Sun-Ki Hong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sun-Ki Hong has a total of 24 co-authors. Here are the top co-authors by number of shared publications:
- Ji-Hoon Han
Department of System and Control Engineering, Hoseo University, Asan, Korea
4 shared publications Ji-Hoon Han
Department of System and Control Engineering, Hoseo University
3 shared publicationsJong-Hoon Park
Hoseo University, Asan, South Korea
3 shared publications- Jong-Hoon Park
Department of System and Control Engineering, Hoseo University, Asan, Korea
3 shared publications Seung-Min Song
Department of System and Control Engineering, Hoseo University, Asan, Korea
3 shared publications- Ji-Hoon Han
Hoseo University, Asan, South Korea
2 shared publications
Latest Publications
- Open Access
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Industry Applications
- Vol. 7, 338-350
Study on Efficiency Map Prediction of Synchronous Motors Using Tandem Variational Autoencoder-Based Surrogate Model
To reduce the computational cost of finite element analysis for rapid optimal motor design, surrogate modeling techniques have been widely studied. While conventional surrogate models mainly predict scalar performance metrics, propulsion motors must be evaluated over a wide operating range, requiring the prediction of high-dimensional outputs such as efficiency maps. Owing to the large dimensional gap between design variables and efficiency maps, conventional neural networks often suffer from limited accuracy and unstable convergence. This paper proposes a Tandem Variational Autoencoder (TD-VAE) that directly predicts high-dimensional efficiency maps from motor design variables. The model employs a tandem architecture in which a variational autoencoder compresses and reconstructs efficiency maps, while an auxiliary autoencoder reconstructs the design variables to enforce latent consistency. This structure enables stable training and allows previously collected simulation data to be used without modification. The TD-VAE is further extended to predict physically interpretable loss components, including copper and iron losses. The model performance is evaluated using mean absolute percentage error (MAPE) and the coefficient of determination (R²), and predictive uncertainty is analyzed through pixel-wise confidence and prediction intervals. When applied to an interior permanent magnet synchronous motor, the proposed TD-VAE achieves a maximum MAPE of 0.18% compared with finite element analysis, demonstrating accurate prediction, stable convergence, and significantly reduced design iteration time.
- 16 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Open Journal of Industry Applications
- Vol. 7, 338-350
Open AccessStudy on Efficiency Map Prediction of Synchronous Motors Using Tandem Variational Autoencoder-Based Surrogate Model
To reduce the computational cost of finite element analysis for rapid optimal motor design, surrogate modeling techniques have been widely studied. While conventional surrogate models mainly predict scalar performance metrics, propulsion motors must be evaluated over a wide operating range, requiring the prediction of high-dimensional outputs such as efficiency maps. Owing to the large dimensional gap between design variables and efficiency maps, conventional neural networks often suffer from limited accuracy and unstable convergence. This paper proposes a Tandem Variational Autoencoder (TD-VAE) that directly predicts high-dimensional efficiency maps from motor design variables. The model employs a tandem architecture in which a variational autoencoder compresses and reconstructs efficiency maps, while an auxiliary autoencoder reconstructs the design variables to enforce latent consistency. This structure enables stable training and allows previously collected simulation data to be used without modification. The TD-VAE is further extended to predict physically interpretable loss components, including copper and iron losses. The model performance is evaluated using mean absolute percentage error (MAPE) and the coefficient of determination (R²), and predictive uncertainty is analyzed through pixel-wise confidence and prediction intervals. When applied to an interior permanent magnet synchronous motor, the proposed TD-VAE achieves a maximum MAPE of 0.18% compared with finite element analysis, demonstrating accurate prediction, stable convergence, and significantly reduced design iteration time.
- 13 January 2026
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 21 (2) , 1633-1640
Study of Reducing Analysis Time on Minimizing Detent Force in Linear Motors Using Surrogate Model
- 13 January 2026
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 21 (2) , 1633-1640
Study of Reducing Analysis Time on Minimizing Detent Force in Linear Motors Using Surrogate Model
- 8 January 2026
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 21 (2) , 1793-1803
Efficiency and Noise Characteristics of Over-Excitation in a Hysteresis Motor Using an On-Time Control Method
- 8 January 2026
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 21 (2) , 1793-1803
Efficiency and Noise Characteristics of Over-Excitation in a Hysteresis Motor Using an On-Time Control Method
- Open Access
- 1 November 2025
- Published by Institute of Electrical Engineers of Japan (IEE Japan) in IEEJ Journal of Industry Applications
- Vol. 14 (6) , 796-800
Evaluation of Feasibility for Optimal Motor Design using Deep Q-Network
The use of the reinforcement learning algorithm DQN(Deep Q-Network) can increase the design variables and offers the advantage of enabling more versatile motor optimization design. This paper evaluates the potential and advantages of reinforcement learning-based motor optimization design, which is still in its early stages globally. To demonstrate the performance of the proposed method, the geometry and drive conditions for reducing torque ripple of a FWSM(Field-Wound Synchronous Motor) using DQN were optimized. The proposed method exhibited excellent optimization performance, and it is expected that a more versatile optimization model can be developed by increasing the number of analysis cases.
- 1 November 2025
- Published by Institute of Electrical Engineers of Japan (IEE Japan) in IEEJ Journal of Industry Applications
- Vol. 14 (6) , 796-800
Open AccessEvaluation of Feasibility for Optimal Motor Design using Deep Q-Network
The use of the reinforcement learning algorithm DQN(Deep Q-Network) can increase the design variables and offers the advantage of enabling more versatile motor optimization design. This paper evaluates the potential and advantages of reinforcement learning-based motor optimization design, which is still in its early stages globally. To demonstrate the performance of the proposed method, the geometry and drive conditions for reducing torque ripple of a FWSM(Field-Wound Synchronous Motor) using DQN were optimized. The proposed method exhibited excellent optimization performance, and it is expected that a more versatile optimization model can be developed by increasing the number of analysis cases.
- 15 April 2025
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 20 (5) , 3335-3341
Enhancing Learning Efficiency in Physics Informed Neural Network through Data Comparison and Transfer Learning
- 15 April 2025
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 20 (5) , 3335-3341
Enhancing Learning Efficiency in Physics Informed Neural Network through Data Comparison and Transfer Learning
- 26 March 2025
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 20 (4) , 2509-2517
Compensation of Phase Errors in Current Sensors Induced by Eddy Currents Using a Deep Learning-Based Surrogate Model
- 26 March 2025
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 20 (4) , 2509-2517
Compensation of Phase Errors in Current Sensors Induced by Eddy Currents Using a Deep Learning-Based Surrogate Model
- Open Access
- 23 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 62730-62737
Analysis and Compensation of Phase Shift Errors of an Open-Loop Current Transducer Considering Eddy Current
The open-loop current transducer comprises a magnetic core made of soft magnet material and includes a Hall sensor. Traction motors in applications such as electric vehicles tend to be driven at high speeds. It pushes the inverters to increase the fundamental frequency of the output current above 1 kHz. However, this high frequency current generates eddy currents in the magnetic core of the current transducers. They deteriorate the measurement data thereby resulting the amplitude error and phase shift error, which is applied open-loop current transducer with the lamination magnetic core. In this study, the errors of an open-loop current transducer due to eddy currents were analyzed through mathematical modeling and finite element method (FEM). Furthermore, the impact of the error on permanent magnet synchronous motor current control was studied. The phase shift error of the current transducer used in this study reaches 3.6 degree when the frequency of current was 1.5 kHz, resulting in about a 20% torque reduction during field-weakening operation at 200 kW, 22500 rpm. On the other hand, the amplitude error of the current transducer is only 0.18% in the same condition. In order to correct the error, an online compensation method is proposed based on FEM simulation results. The phase shift error was verified through the experiments with a $300~\mu H$ , 3-phase inductive load. The proposed method was also validated through the experimental results conducted for variable frequencies from 500 Hz to 1.5 kHz.
- 23 April 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 12, 62730-62737
Open AccessAnalysis and Compensation of Phase Shift Errors of an Open-Loop Current Transducer Considering Eddy Current
The open-loop current transducer comprises a magnetic core made of soft magnet material and includes a Hall sensor. Traction motors in applications such as electric vehicles tend to be driven at high speeds. It pushes the inverters to increase the fundamental frequency of the output current above 1 kHz. However, this high frequency current generates eddy currents in the magnetic core of the current transducers. They deteriorate the measurement data thereby resulting the amplitude error and phase shift error, which is applied open-loop current transducer with the lamination magnetic core. In this study, the errors of an open-loop current transducer due to eddy currents were analyzed through mathematical modeling and finite element method (FEM). Furthermore, the impact of the error on permanent magnet synchronous motor current control was studied. The phase shift error of the current transducer used in this study reaches 3.6 degree when the frequency of current was 1.5 kHz, resulting in about a 20% torque reduction during field-weakening operation at 200 kW, 22500 rpm. On the other hand, the amplitude error of the current transducer is only 0.18% in the same condition. In order to correct the error, an online compensation method is proposed based on FEM simulation results. The phase shift error was verified through the experiments with a $300~\mu H$ , 3-phase inductive load. The proposed method was also validated through the experimental results conducted for variable frequencies from 500 Hz to 1.5 kHz.
- 18 August 2022
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 17 (6) , 1-7
A Study on the Effectiveness of Current Data in Motor Mechanical Fault Diagnosis Using XAI
- 18 August 2022
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 17 (6) , 1-7
A Study on the Effectiveness of Current Data in Motor Mechanical Fault Diagnosis Using XAI
- 28 October 2021
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 17 (2) , 1371-1378
A Current Transducer with Digital Output for Motor Drive
- 28 October 2021
- Published by Springer Nature in Journal of Electrical Engineering & Technology
- Vol. 17 (2) , 1371-1378
A Current Transducer with Digital Output for Motor Drive
- 30 November 2020
- Published by Institute of Control, Robotics and Systems in Journal of Institute of Control, Robotics and Systems
- Vol. 26 (11) , 932-939
DT-CNN based Motor Failure Prediction Considering Outlier Data
One of the major problems with the existing motor failure prediction system is to assume that all motors with the same fault condition have the same or a similar signal. This is a problem that arises because it is impossible to measure all the countless types of motors and data of driving conditions and failures. It is difficult to implement a general-purpose failure prediction system with an existing system having limited data and limited output. Data that have a large difference because they do not exist in the existing system are called outlier data. In previous studies, the problem arising from the outlier data has not been considered. To solve this problem, a system designed by separating the failure diagnosis model and the failure prediction model is proposed. The diagnostic model of the proposed system can detect data that are not inside big data using a decision-tree convolution neural network (DT-CNN). By using the diagnostic model and the predictive model in series, it is possible to analyze data in a non-measured state more efficiently. Additionally, a method for averaging the outputs of the diagnostic and predictive models is proposed. Through this, the deep learning algorithm can obtain in effect of applying the filter. Furthermore, the average values can be used to confirm the long-term signal change trend. The proposed system improves the problems of the existing failure prediction and enables more practical failure prediction.
- 30 November 2020
- Published by Institute of Control, Robotics and Systems in Journal of Institute of Control, Robotics and Systems
- Vol. 26 (11) , 932-939
DT-CNN based Motor Failure Prediction Considering Outlier Data
One of the major problems with the existing motor failure prediction system is to assume that all motors with the same fault condition have the same or a similar signal. This is a problem that arises because it is impossible to measure all the countless types of motors and data of driving conditions and failures. It is difficult to implement a general-purpose failure prediction system with an existing system having limited data and limited output. Data that have a large difference because they do not exist in the existing system are called outlier data. In previous studies, the problem arising from the outlier data has not been considered. To solve this problem, a system designed by separating the failure diagnosis model and the failure prediction model is proposed. The diagnostic model of the proposed system can detect data that are not inside big data using a decision-tree convolution neural network (DT-CNN). By using the diagnostic model and the predictive model in series, it is possible to analyze data in a non-measured state more efficiently. Additionally, a method for averaging the outputs of the diagnostic and predictive models is proposed. Through this, the deep learning algorithm can obtain in effect of applying the filter. Furthermore, the average values can be used to confirm the long-term signal change trend. The proposed system improves the problems of the existing failure prediction and enables more practical failure prediction.