Nonlinear Wiener Process Remaining Life Prediction Based on Adaptive Drift Parameters

Existing strategies for forecasting the remaining lifespan of deteriorating equipment are not suitable for situations where the measurement intervals for the deteriorating equipment are unevenly distributed, and the monitoring data's measurement frequency does not align with historical data. Moreover, they have not taken into account the fluctuation of adaptive drift parameters during the period of decline. In light of this, a nonlinear prediction model for the remaining lifespan of deteriorating equipment that incorporates historical observational data, based on adaptive drift parameters, is proposed. Initially, the degradation process is described using a real-time state-space model with drift parameters adjusted using historical data to establish a nonlinear degradation model following the Wiener process. Subsequently, an analytical expression for the remaining life in terms of first-passage time is derived, and the parameters of the model are determined through the expectation maximization algorithm. Finally, the proposed method's efficacy and applicability are confirmed using degradation data from a reset spring under impact.