EFFECTS OF FILTERING THE ANGULAR MOTION OF THE CRANKSHAFT ON THE ESTIMATION OF THE INSTANTANEOUS ENGINE FRICTION TORQUE

The focus of this study is to investigate the effects of filtering the actual angular displacement, velocity and acceleration of the crankshaft on the computation of the instantaneous engine friction torque. These effects are isolated from those of measurement errors and/or noise by relying on a detailed model of the crank-slider mechanism to generate the rigid and flexible motions of the piston/connecting-rod/crankshaft mechanism along with the engine friction torque. The (P−ω) method is used herein to estimate the instantaneous engine friction torque based on the actual and the filtered angular displacement, velocity and acceleration of the crankshaft. The digital simulation results have demonstrated that the (P−ω) method cannot produce an acceptable estimation of the instantaneous engine friction torque in spite of filtering the actual angular motion of the crankshaft. It should be mentioned that the low-pass filter is commonly implemented to attenuate the measurement noise and the effects of structural deformations on the measured angular velocity of the crankshaft. However, the ineffectiveness of the low-pass filter stems from the non-linearities of the crank-slider mechanism that induced superharmonic and combination resonance frequencies in the angular displacement, velocity and acceleration of the crankshaft. The filter has severely attenuated some of the superharmonic resonance frequencies, which constitute an important part of the rigid-body behavior of the crankshaft that is needed by the (P−ω) method to accurately predict the engine friction torque. Moreover, the filtered signals would still be contaminated by the combination resonance frequencies that may appear in the low-frequency range commonly assumed to be dominated by the frequency components of the rigid-body motion of the crankshaft.