A5016477281- Risk and Safety Analysis
- Reliability and Maintenance Optimization
- Infrastructure Resilience and Vulnerability Analysis
- Software Reliability and Analysis Research
- Probabilistic and Robust Engineering Design
- Smart Grid Security and Resilience
- Software-Defined Networks and 5G
- Fault Detection and Control Systems
- Power System Reliability and Maintenance
- Machine Fault Diagnosis Techniques
- Supply Chain Resilience and Risk Management
- Statistical Distribution Estimation and Applications
- Software System Performance and Reliability
- Occupational Health and Safety Research
- Global Energy Security and Policy
- Multi-Criteria Decision Making
- Optimal Power Flow Distribution
- Technology and Data Analysis
- Quality and Safety in Healthcare
- VLSI and Analog Circuit Testing
- Network Security and Intrusion Detection
- Risk Perception and Management
- Gear and Bearing Dynamics Analysis
- Evaluation and Optimization Models
- Hydrocarbon exploration and reservoir analysis
CentraleSupélec
2016-2025
Université Paris-Saclay
2016-2025
Laboratoire Génie Industriel
2016-2024
China University of Geosciences
2023-2024
Électricité de France (France)
2016-2023
Bouygues (France)
2016-2023
Supélec
2015-2022
Orange (France)
2022
Beihang University
2012-2015
The increasing availability of condition-monitoring data for components/systems has incentivized the development data-driven Remaining Useful Life (RUL) prognostics in past years. However, most studies focus on point RUL prognostics, with limited insights into uncertainty associated these estimates. This limits applicability such to maintenance planning, which is per definition a stochastic problem. In this paper, we therefore develop probabilistic using Convolutional Neural Networks. These...
In this paper, a systematic review of non-probabilistic reliability metrics is conducted to assist the selection appropriate model influence epistemic uncertainty. Five frequently used are critically reviewed, i.e., evidence-theory-based metrics, interval-analysis-based fuzzy-interval-analysis-based possibility-theory-based (posbist reliability) and uncertainty-theory-based (belief reliability). It pointed out that qualified metric able consider effect uncertainty needs (1) compensate...
Traditional quantitative risk assessment methods (e.g., event tree analysis) are static in nature, i.e., the indexes assessed before operation, which prevents capturing time-dependent variations as components and systems operate, age, fail, repaired changed. To address this issue, we develop a dynamic (DRA) method that allows online estimation of using data collected during operation. Two types considered: statistical failure data, refer to counts accidents or near misses from similar...
In this article, we develop a mixture of Gaussians-evidential hidden Markov model (MoG-EHMM) to fuse expert knowledge and condition monitoring information for remaining useful life (RUL) prediction under the belief function theory framework. The evidential expectation-maximization algorithm is implemented in offline phase train MoG-EHMM based on historical data. online phase, trained used recursively update health state reliability particular individual system. predicted RUL is, then,...
Selective maintenance is an important condition-based strategy for multi-component systems, where optimal actions are identified to maximize the success likelihood of subsequent missions. Most existing works on selective assumed that after each mission, components' states can be precisely known without additional efforts. In engineering scenarios, components in a system need revealed by inspections usually inaccurate. Inspection activities also consume limited resources shared with...
We consider reliability engineering in modern civil aviation industry, and the related activities methods. a broad sense, referring to other system characteristics that are it, like availability, maintainability, safety durability. covered entire lifecycle of equipment, including requirement identification, analysis design, verification validation requirements (typically involved equipment design development phase), quality assurance (which typically enters manufacturing fault diagnosis...
Model-based reliability analysis and assessment methods rely on models, which are assumed to be precise, predict reliability. In practice, however, the precision of model cannot guaranteed due presence epistemic uncertainty. this paper, a new metric, called belief reliability, is defined explicitly account for uncertainty in model-based assessment. A method developed quantify by measuring effectiveness engineering activities related To evaluate an integrated framework presented where...
Existing Physics-of-Failure-based (PoF-based) system reliability prediction methods are grounded on the independence assumption, which overlooks dependency among components. In this paper, a new type of dependency, referred to as failure collaboration, is introduced and considered in predictions. A PoF-based model developed describe behavior systems subject collaboration. Based model, Bisection-based Reliability Analysis Method (BRAM) exploited calculate reliability. The applied predicting...
A sequential Bayesian approach is presented for remaining useful life (RUL) prediction of dependent competing failure processes (DCFP). The DCFP considered comprises soft due to degradation and hard random shocks, where dependency arises the abrupt changes brought by shocks. In practice, shock are often unobservable, which makes it difficult accurately estimate intensities predict RUL. proposed method, problem solved recursively in a two-stage framework: first stage, parameters related...