Parametric identification and fault detection in vehicle models with nonlinear suspension

Abstract

Systematic methodologies are applied for performing parametric identification and fault detection in nonlinear vehicle systems. The dynamic response of the vehicle models examined is caused by road excitation. In these models, the nonlinearities arise due to the function of the suspension dampers, which assume a different damping coefficient in tension and in compression. This leads to oscillator models with parameter discontinuities. First, emphasis is put on investigating some issues of fundamental importance, by employing a classical two degree of freedom quarter-car model. Since the dominant nonlinearities are due to switches between constant damping coefficients, appropriate methodologies are applied for obtaining exact motions of these models. Moreover, a statistically based methodology is applied for parametric identification and fault detection in vehicle suspensions with nonlinear characteristics, using dynamic test data. This methodology is then extended and applied to more involved and complete vehicle models. The numerical results presented demonstrate that this methodology provides an effective and practical way of detecting the type, location and severity of faults in vehicle suspensions.