Modeling heat transfer through corrosion product deposits on fuel rods in pressureized water reactors

Abstract

CRUD is the deposit that is found on the fuel cladding surface, and it has been suspected as one of the challenges regarding higher burnup operation and power uprates. Two major phenomena caused by CRUD – CRUD Induced Power Shift (CIPS) and CRUD Induced Localized Corrosion (CILC) are highly related to the heat transfer mechanism within the CRUD since these are highly dependent on the boiling rate and the temperature profile within the CRUD. The CRUD is the porous deposit, which contains “chimney” where vapor is escape through during boiling regime. Currently, most of studies were done for the nucleate boiling regime of the CRUD, but the heat transfer regimes of the CRUD are still not well clarified. Therefore, the heat transfer regimes of the CRUD was carefully investigated using existing database from the CRUD heat transfer experiments. From the study, it was found that there are five heat transfer regimes in the CRUD, and these are named as liquid-saturated regime, uninhibited nucleate boiling regime, capillary nucleate boiling regime, confined film boiling regime and spilled-over film boiling regime. Among them, the nucleate boiling regimes and the confined film boiling regime are related to the phenomena observed in the plant – CIPS and CILC. In this paper, the model for the confined film boiling model is suggested. Unlike the traditional approaches that use a capillary force as a unique source of the liquid supply in pores, we adopted the disjoining pressure term for the liquid supply terms. The disjoining pressure is known to be proportional to the temperature difference of the channel wall and the bulk vapor of the pore. This temperature difference was expressed in terms of the heat flux, vapor film thickness, and the thermal conductivity of the vapor film. As a result, by comparing the supply terms and demand terms of the liquid, the explicit expression for the vapor film thickness could be obtained. As the first step of the validation, the prediction from the model was compared to the experimental results from the heat pipe experiment where measured the vapor film thickness directly. The researchers of the experiment also suggested their own model that only used the capillary pressure as liquid supply term. From the comparison, we found that the present model predicts the vapor film thickness well, which could not be predicted from the model by the researchers of the experiment. For the next step, the validation to the CRUD heat transfer experiment was done. As a result, we found that the present model predicts the wall superheat with normalized root mean square of 30%. Furthermore, it was found that the present model was not sensitive to the selection of the effective thermal conductivity model for the vapor film.