Investigation of the Relationship between ...

Selecting the fuel properties as the independent variables, and the emissions from engines as the dependent variables, the fuzzy linear regression equations are constructed in this paper with the fuzzy linear regression; as well as the simulation and prediction results calculated with the equations are analyzed. Satisfactory results are achieved by selecting an appropriate degree of membership. The effect of the degree of the membership on the fuzzy regression coefficients is also analyzed. There was no effect on the fuzzy centers of the fuzzy regression coefficients of the equations for total hydrocarbons (HC), carbon monoxide (CO), nitrogen oxides (NOx). There was a little effect on those of the equation for particulates (PM). The fuzzy widths of the fuzzy regression coefficients increase when the degree of the membership increases. When the biases of the center values of the simulation results to the actual values are big or the intervals of the simulation results are too big to have meaning, additional restrictions should be used to improve the simulation accuracy of the equations.