Probabilistic RUL Prediction Based on Nonlinear Wiener Process and Monte Carlo Simulation

Remaining useful life (RUL) prediction is the core of intelligent operation and maintenance, improving operation and maintenance efficiency and reducing maintenance costs. Aiming at the limitations of RUL prediction methods based on mathematical derivation in terms of adaptability and flexibility, this paper proposes a probabilistic RUL simulation method for complex systems or complex degradation processes. The method constructs a degradation model using a complex nonlinear Wiener process, estimates the model parameters using maximum likelihood estimation, and updates the stochastic parameters based on the Bayesian updating strategy. On this basis, the prediction of probabilistic RUL is realized by the simulation algorithm. Compared with the mathematical derivation method, the prediction of the method proposed in this paper is accurate, which verifies the effectiveness and practicality of the simulation method. In a complex system or complex degradation process, the method in this paper can construct the probabilistic RUL prediction model faster.