Thermal–Hydraulic Design in the Transition Laminar–to–Turbulent Flow Regime

Difficulties are experienced during the thermal–hydraulic design of a nuclear reactor operating in the transition flow regime and are the result of the inability to accurately predict the heat transfer coefficient (HTC). Experimental values for the HTC in rectangular channels are compared with the calculated by correlations usually used for the design of material testing reactors (MTR). The values predicted by Gnielinski and Kreith correlations at Reynolds numbers below 5000 are not necessarily conservative. The Al-Arabi-Churchill correlation with the correction proposed by Jones has proved to be conservative for Reynolds between 2100 and 5000. Two alternative design approaches are proposed to solve a specific thermal–hydraulic design problem for a MTR operating at Reynolds 2500. The conservative approach comprises two alternatives: the use of Al-Arabi correlation with no uncertainty factors, as it has proved to be conservative, or the use of Kreith correlation with a maximum uncertainty. In this conservative approach, maximum deviations in other input parameters are also taken into account. The best estimate plus uncertainty approach considers an uncertainty distribution in input parameters to generate a random sample of 59 inputs. An uncertainty distribution based on the ratio between the experimental and the calculated HTC, when using Kreith correlation, is considered. Results are given in terms of maximum and minimum bounds for the figure of merit used as design criterion with 95% probability and 95% confidence level. The best estimate plus uncertainty approach offers a less penalizing design and its use depends on regulator's acceptance.