Additively Manufactured Molten Salt to Supercritical Carbon Dioxide Heat Exchanger

Authors:

Abstract
Design, fabrication, and experimental characterization of the performance of a molten salt (MS)-to-supercritical carbon dioxide (sCO2) heat exchanger (HE) for concentrating solar power (CSP) applications using laser powder bed fusion (LPBF) additive manufacturing (AM) of nickel-based superalloys is presented in this report. The hypothesis is that such additively manufactured HEs will result in a monolithic, compact, low-pressure drop HE that is durable under cyclic operation at high temperature and high pressure in a corrosive salt environment. Two nickel-based superalloys were considered in the project- Haynes 230 (H230) and Haynes 282 (H282). Process windows for LPBF AM were established to provide dense low porosity parts for both alloys. Creep tests results on heat treated AM H230 and H282 samples showed that the AM fabricated specimens were as robust as the wrought specimens and showed no deficit in creep strength at high temperatures. Haynes 282 was down selected for the HE fabrication based on better dimensional tolerance, better corrosion resistance to chloride salt, and higher creep strength. The HE was designed for an inlet temperature of 720 °C on the chloride salt side and 500 °C on the sCO2 side. The design system pressure on the sCO2 side was 200 bar, with a pressure drop less than 2% of line pressure on the sCO2 side. To reduce parasitic losses in the MS loop, it was desired to have a low exit temperature on the MS exit. Two designs (generation 1 and 2) of the HE were developed by a co-design effort considering mechanical stresses, thermal and fluidic performance, and AM fabrication constraints such as overhangs, powder removal and build dimensions. The HE core designs consist of a pin array on the sCO2 side and either straight or 3D periodic lattice on the MS side. Additional design considerations included creep, fatigue and thermal stresses. A key concern for use of chloride salts in CSP pertains to the high corrosion rate. To assess whether AM printed specimens would have different corrosion rates than wrought samples, corrosion tests on printed H230 and H282 specimens were performed in purified chloride salt with baseline wrought samples as reference. AM H282 was superior to AM H230 with a best corrosion rate of around 110 µm/year compared with heat treated H230 corrosion rate of 268.3 ± 28.8 µm/year. One of the key issues with LPBF is the roughness and feature accuracy of the geometry, both of which can cause excessive pressure drop. Two methods of smoothening the internal flow passages of the HE were explored- a physical method of abrasive liquid slurry flow, and a novel iodine-based chemical treatment. The pressure drop through the pin array is reduced by 15-30% with the liquid abrasive slurry treatment, with varying reductions depending on the extent of the treatment, geometry of the pin array, and mass flow rate. The iodine-based is a Self-Terminating Etching Process (STEP) to uniformly remove material from all fluid-accessible surfaces while preserving the intricate geometrical features. Pressure drop reductions of up to 21% were measured between the baseline and chemically treated pin array. A simplified thermo-fluidic sectional model, based on correlations for pin arrays in open literature, was developed for design and analysis of the HE performance. With account for the larger flow frontal area caused due to the down-skin on the pins, it is possible to get reasonable predictions for heat transfer using the Rasouli et al. (2018) correlation for the designs experimentally characterized in this study. The Prasher et al. (2007) correlation provided the best fit to experimental pressure drop data. A process-based cost model (PBCM) that included powder, LPBF print time, and post processing steps was developed to estimate the cost of the HE. The PBCM was used to assess the impact of design changes, process parameters, and innovations in LPBF technology on the cost of the HE. To experimentally demonstrate the performance of the HE, a 20 pair unit was fabricated and characterized. Heated air was used as a surrogate to the chloride salt on the hot side and sCO2 in the 100-200 bar range was used on the cold side. The results of exit temperatures and heat transfer rate were compared against the model. The model was able to predict experimental results to within 0.8% percent for heat transfer rate, 0.97% percent for exit sCO2 temperature, and 1.5% percent for exit air temperature. Preliminary work on potential to scale up the HE was undertaken through demonstration of taller builds in the LPBF machine. The validated HE model was used to determine the performance of a scaled-up design. With a 20 cm tall build that covers 24 cm x 24 cm of the build plate, a 166 kW HE can be built in a single laser machine with a bed size of 25 cm x 25 cm. The impact of system-level variations in inlet temperature on the performance of this HE was established. In summary, it was demonstrated in this project that it is feasible to design and fabricate, using metal AM, compact primary HEs that meet the performance targets at the stringent operating conditions encountered in the liquid pathway for Gen3 CSP power plants, and established cost metrics.
Funding Information
  • USDOE Office of Energy Efficiency and Renewable Energy (EERE), Renewable Power Office. Solar Energy Technologies Office
  • USDOE Advanced Research Projects Agency - Energy (ARPA-E)
  • DOE (EE0008536)
  • DOE (AR0001127)

This publication has 35 references indexed in Scilit: