Machine Learning-Based Critical Heat Flux Predictors in Subcooled and Low-Quality Flow Boiling

摘要

The subcooled and low-quality flow boiling critical heat flux (CHF) corresponding to the departure from nucleate boiling (DNB) is one of the major limiting factors in the design and operation of pressurized water reactors (PWRs). The path for an accurate and robust prediction of CHF is elusive in that no general agreement on its physics has been established. Over the past few decades, the reactor thermal-hydraulics community has proposed various predictive tools under steady-state conditions. The data-driven approaches -namely the best-fit empirical correlations and look-up tables- result in relatively good agreement with specific experimental datasets but often fail to extend beyond their ranges of validity, while the physics-based mechanistic models rely on reasonable yet limited understanding of the underlying physics and apply constitutive relations to close the conservation equations.