Experimental investigation of a printed circuit heat exchanger using supercritical carbon dioxide and water as heat transfer media

Abstract

The Secure Transportable Autonomous Reactor – Liquid Metal system combines a Generation IV nuclear reactor with an advanced Supercritical Carbon Dioxide (S-CO[subscript]2) Brayton power conversion cycle. The Brayton cycle was selected as the power conversion cycle due to its high efficiency, small turbomachinery size, and competitive cost due to reduced complexity as compared to a traditional Rankine cycle. Overall system thermal efficiency is closely tied to the performance of the precooler and recuperators. The Printed Circuit Heat Exchanger (PCHE) manufactured by Heatric is being considered for use as both the precooler and recuperator in the STAR-LM system due to its high effectiveness, wide temperature and pressure operating range, small size, and low cost. PCHEs have been used primarily in the hydrocarbon processing industry to date, and are relatively new in being considered for nuclear applications. In this study, a PCHE is investigated using S-CO[subscript]2 and water as the heat transfer media in conditions relevant to the precooler in the STAR-LM system. Experiments conducted with small temperature differences across the PCHE revealed that the heat transfer coefficient is strongly correlated with the temperature-dependent specific heat near the pseudocritical point. The STAR-LM precooler outlet temperature is near the pseudocritical point, making this region of interest to this work. Testing was conducted to determine the effect of property variation near the precooler outlet in conditions with large temperature differences in the PCHE. These tests revealed that maintaining the precooler outlet temperature near the pseudocritical point does not have a significant effect on heat transfer coefficients in the PCHE under large temperature difference test conditions. Computational Fluid Dynamics (CFD) models were developed to simulate fluid flow and heat transfer in the PCHE. A 2D, 4-channel, zig-zag model was found to reproduce the outlet temperatures to within approximately 15% relative error. The 3D straight channel model reproduced the experimental data to within 3% relative error for the cases simulated. Both of these models predicted the water side outlet temperatures to within 20% relative error.