Quantitative analysis of iron loss taking magnetic hysteresis behavior into account is essential to development of high-efficiency electric machines. Among various hysteresis models, a play model has attracted attention in recent years, since it needs no convergence calculation. However, to derive the play model, a large number of dc hysteresis loops with different maximum flux densities have to be measured in the preliminary experiment, which is difficulty in practical use. To overcome this problem, we presents an efficient method for obtaining the play model from the LLG equation, which requires only one or two measured dc hysteresis loops. Furthermore, in order to calculate iron loss including magnetic hysteresis behavior, we propose that the play model obtained from the LLG equation is incorporated into a reluctance network analysis (RNA) by using a reactor as an object of discussion.

A number of machines have been fitted with electric motors and their inverters. In order to reduce the space of the motors, we propose a two-controllable-rotor motor which consists of 2 rotors and 1 stator. However, the two-controllable-rotor motor is driven by a 6-phase inverter, and the space of its inverter is almost the same as that for 2 motors driven by 3-phase inverters. In order to reduce the space of the inverter, we propose a two-controllable-rotor motor using a 5-phase inverter, and its operation principle is investigated using JMAG-Designer.

This poster reports a variable flux PM motor based on the new principle. The principle of this proposed motor is based on the self-excitation utilizing space harmonics that are inevitably generated in the concentrated winding structure. In addition, it has a passive variable flux function. Its rotor is composed of consequent poles for each N-pole and S-pole pair and self-excited wound-field poles. The effects of automatic variable flux function on driving performance and its efficiency map are demonstrated via the prototype.

Among various types of electric motors, variable flux motors (VFM) receive much attention in terms of expanding high efficiency area and operating area required for automotive traction drive application. As one of the types of VFM, this project focuses on hybrid excitation flux switching motor (HEFSM) which can realize variable flux capability utilizing both the field excitation coils (FECs) and the permanent magnets (PMs) arranged on a stator body. To improve the motor efficiency of the conventional HEFSM, a new HEFSM in which a part of FEC slots is replaced with “Variably Magnetizable PMs (VM-PM)” is proposed. When designing the motor with VM-PMs, it is essential to estimate the magnetization characteristics of VM-PMs at the design stage. In this presentation, the magnetization analysis using JMAG-Designer based on the B-H data of the VM-PM is examined from a viewpoint of the estimation accuracy.

Recent improvements in GPU performance have been remarkable, for example, the NVIDIA Tesla K20 released in 2012 had a single-precision arithmetic performance of 3.52 Tflops, while the NVIDIA Quadro GV100 released in 2018 has more than four times the computational performance at 14.8 Tflops.
This paper reports on the transient response analysis of IPM/SPM motors using the NVIDIA Quadro GV100 with three different methods: non-parallel CPU, shared memory parallel, and using the GPU.

The transient response analysis is known to be more accurate than the frequency response analysis for the evaluation of induction motors. On the other hand, in the case of circuit-coupled analysis using a voltage source, a great number of steps are required to reach a steady-state solution due to the nonlinear magnetic properties of the magnetic steel sheet. When creating an efficiency map, not only are there many operating points to evaluate, but it is also necessary to find the drive conditions that will result in the highest efficiency at each operating point. In order to achieve this, parallel computing under a supercomputer environment is effective. In this poster, we introduce the loss evaluation and efficiency mapping of an induction motor using the supercomputer “TSUBAME” at Tokyo Institute of Technology.

Recently, IPMSM (Interior Permanent Magnet Synchronous Motor) are widely used for driving source of electronic vehicle.　 Electric vehicles are required to be quietness and silence. However, IPMSM cause torque ripple that leads vibration and noises. So it is important to reduce the torque ripple.
In developing a control that reduces torque ripple, it takes a lot of development terms to build control after it verify actual machine. Therefore, control development by simulation using a motor model is necessary.
This presentation shows that　torque ripple reduction by feed-forward control was simulated using a motor model suitable for the actual machine and reduction　effect was verified using interaction analysis by MATLAB/Simulink and JMAG-RT for motor model.

Electrification in automotive industry has brought about an increase of the number of modules and therefore, down-sizing of components becomes important. In this project, a compact isolated DC-DC converter for hybrid vehicles utilizing gallium nitride in Silicon (GaN/Si) power devices is studied. In the conventional converter, the high step-down ratio from the high voltage main battery to the low voltage sub-battery for auxiliary system increases the secondary current after passing through the transformer, resulting in a large conduction loss. Therefore, a matrix transformer for dividing the secondary current is examined. A flux cancellation method to reduces the core size and loss is introduced in the transformer design. The prototype transformer can integrate the secondary rectification board. In this presentation, the comparison between the design results of the inductance values of matrix transformer using JMAG-Designer and the measurement results of those values obtained from the prototype is reported.