Optimization of Engine Torque Model Based on the Multiple Linear Regressions

摘要

The engine torque is a very important parameter in the calculation of automotive dynamical performance and economical efficiency. In this paper, the experimental data of a certain engine is taken as a basis. The model of the engine toque, which is based on the multiple linear regressions, is improved. What's more, the regression coefficient is for the first time checked, and the independent variable, which plays a faint role in the result, is eliminated. It firstly tests the degree of fitting by means of multiple correlation coefficients, which is modified by the degree of freedom, and thus it avoids the phenomenon of blind increase of the independent variable, which is the result of the pursuit of the fitting effect. The relative error is summed up for the sake of making the model more practical. The final improved regression project has a noticeable effect of regression, a higher degree of fitting and less calculation, which is more suitable to the calculation, simulation and test analysis of the engine as well as other capabilities of the entire vehicle. The mat lab calculation procedure is universal.