Heat transfer behavior of molten nitrate salt

Abstract

Abstract. The usage of molten nitrate salt as heat transfer fluid and thermal storage medium decouples the generation of electricity from the variable nature of the solar resource, allowing CSP plants to avoid curtailment and match production with demand. This however brings some unique challenges for the design of the molten salt central receiver (MSCR). An aspect critical to the use of molten nitrate (60wt%/40wt% -NaNO 3 /KNO 3 ) salt as heat transfer fluid in the MSCR is to understand its heat transfer behavior. Alstom collaborated with the University of Wisconsin to conduct a series of experiments and experimentally determined the heat transfer coefficients of molten nitrate salt up to high Reynolds number (Re > 2.0E5) and heat flux (q″ > 1000 kW/m 2 ), conditions heretofore not reported in the literature. A cartridge heater instrumented with thermocouples was installed inside a stainless steel pipe to form an annular test section. The test section was installed in the molten salt flow loop at the University of Wisconsin facility, and operated over a range of test conditions to determine heat transfer data that covered the expected operating regime of a practical molten salt receiver. Heat transfer data were compared to widely accepted correlations found in heat transfer literature, including that of Gnielinski. At lower Reynolds number conditions, the results from this work concurred with the molten salt heat transfer data reported in literature and followed the aforementioned correlations. However, in the region of interest for practical receiver design, the correlations did not accurately model the experimentally determined heat transfer data. Two major effects were observed: (i) all other factors remaining constant, the Nusselt numbers gradually plateaued at higher Reynolds number; and (ii) at higher Reynolds number a positive interaction of heat flux on Nusselt number was noted. These effects are definitely not modeled by the existing correlations. In this paper a new Nusselt number correlation is shown to accurately model the experimental data. A theory to explain the heat transfer trends at higher Reynolds number and heat flux is proposed. Finally, the impact of these results on receiver design is discussed, with comparison between correlations available from literature and that proposed in this study.

Experimental analysis of forced convective heat transfer of nitrate salt in a spirally grooved tube at high Reynolds numbers and temperatures

“…However, the available experimental data for molten salts is limited to Reynolds numbers up to 30,0 0 0. This is considerably lower than the Reynolds number used for example in concentrating solar power (CSP) applications, which are commonly operated at Reynolds numbers ranging between 10,0 0 0 and more than 20 0,0 0 0 [ 4 ]. Also, compared to the experimental data, the general correlations overpredict the measured Nusselt number ratio of other authors by up to 90% and the friction factor ratio by up to more than a factor of two.…”

“…As a consequence, high flow velocities are employed in nitrate salt tubular heat exchangers. This commonly leads to Reynolds numbers larger than 20 0 0 0 0 in smooth tubes in order to achieve the desired high forced convective heat transfers [4] . On the other hand, this practice leads to high pressure drops in the heat exchanger and consequently high pumping costs.…”

Using corrugated tubes in external molten salt receivers

“…However process flow conditions in molten salt receiver tubes exceed the experimentally explored range of Reynolds number for molten salts both for smooth and enhanced tubes by a large margin. Das et al [11] found that at lower Reynolds number conditions, the available experimental molten salt heat transfer data followed the commonly used correlations for smooth tubes [12], [13]. However, in the region of interest for practical receiver design (Re>50'000), the correlations overestimate the experimentally determined heat transfer data by a factor of up to 3.…”