Under variable operating conditions damaged bearings generate nonstationary vibrations. They are characterized by repetitive impacts between the rolling elements and the damaged surface. This paper presents an approach for bearing fault detection based on the theory of stochastic point processes. Within the concepts of point processes, bearing vibrations can be described as a point process with inverse Gaussian interevent distribution. The effectiveness of the approach is shown on a data-set acquired from a two–stage gearbox with various bearing faults operating under different rotational speeds and loads.
COBISS.SI-ID: 25258023
The paper presents a novel approach for prognostics of faults in mechanical drives under non-stationary operating conditions. The feature time series is modelled as the output of a dynamical state-space model. In this model the variable operating conditions are treated as the known model inputs. An algorithm for on-line model estimation is adopted to find the optimal model at the current state of failure. This model is then used to determine the presence of the fault and predict the future behaviour and remaining useful life of the system. The approach is validated using the experimental data on a single stage gearbox.
COBISS.SI-ID: 25257767