Genetic algorithm optimization applied to the fuel supply parameters of diesel engines working at plateau

In order to enhance the power performance for the diesel engines working at plateau, the method to adjust fuel injecting parameters had been adopted. However, the diesel engine is considered as a complicated nonlinear multiple-inputs and multi-boundary system. Hence, it is difficult to find out the appropriate value for fuel injecting parameters for all conditions, this is the reason why we study the Genetic Algorithm method for optimization. Firstly, the numerical model of a turbocharged diesel engine with the predictable combustion model was established and then verified by experimental data. Base on the engine model, the relation between injecting parameters and performance was studied. Secondly, the optimization model is constructed, including the objective and the boundary conditions with a novel parameter introduced, measuring the surge margin of the operating points. Then, the Fitness function is proposed employing penalty functions to express constraints. Based on the impact of injecting parameters on constraint conditions, the method was put forward about how to choose the penalty parameter values, named "Fitness Equal to Zero at the Worst Point". In order to explain this method, 4500 m rated operation point was illustrated and four schemes with different plenty values were compared. After the comparison of the population distributions and the optimizing processes, the Scheme II is proofed to be accurate and efficient, which adopted the plenty value chosen method (Fitness (w) = 0). Finally, this GA model was used for the fuel supply parameters optimization of full-load operation at 4500 m altitude. The result demonstrates that the rated engine power is enhanced by 22.7% and the fuel consumption reduces by 6.4%.