,bench tests need to be carried out to obtain key control MAP.Aiming at the problem that the key control in the control strategy is not easy to obtain,a MAP acquisition method combining the calibration of MAP based on virtual calibration method and the optimization of the calibration MAP by using multi-objective genetic algorithm is pro-posed.The main research contents are as follows:（1）Data collection based on bench testBy using space filling test design,a suitable test site was selected for bench test,and the data of the diesel engine’s original and tail row required by SCR system model were collected,and the original row data were briefly analyzed.（2）The SCR system model is establishedThe catalytic reaction mechanism of SCR system is studied,and a simplified SCR system model which can basically meet the research needs is established according to the research situation and data volume.The key parameters of SCR system model are identified by nonlinear least square method.（3）Urea injection control strategy for virtual calibration was designedUrea injection control strategy of SCR system was studied,and urea injection con-trol strategy model oriented to virtual calibration was designed according to the research requirements,including:spray starting temperature judgment model,calculation model of urea basic injection volume,steady-state correction model,ammonia leakage correc-tion model,ammonia storage correction model,etc.The strategy verification shows that the control strategy can basically meet the research requirements.（4）Propose a virtual calibration method for key control MAPBased on the co-simulation of SCR system model and urea injection control strat-egy model,the calibration method of SCR system was studied,and a virtual calibration method was proposed to study the acquisition of key control MAP in the control strategy.The target urea injection amount,target NO_x conversion rate,target ammonia leakage and target ammonia coverage rate were calibrated.（5）A multi-objective genetic algorithm is proposed to optimize the calibration MAPMulti-objective genetic algorithm was used to optimize the calibration parameters,and the optimal solution of NO_x conversion rate and NH₃ leakage was obtained by solv-ing the multi-objective optimization problem.The optimized MAP was applied to the urea injection volume control strategy and WHTC simulation verification was carried out.The simulation results showed that the calculated results of the specific NO_xemission of the SCR system model had little difference with the calculated results of the experimental value NO_x specific emission;the simulated value of the tail discharge NO_x was basically consistent with the experimental value,and the average ammonia leakage was less than 10ppm.