WP3: Heat pump design and construction

In this work package the combination of the different components of magnetocaloric heat pumps into an optimised overall design will be studied. The knowledge gained from the study of materials, magnet assemblies as well as the system modelling and dimensioning will be consolidated towards the objective of dimensioning and constructing a magnetocaloric heat pump device. This will also draw heavily on the experience in this area gained through the previous successful design and construction of a magnetocaloric refrigeration prototype.

WP 3.1: Design of magnetocaloric regenerators for heat pumps – PhD project at DTU

The aim of this work package is to optimise the heat transfer in the regenerator. A regenerator geometry will be designed using the measured properties of the materials. Shaping of the materials and design of the flow system will be done allowing a high fluid flow with a minimum flow resistance while retaining high heat transfer performance.


Initially the existing one-dimensional numerical model available at DTU Energy was improved by implementing a new spatial discretization method to avoid any misleading oscillations of the results that would otherwise occur.The impacts of flow arrangement and dead volume were investigated and a study of AMR regenerator geometry was carried out. Multi-layer active magnetic regenerators using the solid-state refrigerant La(FeSiMn)13Hy, which presents a large but quite sharp isothermal entropy change, were studied. The impact of the number of layers and the sensitivity to the working temperature as well as the temperature span were quantified using a one dimensional numerical model. A study of the sensitivity of variation in Curie temperature through a uniform and normal distribution was also carried out.

An experimental passive regenerator test system for investigating the sensitivity of regenerators to the geometries, operation parameters and oscillating AMR flow has been developed.

Task 3.2: Heat transfer fluids – Partial PhD project at Univ. Ljubljana

The effect of different heat transfer fluids will be studied theoretically and experimentally in order to decide which is the optimal for device usage. The balance between viscous and thermal properties will be studied. Using additives to increase the thermal properties will be considered. Importantly, the chosen fluid must operate over the full range of temperatures relevant to heat pumps.


Firstly, the applicability of different possible heat transfer fluids was studied by numerical modelling. Twenty different fluids were considered in the study divided into five sections: water, alcohols, silicone oils, nano-fluids and liquid metals. Each fluid’s thermal and viscous properties were implemented into the AMR numerical model to simulate the performance of each fluid for a given AMR geometry. The results of numerical modelling showed the potential of two specific fluids to be further considered in the experimental evaluation. These fluids were water and the liquid metal Galinstan. Water and Galinstan were experimentally tested in the AMR testing device at UL. High price, difficulty of handling and disappointing test results led to the decision to discontinue the work with Galinstan. So water, with suitable additives to avoid corrosion and freezing, will be used as the heat transfer fluid for the magnetocaloric heat pump developed in ENOVHEAT.

WP 3.3: Design, construction and test of heat pump – Postdoc at DTU

Testing the components as well as designing and constructing the heat pump will be done in close collaboration with the industrial partners. Novel ideas for magnetocaloric devices, including magnet-motor integration and a shifted parallel design, will be investigated.


A review of the different configurations existing in literature has been conducted, to aid in the design choices for the prototype device to be built. From the experiments on the recently developed AMR described above, the performance was found to be very sensitive to the difference in flow rates in each regenerator, showing very quick degradation as the difference increases. Thus, the best way to control the flow system inside the regenerators, in order to have the same flow rate in all the channels is being considered.