Residual useful life prediction of gearbox based on particle filtering parameter estimation method

To solve the problem of predicting equipment residual useful life (RUL) which is non-linear and non-Gaussian, a particle filtering framework for system's RUL prediction was proposed. The framework uses a non-linear state-space model of the system (with unknown time-varying parameters) and a particle filtering (PF) algorithm to estimate the probability density function (PDF) of the state. The state PDF estimate was then used to predict the evolution of the fault indicator and ad a result obtain the PDF of the remaining useful life (RUL) for the faulty subsystem. The approach provides informations about the effectiveness and accuracy of the predictions, RUL expectations, and 95% confidence intervals for the condition under study. Data from a full life test for a gearbox were used to validate the proposed methodology, and comparisons were made between proportional hazard model (PHM) and PF method. The outcome shows that the PF method has a better effect than PHM on RUL prediction.