Optimization On Emission Performance Of Non-road Single-Cylinder Diesel Engine Based On Surrogate Model

The problem of pollutant emission from non-road diesel engines has become increasingly prominent,especially for low-power single-cylinder diesel engines limited by structure and cost.With the increasing market demand and the emission limits’ becoming stricter,the upgrading of key technology and renewal for non-road diesel engines have reached an urgent stage.In this paper,the 186 FA non-road single cylinder diesel engine is taken as the research object.the influence of main parameters of intake and combustion system on engine power and emission performance is studied by combination of one-dimensional and three-dimensional numerical simulation.On this basis,a multi-objective optimization method based on surrogate model is introduced to reduce NOx and Soot emissions of diesel engine by optimizing the match of key factors including oil,gas and chamber.The main research work and conclusions are as follows:(1)One-dimensional and three-dimensional simulation models of in-cylinder working process of 186 FA diesel engine were established by GT-POWER and CONVERGE software.The grid independence of the three-dimensional model was tested.Finally,the reliability of the one-dimensional and three-dimensional simulation models was verified by bench tests.The initial and boundary conditions of the three-dimensional model are derived from the simulation results of the one-dimensional model.(2)Based on the three-dimensional numerical simulation model of engine working process,the effects of four parameters,including intake swirl,injection angle,injection timing and injection duration,on engine power and emission performance were studied,and the influence rules of each parameter on engine performance were obtained.(3)Taking the swirl ratio,injection angle,injection timing and injection duration as design variables,the engine average indicated pressure,NOx and Soot emission index as response variables,the sample library was constructed by experimental design and three-dimensional numerical simulation of engine combustion process.A BP neural network(GA-BP)proxy model based on genetic algorithm optimization was established based on MATLAB software.Sample test results show that the model has strong robustness and generalization ability,and can approximately replace the three-dimensional simulation model for further research.(4)Coupled the second generation non-dominated sequencing genetic algorithm(NSGA-II)with the above surrogate model to form a complete multi-objective optimization model,the dual-objective optimization of diesel engine Soot and NOx emissions is realized,and the Pareto optimal solution set is obtained,which is the optimal matching scheme set of four parameters: intake swirl ratio,injection angle,injection timing and injection duration.Finally,three typical four-parameter matching schemes,Soot emission optimization,NOx emission optimization and Soot-NOx weighted emission optimization,are selected in Pareto optimal solution set.The reliability of multi-objective optimization results is verified by three-dimensional numerical simulation of engine combustion process.The results show that the three matching schemes can effectively reduce Soot and NOx emissions.(5)In order to achieve the optimal matching of the above parameters of the intake and injection system with the combustor,the structure of the original engine combustor was improved,including reducing the diameter-depth ratio and increasing the pit diameter depth.Then,based on the above multi-objective optimization method,reasonable parameters of the intake and injection system were matched for the improved combustor to realize the overall optimization of oil,gas and chamber thus reducing emissions.The results show that for different combustors,the reasonable range of intake and injection parameters is different for different combustors,and the emission performance of the improved combustor is better after the optimization of parameters of intake and injection system.Taking the weight optimization schemes of Soot and NOx emissions as examples,the emission performance of Soot and NOx is 27.9% and 27.2% higher than that of the original combustor respectively.