DNB Prediction Method Using Local Void Distribution in Rod-Bundle

Abstract

A computational multi-fluid dynamics (CMFD) code that predicted the void distribution in sub-channel was coupled with sub-channel analysis code to predict departure from nucleate boiling (DNB). The main assumption was that the void fraction near heated wall was the dominant parameter in DNB. A sub-channel analysis code was used to calculate three dimensional distribution of sub-channel averaged values of mass flux, void fraction, density and quality. These were used as a boundary condition in the CMFD code to predict local void fraction in a subchannel. A bubble diffusion equation was used assuming the wall peak void distribution caused by turbulence. The present method was applied to the analysis of DNB tests. The coupled codes showed a reasonable profile of void fraction in a rod bundle and reproduced DNB heat flux at low void fraction. To investigate this analysis result, the local condition were compared with a DNB flow regime map. This suggested an approach to improve the predictability: the critical void fraction should be modified at low void fraction condition; the bubble diffusion model should be modified to handle the flow regime transition from the isolated nucleation type to churn turbulent flow type.