Data consolidation, correlations assessment, and new correlation development for pool boiling critical heat flux specific to cryogenic fluids

Abstract

A common and crucial process for cryogenic fluid management is the need to fill the cryogenic fluids into warmer storage tanks efficiently while minimizing the evaporative and entrainment losses. However, no good predictive tool that can estimate the heat transfer behavior accurately during this fill process is available in the literature. In this study, we start with developing a new generalized correlation set for predicting critical heat flux (CHF), an important parameter on the boiling curve. First, a new pool boiling test facility is built to generate CHF data using a flat plate geometry. Based on our assessment of prior pool boiling data and correlations, 30 CHF datapoints in three different orientations with respect to gravity are obtained and validated against the existing CHF data from published literatures. Second, our experimental data is then combined with worldwide data for cryogenic fluids to generate a consolidate database that includes 2205 datapoints for 6 cryogenic fluids for 3 boiling surfaces (flat plate, cylinder and sphere). Third, a comprehensive assessment of prior pool boiling CHF correlations is conducted and it is observed that no one correlation or correlation form can capture all important parametric effects important to CHF. Fourth, a new correlation model for flat plate is setup that gave a mean absolute error (MAE) of 15.96% based on 1259 datapoints. In addition, multipliers to the flat plate correlation form are used to predict CHF for curved surfaces and gave MAE of 22.38% based on 355 datapoints and MAE of 17.16% based on 177 datapoints for cylinders and spheres, respectively. Values of CHF based on quenching is observed to be smaller than CHF based on boiling for the same test conditions regardless of the boiling surface material. Finally, a parameter to account for the thickness & thermal properties of the quenching surface is generated and added to the original correlation form and gave an MAE of 14.05% based on 245 datapoints.