Hybrid Degradation Equipment Remaining Useful Life Prediction Oriented Parallel Simulation considering Model Soft...

Area of Science:

• * Engineering
• * Reliability Engineering
• * Predictive Maintenance

Background:

• * Equipment degradation often involves a mix of continuous and discrete shock processes.
• * Existing parallel simulation methods primarily address single continuous degradation modes.
• * Accurate remaining useful life (RUL) prediction requires modeling hybrid degradation patterns.

Purpose of the Study:

• * To develop a hybrid degradation parallel simulation approach for equipment RUL prediction.
• * To enable adaptive model selection and parameter evolution based on real-time data.
• * To enhance RUL prediction accuracy and reduce uncertainty in complex degradation scenarios.

Main Methods:

• * State space model (SSM) framework incorporating Wiener process and Poisson effect.
• * Forward interactive multiple model (FIMM) filtering for model soft switching and data assimilation.
• * Expectation maximization (EM) algorithm for model parameter evolution.
• * Integration of inverse Gaussian distribution for RUL probability density function (PDF).

Main Results:

• * The proposed method effectively models hybrid degradation processes, demonstrated with bearing data.
• * Adaptive model soft switching improves simulation fidelity by continuously approaching actual equipment states.
• * The hybrid approach yields higher RUL prediction accuracy and lower uncertainty compared to single-model methods.

Conclusions:

• * Parallel simulation with adaptive model soft switching is crucial for hybrid degradation RUL prediction.
• * The developed method offers a robust framework for real-time RUL estimation under mixed degradation.
• * This approach enhances the reliability and predictive maintenance capabilities for industrial equipment.