Abstract
In this work, the objective is to quantify the uncertainty in crack-growth propagation with the sparse available fatigue crack growth data of a Carbon-Steel Nuclear piping. Using the Bayesian Model Updating framework, we perform a model update on the established Paris-Erdogan Crack-growth rate model with the available data and compared the results of the model updating with the uncertain bounds determined using an Interval Predictor Model (IPM). In doing so, this allows for the provision of a "Reliability Certification" on the resulting probabilistic model updating which illustrates how likely the next data would fall within the stipulated bounds.
Original language | English |
---|---|
Number of pages | 1 |
Publication status | Published - 4 Nov 2020 |
Event | Modelling in Nuclear Science and Engineering Seminar 2020 - Virtual, Bangor, United Kingdom Duration: 4 Nov 2020 → 5 Nov 2020 https://www.nuclearinst.com/Events-list/Nuclear-Modelling-Conference-2020/72300 |
Conference
Conference | Modelling in Nuclear Science and Engineering Seminar 2020 |
---|---|
Country/Territory | United Kingdom |
City | Bangor |
Period | 4/11/20 → 5/11/20 |
Internet address |
Keywords
- Bayesian regression
- sparse fatigue crack growth data
- nuclear piping
- crack-growth propagation
- Bayesian model updating