This doctoral thesis deals with the heat transfer in magnetic cooling using solid-state thermal diodes and fluidic thermal switches. In the first part we developed a 1D numerical model of a magnetocaloric device with two built-in thermal diodes. We focused on the properties of the materials suitable for thermal diodes and thermal rectification in a transient operating regime. The rectification ratio was also determined using neural networks. In the second part, we developed a 1D numerical model for the simulations of static thermal switches. We determined the most influential parameters with a sensitivity analysis, and the upgrade of the model with a genetic algorithm allowed the optimization of the dimensions of the thermal switch and the magnetocaloric material according to the maximum temperature span established between the heat source and the heat sink. The operation of two concepts of thermal switches was analysed experimentally: one was based on electrowetting and the other on ferrohydrodynamics. The operation of the latter was also tested in a magnetocaloric refrigeration device.
The results of the numerical analyses confirmed that magnetocaloric devices with thermal switches or thermal diodes can operate at higher frequencies than the currently used active magnetic regenerators, but we were not able to confirm this experimentally within the scope of this work. The reason for this in the case of thermal diodes is the limited state of the art and the unavailability of suitable materials, and in the case of thermal switches, the switching ratio was smaller and the response time longer than expected. Nevertheless, the guidelines for the rapid development of micro-technology, new material processing and machining, indicate the possibility of an experimental confirmation of the concept in the coming years.
|
