Prediction and optimisation of geometrical properties of the refrigerant bed in an AMRR cycle

Abstract Active magnetic regenerative refrigeration cycle is nowadays investigated in magnetic refrigeration prototypes in order to increase the temperature span between the Hot source and the Cold one. The AMRR model can give the magnetothermal behaviour of the regenerative refrigerant bed and the system cooling power considering the geometrical and the physical properties of the fluid and the material. In this paper an inverse approach is considered. Starting from the system performance required (cooling power, temperature profile, etc.); we go up by optimisation processes to the dimensions of the system. Let's consider for example, a desired temperature profile. These curves are called the characteristics of the system. The dimensions of the system can be found out by using temperature characteristics for different geometrical configurations of the bed. In addition, the optimisation can be performed under external constraints, i.e. limited variation band of the parameters. When obtained, the results can be used to optimise the whole shape of the system.