Characterisation of the Dynamics of an Automotive Suspension System for On-line Condition Monitoring

As the most critical system that determines the driving performance, passenger comfort and road safety of a vehicle, the suspension system has not been found to have adequate monitoring systems available to provide early warnings of possible faults online. To fill this gap, this study has focused on the investigation of the dynamic behaviour of the suspensions upon which a new on-line condition suspension monitoring approach was proposed and verified under different conditions. Specifically, the approach quantifies the modal shapes which are obtained based on an improved modal identification applying to acceleration responses at the four corners of the vehicle. To achieve this, the research was carried out by the means of dynamic modelling, numerical simulations, optimal measurement optimisations and subspace identification improvements based on a representative vehicle system, cost-effective measurement techniques and road standards.
Firstly, a mathematical model with a seven degree-of-freedom (7-DOF) was developed in account of variable stiffness and damping coefficients, being applicable for computer simulation of the dynamic interaction between a vehicle and a road profile. To validate the proposed model during real operation, this study investigates a set of on-road experiments, to measure the acceleration of the vehicle body. Comparisons between the experimental and simulation paths demonstrated that, simulation results and measured on road results were found to be almost have similar trend.
In the simulations the modal parameters (obtained theoretically) of a vehicle are: natural frequency, damping ratio and modal shapes and their characteristics are characterised under the influence of different suspension faults and operating conditions (loads and speed). It has found that the modal shapes are more independent of operating conditions and thereby reliable as indicators of faulty suspensions, compared with modal frequency and damping which are influenced more by operating conditions. Furthermore, the modal shape difference between left and right side responses are developed as the fault severity indicator.
To obtain the modal shapes online reliably, an improved stochastic subspace identification (SSI) is developed based on an average correlation SSI. Particularly the implementation of optimal reference channels is achieved by comparing the average correlation signals which can be more efficient due to much smaller data sizes, compared with that raw data based spectrum analysis method used in original development.
On road verification based on a commercial vehicle operating in normal road conditions shows that common suspension faults including inadequate damping faults and under-inflation of the tyre, induced one of the four shock absorbers, can be detected and diagnosed with acceptable accuracy. Therefore, it can be deduced that the SSI modal shape based detection techniques are effective and therefore promising to be used to diagnose and monitor the suspension system
online.