Search Results

A finite element model of the pushrod was derived and integrated into a lumped parameter valve train model. The pushrod model included the nonlinear effects of material dissipation, axial inertia, and coupling of the transverse and axial vibration. Experimental valve motion and pushrod vibration data were obtained, from a valve train test fixture, using proximeters and a strain gauged pushrod. The experimental results were used to identify model parameters and validate the model. Once the effectiveness of the valve train model had been determined it was used to investigate the effect of pushrod bending on the valve train. The valve train model proved to be very effective in predicting valve train response at high speeds. In addition to this, the pushrod model itself was found to accurately simulate the actual response of the pushrod. However, no distinct relation was found between pushrod vibration and valve response for the system studied.

An experimental method was developed to dynamically measure rigid-body camshaft torsional vibrations in a high speed internal combustion engine. A mathematical model of the system was also created which required an accurate dynamic model of the entire valvetrain to predict the forcing function. Techniques for limiting the vibration were also implemented and tested. While attempting to correlate the experimental and theoretical results, it was discovered that the rubber timing belt exhibits viscoelastic behavior. The system stiffness and damping coefficients are frequency dependent. A system identification was required to accurately determine these parameters. No distinct relation between camshaft torsional vibrations and valve response was determined for the system studied. However, significant vibration reduction was achieved with a camshaft damper and an inertia addition to the camshaft.

A numerical modeling technique is proposed for computer simulations of high speed valve train dynamics. The dynamic terms in the valve spring reaction forces are calculated using linear vibration theory for given kinematic valve motions. Because the spring dynamics are analyzed before the time stepping integration, spring surge phenomena can be included without using additional computer time. Consequently, valve train dynamics can be simulated very quickly without noticeable errors in accuracy. The experimental results prove the computer model developed here is accurate and also computationally efficient.