PREDICTION OF SUBCOOLED BOILING WITH UNCERTAINTIES FROM BOUNDARY CONDITIONS AND INTERACTION MODELS

Tenglong Cong,Minjun Peng,Xiang Zhang
DOI: https://doi.org/10.1115/icone26-81101
2018-01-01
Abstract:Subcooled boiling has been investigated by using the RPI wall boiling models in the last two decades. High accuracy of such models has been achieved by improving the submodels for interphase actions or tuning the model parameters. However, the reliabilities of the models are still suspicious due to the limited model validation and the experimental data based model calibration. The applicability of calibrated model parameters in the new experiment data can not be assured. The effects of model parameters need to be calibrated were treated as the uncertainties of these parameters. The critical parameters that dominate the prediction of subcooled boiling were selected by using the hierarchy analysis. After that, the input samples for uncertainty analysis were obtained by an efficient Monte-Carlo sampling technology Latin Hypercube Sampling based on the hypothetic normal distribution. Then, the samples were transferred into the FLUENT code for CFD calculations. Results from CFD code were extracted for statistical analysis. Besides, the uncertainties from boundary conditions were also analyzed to quantify the effects of experimental uncertainties. The dependency of the predicted subcooling parameters and the input parameters can be obtained. A PIRT table can be drawn from the generated correlation coefficients between inputs and outputs to quantify the importance of model parameters on subcooled boiling.
What problem does this paper attempt to address?