A Numerical Model of an Active Magnetic Regenerator Refrigeration System

Abstract

Active magnetic regenerator refrigeration (AMRR) systems are an environmentally attractive space cooling and refrigeration alternative that do not use a fluorocarbon working fluid. Two recent developments have made AMRRs appear possible to implement in the near-term. A rotary regenerator bed utilizing practical and affordable permanent magnets has been demonstrated to achieve acceptable COP. Concurrently, families of magnetocaloric material alloys with adjustable Curie temperatures have been developed. Using these materials it is possible to construct a layered regenerator bed that can achieve a high magnetocaloric effect across its entire operating range, resulting in an improved COP.There is currently no tool capable of predicting the performance of a layered AMRR. This project provides a numerical model that predicts the practical limits of these systems for use in space conditioning and refrigeration applications. The model treats the regenerator bed as a one dimensional matrix of magnetic material with a spatial variation in Curie temperature and therefore magnetic properties. The matrix is subjected to a spatially and temporally varying magnetic field and fluid mass flow. The variation of these forcing functions is based on the implementation of a rotating, multiple bed configuration. The numerical model is solved using a fully implicit (in time and space) discretization of the governing energy equations. The nonlinear aspects of the governing equations (e.g., fluid and magnetic property variations) are handled using a relaxation technique.The model is used to optimize AMRR applications by varying model inputs such as matrix material, fluid mass flow rate, working fluid, reservoir temperatures, and the variation of the Curie temperature across the bed. The preliminary model has been verified qualitatively using simple cycle parameters and constant property materials and quantitatively by comparing the results with prior solutions to the regenerator governing equations in the limits of constant properties and no magnetocaloric effect. A second goal of this project is to create a cost estimate for a future project that will design, build, and test a prototype AMRR to be used to verify the numerical model.

Exploring the efficiency potential for an active magnetic regenerator

“…Since then, an increasing number of AMR prototypes based on permanent magnets have been reported by various groups (Bjørk et al 2010, Kitanowski et al 2015. Besides this, numerical AMR modelling has indicated the possibility of obtaining competitive COPs (Engelbrecht et al 2006, Engelbrecht 2008. However, technological and scientific challenges, spanning from development and basic understanding of the MCMs themselves to actual machine design issues, have made the road towards the ultimate goal of a magnetic refrigerator with a competitive COP long and bumpy -and we are not quite there yet.…”

Modeling of in-vehicle magnetic refrigeration

“…Extreme material properties have been invoked to fix performance boundaries and aid to find the most promising development paths. Both sets of simulations uses algorithms based on a 1D approach (Engelbrecht et al, 2005) employ the use of Curie temperature graded magnetocaloric materials stressing the importance of accurate control of the Curie temperature in order to employ multiple materials in reaching the goal. The main factors influencing the performance in both approaches have been studied and compared.…”

The effect of flow maldistribution in heterogeneous parallel-plate active magnetic regenerators

“…The influence on AMR performance due to inhomogeneity of parallel plate stacks is investigated using a 1D AMR model previously published [19,22]. The model is well suited for this task since it is possible to directly alter the heat transfer coefficient describing the convective heat transfer between the solid and the fluid.…”

“…In this paper an established 1D numerical AMR model [19] is used to quantify the influence on AMR performance due to inhomogeneity of the stack. The Nusselt number scaling factor (Nu scl ) is applied as a function of Reynolds number in the model.…”