Rubber mounts are used to mount the engine/power train to the chassis in most automobiles. The primary function of these mounts is to isolate the passenger from the engine induced vibrations transmitted to the frame as well as to isolate the engine from road undulations. The design of these mounts is usually a trade-off between performance and durability. Rubber mounts are viscoelastic which is helpful in vibration attenuation. It is challenging to characterize the rubber mounts because their dynamic properties (stiffness and damping) depend on static preload, excitation amplitude, frequency, temperature and orientation angle. Most of the parameters vary under operating conditions. Thus a mathematical model is required to predict the response of the mount for effective design. Amplitude and frequency of excitation are the most common parameters that vary under operating conditions. The objective of this thesis is to develop a procedure to identify the frequency and amplitude-dependent dynamic properties (stiffness and damping) of the engine rubber mount in both shear and compression.

To carry out system identification, a single-degree-of-freedom system was fabricated. A rubber mount provided compliance and damping while a cylindrical mass of 4.6 kg attached to the mount using mount bolts supplied the inertial mass. A harmonic base excitation was provided to the system in compression and shear directions using an electrodynamic shaker. Sine sweep test was carried out for various amplitudes of excitation. Apart from amplitude, and frequency-dependent stiffness and damping, experimental FRF indicates that the mount stiffness is non-linear (cubic softening type), as sudden jumps in response amplitudes were observed during sweep up and sweep down of excitation frequency.

To identify compressive stiffness and damping, the Kelvin-Voight model with cubic stiffness nonlinearity was used. To identify shear stiffness and damping, in addition to cubic stiffness nonlinearity, shear stiffness was assumed to be quadratic in frequency. Parameters were estimated for various amplitudes of excitation from experimental FRF using the harmonic balance method. The model response generated from identified parameters using harmonic balance and Newton Raphson method agree well with experimental FRFs.