Research On Multi-objective Optimization Method Of Marine Diesel Engine Combustion Performance Based On Identification Model

The marine diesel engine is one of the most important power equipment in Marine traffic,carrying 90% of the global trade traffic.However,with the growing awareness of energy conservation and environmental protection,as well as the continuous improvement of emission standards,optimizing the combustion process of diesel engines and reducing their emissions have become important research areas.Traditional combustion optimization methods often require a significant amount of time and resources,but with the development of computer,combustion prediction model based on intelligent algorithm and multi-objective optimization method can effectively solve these problems.In addition,the rotational speed stability of Marine diesel engine is also an important aspect,but few studies have considered how to optimize combustion while keeping the speed of diesel engine stable.Therefore,this paper proposes a novel double closed-loop control scheme to maintain diesel engine speed stability while optimizing combustion.Firstly,a simulation model of the marine diesel engine was established and calibrated using the GT-Power software.The combustion process analysis was conducted to select fuel injection timing,injection pressure,cyclic fuel injection quantity,and intake valve opening timing as control parameters for combustion optimization.The correlation between these parameters and combustion performance parameters was analyzed.Then,in order to achieve fast iteration for multi-objective combustion optimization,a data-driven model for diesel engine combustion performance at full load was developed.Secondly,the Takagi-Sugeno(T-S)fuzzy modeling principle was combined with neural networks to form a global model by combining local neural network models of four typical operating conditions with Gaussian functions marked by speed.Then,a double closed-loop control loop was designed and implemented,with the outer loop being the speed stabilization loop.The speed control model was established in SIMULINK,and the fuzzy PID controller was used to adjust the cyclic fuel injection quantity to achieve rapid speed stability.Selected the power as the inner and outer ring connection,and the power was kept constant as long as the rotating speed remained stable.The inner loop was the combustion optimization loop,as the fuel consumption rate,NOx and power error were optimized by intelligent multi-objective optimization algorithm.After that,TOPSIS method was used to make decisions on the Pareto solution set obtained by NSGA-Ⅲ non-dominant sorting genetic algorithm,selecting the optimal solution.Finally,the optimal solutions obtained under different working conditions were substituted into the established verification model to verify the optimization effect.Through the verification,the accuracy of the diesel engine combustion performance data drive model for combustion optimization could reach 99.84%.The maximum error of the verification data was less than 5%,the minimum was 0.16%,with good generalization ability and global adaptability.When it was applied to the iterative process of multi-objective optimization algorithm,the running speed was 95.5% faster than the GT-Power simulation model,which could save 24851 seconds.After optimization,the fuel consumption rate of each operating condition decreased by an average of 3.647 g/k Wh compared with the original engine,and the average NOx was reduced by 9.38%.The results showed that the double closed-loop control loop can optimize combustion and keep the speed stable at the same time.