Multiple linear regression analysis of some important performance parameters of a conventional diesel engine in different working conditions

Nowadays, the world is facing to increasing loss of fossil resources, energy crisis and environmental problems. On the other hand, diesel engines due to wide application in various sectors such as transport, agriculture, industry, etc., are the main sources of emissions and fuel consumption. Accurate measurement of fuel consumption and engine pollution is time-consuming and costly. Hence, the main objective of this study was to develop proper linear regression models of some important performance parameters of ITM285 tractor engine based on engine torque and engine speed. Experiments were carried out in 11 levels of primary engine speed (1063, 1204, 1346, 1488, 1629, 1771, 1818, 1913 and 2054 rpm) by 10 N.m steps of torque from zero (no load) to full load. The measured parameters include fuel consumption mass flow, exhaust temperature, instantaneous engine speed, maximum and mean exhaust opacities. Four different linear regression models were used to estimate the parameters. The results of regression models performance evaluation showed that quadratic model had the highest efficiency and the lowest RMSE for all parameters. The maximum and minimum effects of engine torque were on exhaust temperature and instantaneous engine speed, respectively; while, this result was completely reverse for primary engine speed. The results of regression models evaluation showed a high adaptation between the output of each model and the desired output. Also, the fuel mass flow and exhaust temperature were highly correlated to the maximum and mean exhaust opacity with correlation coefficients of 0.96 and 0.99, respectively.