Multi-objective optimization of a dual-fuel engine at low and medium loads based on MOEA/D

The diesel/natural gas dual-fuel engine has the problem that it is difficult to compromise and coordinate multiple goals at medium and low loads. The objective of this research is to find the optimal combination of control parameters for a dual-fuel engine that will reduce emissions while reducing fuel consumption. A regression model for predicting emissions and fuel consumption of dual-fuel engines was constructed using BP neural network, and it was verified that the R-2 of the output parameters were all greater than 0.9. The model has the good predictive ability. Then use a multi-objective evolutionary algorithm based on decomposition (MOEA/D) to perform multi-objectiveoptimization on the dual-fuel engine. The algorithm uses an adaptive differential evolution operator to generate offspring solutions. Research shows that MOEA/D can be used for medium and low loads optimization of dual-fuel engines, and the obtained non-dominated solutions are evenly distributed on the Pareto front. Selecting the low-emission optimal solution on the Pareto front, the optimized NOx emissions decreased by 84%, and the fuel consumption decreased by 3.8%.