This paper involves discovering effective and better reaction of the diesel engine at various velocities by having ideal values in a short period. Therefore, gene expression programming is used for modeling and presenting governing expression for the related factors. The effective parameters consist of engine speed, intake air temperature, rate of air over fuel, fuel mass, NOx emission, mechanical efficiency, and immediate infusion diesel engine used as a part of demonstrating. Gene expression programming and its values exactly predict output results and present precise formula. Moreover, the sensitivity analysis was performed to recognize the effectiveness of the factors for reducing NOx and increasing mechanical efficiency. In the sensitivity analysis, the methods such as partial correlation coefficient, standard regression coefficient, and the Sobol'-Jansen and distributed evaluation of local sensitivity analysis are all used. The obtained results show that air/fuel rate is more influential factor in both NOx emission and mechanical efficiency. Moreover, the difference between results of standard regression or partial correlation coefficients and Sobol'-Jansen or distributed evaluation methods is in nonlinearity effect or interactions among the factors.
Keywords: DELSA method; Diesel engine emission; Gene expression programming; Sensitivity analysis; Sobol’-Jansen method.