Remaining useful life prognostic for degrading systems with age- and state-dependent jump-diffusion processes

Area of Science:

• Prognostics and Health Management (PHM)
• Reliability Engineering
• Stochastic Processes

Background:

• Remaining Useful Life (RUL) prediction is crucial for system maintenance and operational planning.
• Existing degradation models often overlook state dependence and environmental jump effects, limiting RUL accuracy.
• Accurate RUL estimation requires models that integrate age, state, and environmental factors.

Purpose of the Study:

• To propose a generalized age-state dependent jump-diffusion (ASDJD) model for enhanced RUL estimation.
• To address the limitations of existing models by incorporating age dependence, state dependence, and degradation jumps.
• To provide a more robust framework for predicting system lifetime.

Main Methods:

• Development of the age-state dependent jump-diffusion (ASDJD) model.
• Derivation of approximate analytic expressions for RUL distribution using first hitting time (FHT).
• Parameter estimation using Expectation Conditional Maximization (ECM) and Maximum Likelihood Estimation (MLE).

Main Results:

• The proposed ASDJD model effectively accounts for age dependence, state dependence, and degradation jumps.
• Validation using simulation and real-world bearing datasets confirms the model's efficacy.
• The study demonstrates the significant impact of state dependence and jumps on RUL estimation accuracy.

Conclusions:

• State dependence and environmental jumps are critical factors that must be included in RUL estimation.
• The proposed ASDJD model offers a more comprehensive approach to RUL prediction in complex systems.
• This research contributes to advancing the field of prognostics and health management through improved degradation modeling.