Multidimensional identification of resonances analysis of strongly nonstationary signals, case study: Diagnostic and condition monitoring of vehicle's suspension system

The paper presents novel methodology for multidimensional analysis of strongly nonstationary signals. This methodology contained the identification of working cycles, dedicated signal processing methods for each of identified segments of signal as time, frequency and TFR distributions of each signal segments. It allows to observe influence of increase of defects based on narrow and precise resonances windows. At the final stage of the signal processing the multidimensional estimators as representations of time, frequency and TFR measures are identified. The paper presents result of investigation conducted on real passenger car. It presents results for the signal processing of unsprung and sprung mass of suspension system separately. This approach allows to evaluate the properties of suspension and influence of the shock absorber defects on safety and comfort of the driving. The obtained results show a very good separation properties of multidimensional estimators for different kind of defects. Thus it can be considered as diagnostics method for condition monitoring of vehicle's suspension system.