Conference Agenda

The conventional magneto-static finite element analysis of an electrical machine is time-consuming and computationally expensive. Since each topology has a distinct set of parameters, design optimization is performed independently. This paper presents a novel method for predicting Key Performance Indicators (KPIs) of two differently parameterized electrical machine topologies at the same time by mapping a high dimensional integrated design parameters in a lower dimensional latent space using a variational autoencoder. After training, via a latent space, the decoder and multi-layer neural network will function as meta-models for sampling new designs and predicting associated KPIs, respectively. This enables parameter-based concurrent multi-topology optimization.

Brushless DC motors are used in many fractional power applications such as drills, garden tools, and small motors. This small motor has permanent magnets in its rotor, and the stator is electronically commutated. To keep the motor size as smaller as possible, it is required to operate it at high speeds. In this digest, the optimal electromagnetic design of a BLDC motor which has a rated speed of 10000 rpm is presented. Second-order response surface models were selected, and the data used from the design experiments were sampled using the Latin Hypercube method. The cost functions are minimized using the NonSorting Genetic Algorithm III, where the converge is fast. In the full paper, more than two objectives will be included, and a prototype will be manufactured to demonstrate the validity of the proposed design.

As a unique power discharging channel for the train, the integrated grounding system plays a crucial role in transmitting grounding currents back to substations. The traction current discharged from on-board traction transformers is transmitted from a working grounding wheel to the rail via a rotating electrical contact. In contrast to fixed grounding techniques normally used in power systems, the grounding impedance of each grounding wheel set varies as the train moves. As the rail forms a common discharging channel, the traction current from the working grounding wheels may flow back to the carriages through the neighboring protective grounding wheels, forming a harmful train-rail circumflux. The fluctuation of the train body potentials may threaten the safety of the on-board equipment. In this paper, an equivalent circuit of the integrated grounding system is created, involving the transiently varying grounding impedances. Using a kriging method, performance is improved through optimizing the grounding distribution and impedances.

In the framework of induction hardening, the coil design task is particularly suitable to be formulated as multi-objective optimization problem. The Pareto front estimation raises the issue of guaranteeing a satisfactory diversity and number of non-dominated solutions to be provided to the decision maker (DM). In this paper, a generative adversarial network (GAN) and a forward neural network (FNN), which is cascade connected to the GAN generator, produce additional Pareto optimal solutions starting from the results of a genetic algorithm used as training set. The FNN ensures an accurate prediction of the objectives of the added solutions, removing the need of further field analyses. This method is first tested against a set of analytical problems and subsequently validated on a 3-objective coil design task to illustrate its utility for a real-world case.

The continuum shape sensitivity of the transient eddy current system is derived by using the material derivative concept and adjoint variable technique. In the sensitivity analysis, the state variable equation is an initial-value problem to be solved forward in time. On the other hand, the adjoint variable equation is obtained as a terminal-value problem, which is solved backward in time. With the state and adjoint variables solved, the sensitivity is evaluated. In a numerical example, the eddy current distribution is optimized to make the uniform dissipated power on the conductor surface, which shows the feasibility of the derived sensitivity formula.