Computer based optimization of permanent magnets

Orateur : Thomas SCHREFL (Department for Integrated Sensor Systems, Danube University Krems, Austria)

Permanent magnets play a decisive role in green technology. Today, high-performance permanent magnets are based on Nd2Fe14B. In many applications, magnets are used at elevated temperatures. Therefore, Nd may be partially substituted with heavy rare earth elements, in order to improve the coercive field at the operating temperature. Current efforts in magnet development aim at the reduction of the total rare-earth content, while maintaining suitable magnetic properties. In the talk I will show how computer simulations and machine learning can be used to find guidelines for the optimization of permanent magnets. The computation of the saddle point for magnetization reversal helps to visualize the onset of magnetization reversal. Micromagnetic simulations show how reversed domains expand. Machine learning can identify the weak grains in the structure. Understanding when and where magnetization reversal starts, helps to improve the magnet by the local modification of the structure or the chemical composition.

Télécharger dans mon agenda