A novel approach for bearing remaining useful life estimation under neither failure nor suspension histories condition

摘要: Remaining useful life prediction methods are extensively researched based on failure or suspension histories. However, for some applications, histories hard to obtain due high reliability requirement expensive experiment cost. In addition, systems' work condition cannot be simulated. According current research, remaining without is challenging. To solve this problem, an individual-based inference method developed using recorded monitoring data date. Features extracted from divided by adaptive time windows. The window size adjusted according increasing rate. in two adjacent selected windows regarded as the inputs and outputs train artificial neural network. Multi-step ahead rolling employed, predicted features post-processed next iteration. Rolling stopped until a value exceeds threshold. proposed validated simulation bearing PHM-2012 Competition data. Results demonstrate that promising intelligent prognostics approach.