Deterministic and Stochastic Methods in Discrete Topology Optimization of a Permanent Magnet Synchronous Machine With Surface Mounted Magnets

To meet the given goals and requirements of the operational characteristics of electrical machines, different optimization concepts and methods are in use. The easiest concept is to find the best adapted geometric dimensions of the active flux guiding parts of an electrical machine assuming a fixed topology. This is the common method used in the design and optimization of electrical machines. An extension of this concept is to assume in some kind adaptable topologies, wherein the edges between the different active parts inside the electrical machine having different materials can be changed to flexible and variable contours. The most flexible concept, but also the most complex and computationally intense concept, is the discrete topology optimization based on a discretization of a locally predefined design space with cells, in which different materials can be freely filled in to reach the given requirements. This latter optimization concept leads to discrete or combinatorial optimization problems. For finding solutions of such discrete or combinatorial optimization problems different search methods exist. Some of these optimization methods will be investigated using an example of a permanent magnet synchronous machine with surface mounted permanent magnets, where the topology of the magnet, especially, the optimal material distribution inside the cells of the design space of the magnet is to be found to meet a given single optimization goal. The suggested and investigated methods can be applied to other topology optimization problems of the active parts of electrical machines as well.