The emergence of deep learning model GAN (Generative Adversarial Networks) is an important turning point in generative modeling. GAN is more powerful in feature and expression learning compared to machine learning-based generative model algorithms. Nowadays, it is also used to generate non-image data, such as voice and natural language. Typical technologies include BERT (Bidirectional Encoder Representations from Transformers), GPT-3 (Generative Pretrained Transformer-3), and MuseNet. GAN differs from the machine learning-based generative model and the objective function. Training is conducted by two networks: generator and discriminator. The generator converts random noise into a true-to-life image, whereas the discriminator distinguishes whether the input image is real or synthetic. As the training continues, the generator learns more sophisticated synthesis techniques, and the discriminator grows into a more accurate differentiator. GAN has problems, such as mode collapse, training instability, and lack of evaluation matrix, and many researchers have tried to solve these problems. For example, solutions such as one-sided label smoothing, instance normalization, and minibatch discrimination have been proposed. The field of application has also expanded. This paper provides an overview of GAN and application solutions for computer vision and artificial intelligence healthcare field researchers. The structure and principle of operation of GAN, the core models of GAN proposed to date, and the theory of GAN were analyzed. Application examples of GAN such as image classification and regression, image synthesis and inpainting, image-to-image translation, super-resolution and point registration were then presented. The discussion tackled GAN’s problems and solutions, and the future research direction was finally proposed.

This paper presents a method for reducing the cogging torque for a sloping notch with two notches applied on the stator teeth. The accuracy of FEA was confirmed by a comparison with a previous model using an asymmetric notch for the experiment data and 3D FEA results, followed by a comparison of the cogging torque of a two notches model and a sloping notch model. The sloping notch model was modified to a step-sloping notch model in consideration of a potential manufacturing process. The optimal design for minimizing the cogging torque was developed considering the sloping degree, angle, position, and size of the notches. As the optimal design result, the cogging torque on the optimal model was reduced. Finally, the analysis and optimal design results were confirmed by FEA.

The aim of this work was to eliminate the backflow power present in an isolated three-port DC–DC converter. Backflow power (which is an inherent property of phase-shifted DC–DC converters) is the major contributor of circulating current the converter, which in turn is known to be the leading cause of system loss. The dual phase shift (DPS) control scheme (which alters the transformer voltage waveform at the different winding terminals) was used to reduce the backflow power. Mathematical relations for the backflow power present in the three-port converter were derived. And from this equation, an operating point in which the backflow power is zero was also obtained. This condition for zero backflow power was confirmed by simulations on PSIM. Added to this were simulation results that show other operating conditions in which zero backflow power could be obtained in the converter. In addition, equations for the power processed at the different ports of the converter were also presented, and 3D plots were made to illustrate the variation of the power and backflow power with respect to the phase shift ratios of the DPS control scheme. It was observed that the backflow power can be totally removed from the three-port converter when using this control scheme.

This paper proposes the influence analysis of silicon carbide (SiC) MOSFET’s parasitic output capacitance on a dual active bridge (DAB) converter. Power converters are required for DC grids and energy storage. Because SiC metal-oxide-semiconductor FETs (MOSFETs) have lower on-state resistance and faster reverse recovery time than Si MOSFETs, they can be controlled with lower losses and higher frequencies. MOSFETs have a parasitic capacitance. Because of the output parasitic capacitance, the switch voltage does not rise instantaneously during switching but has a delay. The output parasitic capacitance of the switch depends on its drain-to-source voltage, and this parasitic capacitance affects the output of the DAB converter by delaying the switch voltage. In this paper, in order to analyze the effect of the parasitic capacitance on the DAB converter output, the delay time was calculated through a formula, and this value was compared with a simulated value. In addition, the effect of the parasitic capacitance of the SiC MOSFET on the output of the DAB converter was presented by comparing the actual output voltage with the ideal output voltage and analyzing the effect of the output voltage according to the delay.

This paper presents a comparative study between the traditional phase shift (also referred to as the Single-Phase Shift (SPS)) and the Dual-Phase Shift (DPS) controlled Triple Active Bridge (TAB) converter. Being a multi-port DC-DC converter with flexible power flow control and characterized by high power density, the TAB converter is applicable in almost any situation where a DC-DC converter is needed. With the availability of multiple control schemes, this work highlights the advantages and disadvantages of the most employed control scheme used on the TAB converter, in comparison with the DPS control scheme that has so far been applied only on Dual-Active Bridge (DAB) converters. As an example, for a TAB converter with a 14 kW maximum power capacity, the work sees the comparison of the backflow power, the maximum possible current, the processed power at the different ports of the converter, the transformer voltage and current waveforms, and the Total Harmonic Distortion (THD). Based on the results obtained, we found that the DPS-controlled TAB converter was more efficient when applied to the TAB converter compared to the traditional phase shift control algorithm.