Assessment of Machine Learning-Based Critical Heat Flux Models in the CTF Subchannel Code for Square Rod Bundle Prediction
Tube-trained machine-learning critical heat flux models inside CTF beat classic correlations on square rod-bundle data.
The paper evaluates machine-learning critical heat flux models inside the CTF subchannel code using the EPRI rod-bundle database. Both pure models and hybrid residual corrections are tested in local and semilocal forms. Tube-trained models generally transfer to rod bundles and outperform the Bowring and W-3 correlations and the 2006 Groeneveld lookup table. A local hybrid lookup-table model is strongest overall, with a semilocal pure ML model close behind.
- ML CHF models trained only on tubes transfer to square rod bundles.
- Local hybrid lookup-table model gives the best overall CTF results.
- Predictions beat Bowring, W-3, and the 2006 Groeneveld lookup table.
- Pure and hybrid residual models are tested in local and semilocal forms.
- Products
- CTF
- Organizations
- Electric Power Research Institute
From the original article
1 min read at arxiv.orgThe prediction of critical heat flux (CHF), a key safety-related quantity in nuclear thermal hydraulics, remains an important challenge due to its direct relationship with fuel performance and reactor safety. Recent studies have demonstrated that relative to traditional empirical correlations and lookup tables (LUTs), machine learning (ML) methods can substantially improve CHF prediction accuracy.
By Aidan Furlong · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.