Session: 07-04: Modeling of Thermal Energy Storage and Receiver Systems

Paper Number: 169963

169963 - Dem Modeling for Dense Moving Bed Heat Exchanger Inform Comprehensive New Nusselt–peclet Number Correlations 

Abstract: 

Dense moving beds of particles have advantages of superior solid-fluid mixing and heat and mass transfer capabilities and are widely employed in applications such as next-generation concentrating solar power plants and particle-based chemical reactors. To generalize the flow and heat transfer behaviors of the dense moving beds, previous research has primarily focused on developing Nusselt number correlations from experimental measurements and analytical derivations. However, the existing empirical and semi-empirical correlations are largely based on continuum flow assumptions or cover limited parameters with short range. Compared to others, this study extends LIGGGHTS, a discrete element method (DEM) simulator, with custom-developed conduction and radiation modules, to model the flow and heat transfer in dense moving bed heat exchanger. Our work comprehensively examines both geometrical and particle thermophysical properties, varying channel length (L = 0.1 – 0.5 m), width (W = 5 – 8 mm), and particle thermal conductivity (k_s = 0.05 – 5 W/m²/K) and diameter (d_p = 0.2 – 0.8 mm), and solid volume fraction (ϕ_s = 0.52 – 0.63). The overall channel-averaged particle-to-wall heat transfer coefficient is calculated under diverse combinations of parameters, revealing non-linear dependencies under different mass flow rates. Our findings examine the strong influence of solid volume fraction in the dense moving beds of particles, which is not captured in the existing Nusselt number correlations, including the Nusselt number correlation from Sullivan and Sabersky, and the Graetz number correlation derived by Muzychka. The high-fidelity data from DEM simulations is directly used to generate new Peclet number and Nusselt number with expansion of four dimensionless terms --- width-to-length ratio (W/L), solid-to-fluid thermal conductivity ratio (k_s/k_g), width-to-particle size ratio (W/d_p) and solid volume fraction (ϕ_s). The corresponding powers for these four terms are calculated simultaneously through a constrained optimization approach – nonlinear support vector machine regression (SVR). The newly developed Nusselt-Peclet number correlations are compared with 6 sets of experimental data obtained from other literatures, each with unique setup parameters. The comparison demonstrates good alignment with most of experimental data, validating the effectiveness of the integration for multivariate parameters. Overall, the new correlation is versatile to offer good estimations and predictions for moving bed heat exchangers with the combined effects of Width, W, length, L, particle size, d_p, thermal conductivity, k_s, and solid volume fraction, ϕ_s.

Presenting Author: Bingjia Li University of Michigan

Presenting Author Biography: Ph.D., Mechanical Engineering, University of Michigan, Ann Arbor, 2020 - 2025
M.S., Mechanical Engineering, University of Michigan, Ann Arbor, 2018 - 2020
B.S., Building Environment & Energy Application Engineering, China University of Mining and Technology, 2014-2018