Gas leakage fault recognition and prognostics of special vehicle hydro-pneumatic spring

Aiming at the demands of condition based maintenance and self-maintenance for special vehicle, the condition recognition and prognostics technology based on data driven method were firstly studied for the hydro-pneumatic spring that is the core element of special vehicle suspension system. The main failure mode and mechanism of the hydro-pneumatic spring were analyzed. Aiming at the gas leakage fault occurring most frequently, a new feature extraction method based on the gas pressure change under the same displacement suitable for different working conditions was proposed. A fault recognition and prognostics architecture based on SVM and SVR was proposed. This architecture is suitable for the real vehicle application environment, can achieve the real time monitoring and preventive maintenance of the gas leakage fault of the hydro-pneumatic spring only using the existing test environment of the vehicle without adding extra transducers. Compared with other methods based on vibration signals, the proposed method has stronger practicality. State degradation simulation test was conducted in the laboratory. The data collected in the test verify the feasibility of the proposed method, which can be used in the online condition monitoring of the hydro-pneumatic suspension in special vehicle.