Active Magnetic Regenerator Research Articles

Development and experimental results from a 1 kW prototype AMR

A novel rotary magnetic refrigeration device has been designed and constructed following the concepts recently outlined in Bahl et al. (2011). The magnet and flow system design allow for almost continuous usage of both the magnetic field and the magnetocaloric material in 24 cassettes, each containing an active magnetic regenerator (AMR) bed. The prototype design facilitates easy exchange of the 24 cassettes, allowing the testing of different material amounts and compositions. Operating with 2.8 kg of commercial grade Gd spheres a maximum no-span cooling power of 1010 W and a maximum zero load temperature span of 25.4 K have been achieved. For the purpose of actual operation, simultaneous high span and high performance is required. At a heat load of 200 W a high temperature span of 18.9 K has been obtained, dropping to a span of 13.8 K at the higher heat load value of 400 W.

Experimental comparison of multi-layered La–Fe–Co–Si and single-layered Gd active magnetic regenerators for use in a room-temperature magnetic refrigerator

We present experimental comparison of different parallel-plate active magnetic regenerators (AMRs) with two different groups of magnetocaloric materials. First, a gadolinium-based single-layered AMR was tested and analysed under different operating conditions. In the next step, three different multi-layered AMRs with different compositions and different Curie temperatures made from LaFe13 − x − yCoxSiy materials were constructed and tested. Seven-, four- and two-layered La-based AMRs were evaluated. The measurements were performed with respect to the maximum measured temperature span under different operating conditions and the cooling load under different temperature spans. In order to find the optimum operating temperature range the AMRs were further compared for different hot-side temperatures. The Gd-based AMR produced a larger temperature span, especially at higher operating frequencies and higher mass-flow rates. Among the multi-layered La–Fe–Co–Si AMRs, the seven- and four-layered AMRs showed very similar characteristics, while the two-layered AMR was much poorer.

Experimental and numerical results of a high frequency rotating active magnetic refrigerator

Experimental results for a recently developed prototype magnetic refrigeration device at the Technical University of Denmark (DTU) were obtained and compared with numerical simulation results. A continuously rotating active magnetic regenerator (AMR) using 2.8 kg packed sphere regenerators of gadolinium (Gd) was employed. With operating frequencies up to 10 Hz and volumetric flow rates up to 600 L h−1, the prototype has shown high performance and the results are consistent with predictions from numerical modelling. Magnetocaloric properties of the Gd spheres were obtained experimentally and implemented in a one-dimensional numerical AMR model that includes also the parasitic losses from the prototype. The temperature span for a thermal load of 200 W as a function of frequency was measured and modelled. Moreover, the temperature span dependence on the cooling capacity as a function of cycle frequency was determined. A detailed study of these parasitic losses was carried out experimentally and numerically.

Performance Analysis and Parametric Study of an Active Magnetic Regenerator Based on the Design of Experiments Approach

Active magnetic regeneration (AMR) is refrigeration concept that provides higher energy efficiency as compared to conventional vapor compression refrigeration techniques. However, AMR poses a complex problem of heat transfer, fluid dynamics and magnetic field, which requires detailed and robust modeling. To solve this complex problem, fast and flexible numerical model is developed to study—in detail—the operation of AMR systems. Once the model is established, a sensitivity analysis has to be done and the performance characteristics have to be investigated. In order to analyze the performance of AMR systems, a new approach—based on the statistical design of experiments (DOE) method—is developed and presented in this paper. It will be demonstrated how the DOE method can help to obtain maximum information and results from the analysis of an AMR system at minimum computation cost. Furthermore, based on the developed approach, some optimal operating-regions are determined and evaluated in this paper.

THERMODYNAMIC ANALYSIS AND PERFORMANCE ASSESSMENT OF A CASCADE ACTIVE MAGNETIC REGENERATIVE REFRIGERATION SYSTEM

In the present study, a thermodynamic model is proposed to analyze and assess the performance, through energy and exergy, of a cascade active magnetic regenerative (AMR) refrigerator operation a regenerative Brayton cycle. This cascade refrigeration system works with Gd x Tb 1–x alloys as magnetic materials where the composition of the alloy varies for different stages. In this model, the heat transfer fluid considered is a water– glycol mixture (50% by weight). The refrigeration capacity, total power consumption, coefficients of performance (COP), exergy efficiency and exergy destruction rate of a cascade AMR refrigeration (AMRR) system are determined. To understand the system performance more comprehensively, a parametric study is performed to investigate the effects of several important design parameters on COP and exergy efficiency of the system.

Performance modeling of AMR refrigerators

A system modeling approach for predicting the performance of active magnetic regenerators using a one-phase approximation is presented. The approach is described for an arbitrary AMR device independent of the magnetic refrigerant, thermal losses or magnetic waveform. A general expression for magnetic work is derived which can be used for cycles where the low-field intensity is not zero. Additionally, a means of treating the varying magnetic field waveform as a single high and low field is described. The model is applied to a permanent magnet magnetic refrigerator using water–glycol as the heat transfer fluid. Simulated results are compared to experimental data which vary by heat load, frequency and utilization. A sensitivity analysis is performed using utilization, adiabatic temperature change, effective conductivity and particle size as independent variables. Comparisons to experimental data show that reducing the calculated magnetocaloric effect by 25% provides good agreement between simulations and experimental results.

A numerical comparison of a parallel-plate AMR and a magnetocaloric device with embodied micro thermoelectric thermal diodes

This article introduces a novel heat-transfer mechanism for magnetocaloric energy conversion based on the introduction of thermoelectric thermal diodes. These can provide the very rapid transport of heat from/to the magnetocaloric material. For the purposes of the study, a new mathematical-numerical model has been developed. The model serves for dynamic heat-transfer simulations and an estimation of the cooling characteristics of magnetocaloric devices, which apply thermal diodes. Micro Peltier elements were considered as the thermal diodes. A theoretical analysis was performed in order to compare the performance of a parallel-plate active magnetic regenerator with a magnetocaloric device that uses thermal diodes. The results show the great potential for magnetocaloric devices with thermal diodes, especially in terms of an increased operating frequency and specific cooling powers.

Magnetic refrigerator for hydrogen liquefaction

This paper reviews the development status of magnetic refrigeration system for hydrogen liquefaction. There is no doubt that hydrogen is one of most important energy sources in the near future. In particular, liquid hydrogen can be utilized for infrastructure construction consisting of storage and transportation. Liquid hydrogen is in cryogenic temperatures and therefore high efficient liquefaction method must be studied. Magnetic refrigeration which uses the magneto-caloric effect has potential to realize not only the higher liquefaction efficiency > 50 %, but also to be environmentally friendly and cost effective. Our hydrogen magnetic refrigeration system consists of Carnot cycle for liquefaction stage and AMR (active magnetic regenerator) cycle for precooling stages. For the Carnot cycle, we develop the high efficient system > 80 % liquefaction efficiency by using the heat pipe. For the AMR cycle, we studied two kinds of displacer systems, which transferred the working fluid. We confirmed the AMR effect with the cooling temperature span of 12 K for 1.8 T of the magnetic field and 6 second of the cycle. By using the simulation, we estimate the total efficiency of the hydrogen liquefaction plant for 10 kg/day. A FOM of 0.47 is obtained in the magnetic refrigeration system operation temperature between 20 K and 77 K including LN2 work input.

Experimental investigation of two-stage active magnetic regenerative refrigerator operating between 77 K and 20 K

This paper presents an experimental investigation on a magnetic refrigerator operating between 77K and 20K. Since the magneto-caloric effect of typical magnetic refrigerant is concentrated in a narrow temperature region, four different rare-earth intermetallic compounds (GdNi2, Dy0.85Er0.15Al2, Dy0.5Dr0.5Al2, and Gd0.1Dy0.9Ni2) are utilized according to their favorable temperature regions. The magnetic refrigerator is composed of two stages each of that is made of two different porous active magnetic regenerators. Helium gas as the heat transfer medium in the magnetic refrigerator is shuttled similar to a GM type pulse tube refrigerator so that the overall cooling capacity is enhanced by gas expansion effect when both gas and magnetic expansions are properly coordinated. The approximate peak magnetic field of 4T is provided for the internal region of the regenerators by the AC superconducting magnet which can fast ramp up to the maximum field for 3s. The fabricated magnetic refrigerator reached the lowest no-load temperature of 24K and the whole temperature span was approximately 56K rather than 57K. The cooling performance of the system is sensitive according to the load match between the first and the second stages in conjunction with helium mass flow rate distribution.

A new approach to prepare spherical La–Fe–Si–Co magnetocaloric refrigerant particles

The drop-tube solidification technique has been employed as a novel approach to prepare spherical LaFe 11 Co 0.8 Si 1.2 magnetocaloric particles, which have been shown to be highly suitable for active magnetic regenerator systems. The rapidly solidified particles have an average value of the size distribution of about 300 μm, a perfectly spherical shape and are essentially void free. Only a very brief annealing at 1373 K for 1 h was required to obtain the La(Fe,Si) 13 phase. A large magnetic entropy change of 15 J kg −1 K −1 caused by a field change of 5 T was observed around room temperature for annealed spheres.

Interface Behavior in Functionally Graded Ceramics for the Magnetic Refrigeration: Numerical Modeling

The active magnetic regenerator refrigerator is currently the most common magnetic refrigeration device for near room temperature applications, and it is driven by the magnetocaloric effect in the regenerator material. In order to make this efficient, a graded configuration of the magnetocaloric material is needed. Tape casting is a common process in producing functional ceramics, and it has recently been established for producing side-by-side (SBS) functionally graded ceramics (FGCs). The main goal of the present work is to study the multiple material flows in SBS tape casting and analyze the influence of the different material properties, i.e. the density and the viscosity, on the interface between the flows, since this is highly important for the efficiency of the device. The Newtonian flow behavior with relatively high viscosity is assumed for each fluid and used in the simulation with a commercial CFD code (ANSYS FLUENT). The results show that the density change does not affect the interface between the adjacent fluids. The viscosity of the fluids plays the most important role in the behavior of the interface. Moreover, increasing the viscosity difference of the adjacent flows, Δμ, leads to increasing the diffusive region between the two fluids.

Performance analysis of a rotary active magnetic refrigerator

Performance results for a novel rotary active magnetic regenerator (AMR) and detailed numerical model of it are presented. The experimental device consists of 24 regenerators packed with gadolinium (Gd) spheres rotating inside a four-pole permanent magnet with magnetic field of 1.24T. A parametric study of the temperature span, cooling power, coefficient of performance (COP) and efficiency of the system was carried out over a range of different hot reservoir temperatures, volumetric flow rates and cooling powers. Detailed modeling of the AMR using a 1D model was performed and compared with the experimental results. An overall mapping of the thermal losses of the system was performed, and good agreement between the experimental and numerical results was found when parasitic heat losses were subtracted from the modeling results. The performance of the system was evaluated via the COP, the exergetic-equivalent cooling power (ExQ), and the overall second law efficiency, η2nd. Losses mapping indicated that friction and thermal leakage to the ambient are the most important contributors to the reduction of the system performance. Based on modeling results, improvements on the flow distributor design and reduction of the cold end thermal parasitic losses are expected to enhance the efficiency of the system. For an operating frequency of 1.5Hz, a volumetric flow rate of 400L/h, a hot reservoir temperature of 297.7K, and thermal loads of 200 and 400W, the obtained temperature spans, ΔTS, were 16.8K and 7.1K, which correspond to COPs of 0.69 and 1.51, respectively. The maximum overall second-law efficiency was 5.6% for a ΔTS of 12.9K at 500L/h and 400W.

Magnetocaloric Effect of Perovskite Eu0.5Sr0.5CoO3

Magnetocaloric properties of the Eu0.5Sr0.5CoO3 system near a phase transition from a ferromagnetic to a paramagnetic state are investigated. It is shown that the magnetic entropy change (ΔSM) peak spanning over a broad range of temperature leads to a remarkably wide working temperature region, yielding a significant performance in terms of refrigerant efficiency. Moreover, ΔSM distribution is very uniform, which is desirable for Ericsson-cycle magnetic refrigerator. Eu0.5Sr0.5CoO3 can be used as a working material of an apparatus based on the active magnetic regenerator cycle that cools hydrogen gas.

Improved modelling of a parallel plate active magnetic regenerator

Much of the active magnetic regenerator (AMR) modelling presented in the literature considers only the solid and fluid domains of the regenerator and ignores other physical effects that have been shown to be important, such as demagnetizing fields in the regenerator, parasitic heat losses and fluid flow maldistribution in the regenerator. This paper studies the effects of these loss mechanisms and compares theoretical results with experimental results obtained on an experimental AMR device. Three parallel plate regenerators were tested, each having different demagnetizing field characteristics and fluid flow maldistributions. It was shown that when these loss mechanisms are ignored, the model significantly over predicts experimental results. Including the loss mechanisms can significantly change the model predictions, depending on the operating conditions and construction of the regenerator. The model is compared with experimental results for a range of fluid flow rates and cooling loads.

New thermodynamic cycles for magnetic refrigeration

Most of the existing prototype devices for magnetic refrigeration are based on a thermodynamic cycle with an active magnetic regenerator (AMR) that operates as a Brayton-type regenerative magnetic refrigeration cycle. However, there are several other cycles that may potentially influence not only the efficiency, but also the cost, compactness and simplicity of magnetocaloric devices. In this article we discuss the possibility of introducing new thermodynamic cycles. This is supported by information about, and a comparison of, the corresponding magnetic field sources. We present the results of numerical analyses and compare the characteristics of different thermodynamic cycles under different operating conditions and for different magnetic field intensities. Guidelines for future work on new magnetic thermodynamic cycles are presented.

Simulation of Magnetocaloric Effect in La0.7Ca0.3MnO3 Ceramics Fabricated by Fast Sintering Process

The enhanced low-field magnetocaloric effect (MCE) is simulated for La0.7Ca0.3MnO3 (LCMO) ceramics that were fabricated by fast sintering process with different Al2O3 contents. It is shown that LCMO exhibits magnetic entropy change (ΔSM) much more uniform than that of gadolinium. The results show that the peak in the MCE at the ferromagnetic to paramagnetic phase transition is improved as the sintering temperature decreases. Furthermore, the samples open up a new way in which to tune the intrinsic properties of mixed-valence manganites. Through these results, LCMO has some potential applications for magnetic refrigerants in an extended high-temperature range. It is suggested that the fast sintering process with different Al2O3 contents for LCMO is an efficient way to obtain a working material of an apparatus based on the active magnetic regenerator cycle that cools hydrogen gas.

Magnetocaloric Effect in Nanopowders of Pr0.67Ca0.33Fe x Mn1−x O3

Magnetocaloric effect in nanopowders of Pr0.67Ca0.33Fe x Mn1−x O3 (x=0,0.1,0.3,0.4), in the vicinity of magnetic phase transitions, was investigated. It is shown that Pr0.67Ca0.33MnO3 exhibits the largest magnetic entropy change (ΔS M ) of 0.61 J/kg/K at 208 K upon 0.5 T magnetic field variation. Furthermore, the ΔS M distribution of the Pr0.67Ca0.33Fe x Mn1−x O3 is much more uniform than that of gadolinium. Because of these results, nanopowders of Pr0.67Ca0.33Fe x Mn1−x O3 have some potential applications for magnetic refrigerants in an extended high-temperature range. Moreover, it can be used as a working material of an apparatus based on the active magnetic regenerator cycle that cools hydrogen gas.

Experimental investigation of the effect of thermal hysteresis in first order material MnFe(P,As) applied in an AMR device

The magnetocaloric first order material MnFe(P,As) is a candidate for room temperature magnetic refrigeration. However, these materials have intrinsic hysteresis and the impact on the refrigeration performance has not yet been thoroughly investigated in the literature.Here, the magnetocaloric effect (MCE) and the thermal hysteresis are studied using vibrating sample magnetometry. The influence on actual refrigeration performance is investigated with an established active magnetic regenerator (AMR) test device (Bahl et al., 2008), utilizing a flat plate regenerator of a single Curie temperature material.We find that the MCE curves are shifted 1.5 K when comparing heating and cooling the material, while the maximum MCE remains constant. The width of the MCE curve peak is seen to increase 0.3 K when cooling compared to heating. These results are confirmed by experiments on the AMR test device.

Numerical and experimental analyses of different magnetic thermodynamic cycles with an active magnetic regenerator

We have analyzed the influence of different magnetic thermodynamic cycles on the performance of a magnetic cooling device with an active magnetic regenerator (AMR) based on the Brayton, Ericsson and Hybrid Brayton–Ericsson cycles. Initially, a numerical simulation was performed using a 1D, time-dependent, numerical model. Then a comparison was made with respect to the cooling power and the COP for different temperature spans. We showed that applying the Ericsson or the Hybrid Brayton–Ericsson cycle with an AMR, instead of the standard Brayton cycle, can increase the efficiency of the selected cooling device. Yet, in the case of the Ericsson cycle, the cooling power was decreased compared to the Hybrid and especially compared to the Brayton cycle. Next, an experimental analysis was carried out using a linear-type magnetic cooling device. Again, the Brayton, Ericsson and Hybrid Brayton–Ericsson cycles with an AMR were compared with respect to the cooling power and the COP for different temperature spans. The results of the numerical simulation were confirmed. The Hybrid Brayton–Ericsson cycle with an AMR showed the best performance if a no-load temperature span was considered as a criterion.

Material properties and modeling characteristics for MnFeP1−xAsx materials for application in magnetic refrigeration

Compounds of MnFeP1−xAsx have received attention recently for their use in active magnetic regenerators (AMR) because of their relatively high isothermal entropy change and adiabatic temperature change with magnetization. However, the materials also generally exhibit a significant magnetic and thermal hysteresis, and it is not well understood how the hysteresis will affect performance in a practical AMR device. The amount of hysteresis shown by a material can be controlled to an extent by tuning the processing conditions used during material synthesis; therefore, knowledge of the practical impact of hysteresis is a key element to guide successful material development and synthesis. The properties of a magnetocaloric MnFeP1−xAsx compound are characterized as a function of temperature and applied magnetic field, and the results are used to assess the effects of hysteresis on magnetocaloric properties. Different methods of building property functions from the measured specific heat, magnetization, and adiabatic temperature change are presented. It is shown that model predictions can be highly dependent on how the properties that are used by the AMR model are calculated.