Combining survival-analysis and data-driven models for predicting remaining useful life

  • Thursday / 12 March / 2026
  • 15:30-15:50

The objective of the approaches for predictive maintenance typically follow one of two paradigms: survival analysis, which considers cumulative wear and usage to capture long-term aging, and data-driven RUL estimation, which relies on time-series sensor measurements to assess the current health state and predict near-future degradation. In this work we show how these two approaches for predicting RUL can be combined.

Speaker
Sepideh Pashami Senior forskare RISE
Room
Teknikscenen