Iterative Condition Monitoring and Fault Diagnosis Scheme of Electric Motor for Harsh Industrial Application

Seungdeog Choi,Pazouki, Elham,Baek, Jeihoon,Bahrami, Hamid Reza Institute of Electrical and Electronics Engineers 2015 IEEE transactions on industrial electronics Vol. No.

<P>This paper presents a robust diagnosis technique by iteratively analyzing the pattern of multiple fault signatures in a motor current signal. It is mathematically and experimentally proved that the proposed diagnosis algorithm provides highly accurate monitoring performance while minimizing both false detection and miss detection rate under high noise and nonlinear machine operating condition. These results are verified on a digital-signal-processor-based motor drive system where motor control and fault diagnosis are performed in real time.</P>

Three-phase three-level four-leg NPC converters with advanced model predictive control

Roh, Chan,Kwak, Sangshin,Choi, Seungdeog The Korean Institute of Power Electronics 2021 JOURNAL OF POWER ELECTRONICS Vol.21 No.10

This paper presents a computational reduction algorithm for applying model predictive control to a three-level four-leg converter. An optimal switching state is selected by only considering 7 voltage vectors located near the reference voltage vector, rather than using 81 voltage vectors in every sampling period, as in the conventional method. The sector, prism, and tetrahedron are sequentially selected using the position of the reference voltage vector. In addition, the seven voltage vectors selected in advance are the vectors constituting the selected tetrahedron. Thus, the proposed method reduces the computational cost and provides an improved model predictive control that does not affect the performance. The proposed method comprises an experimental setup of the proposed three-level four-leg converter to compare its performance with that of the conventional method.

Comparisons of three-phase and five-phase permanent magnet assisted synchronous reluctance motors

Bonthu, Sai Sudheer Reddy,Choi, Seungdeog,Baek, Jeihoon IET 2016 IET electric power applications Vol.10 No.5

<P>This paper presents the comparison of three-phase and five-phase permanent magnet assisted synchronous reluctance motors (PMa-SynRM) in terms of their design and performance characteristics. With higher fault tolerant capability, efficiency, and reliability, the five-phase PMa-SynRM can be a better substitute when compared with the three-phase PMa-SynRM in critical applications where safety is top priority. In this study, for a fair comparison, same design procedure using in lumped parameter modelling is followed in developing the three-phase and five-phase PMa-SynRMs. The models are optimised to minimise cost and maximise efficiency. For further detailed comparison of the three-phase and five-phase PMa-SynRMs, two-dimensional finite element analysis (FEA) is utilised. Performance characteristics such as torque pulsation, back electromotive force, flux linkage, d- and q-axis inductances versus their respective currents, cogging torque, efficiency plot, and torque-speed characteristics are intensively simulated through FEA. The optimised three-phase and five-phase PMa-SynRMs are fabricated with the same power rating (3 kW) and same volume. Experimental tests are conducted on the prototypes to validate the simulation results.</P>

Design of five-phase permanent magnet assisted synchronous reluctance motor for low output torque ripple applications

Baek, Jeihoon,Bonthu, Sai Sudheer Reddy,Choi, Seungdeog IET 2016 IET electric power applications Vol.10 No.5

<P>This paper presents the design of five-phase permanent magnet assisted synchronous reluctance motor (PMa-SynRM) for integrated starter and generator (ISG) of hybrid electric vehicle with low torque ripple. When ISG works as starter in order to operate engine, PMa-SynRM produces 3 kW power at the rated speed of 1800 rpm with reduced vibration by multi-phase structure. PMa-SynRMs are similar to interior permanent magnet motors in structure but are more economical due to reduced permanent magnets. In this study, lumped parameter model (LPM) using magnetic circuit design is used in the approach to initially design the five-phase PMa-SynRM. Numerical equations are integrated with the LPM to design the machine with its given range of design parameter values. Thousands of designs are generated by LPM, which are then converged to optimised model using differential evolution strategy. In this study, optimisation is done with maximum efficiency and minimum torque ripple as objective. The optimised 3 kW five-phase PMa-SynRM is then analysed by finite element method for fine tuning. Simulation results for back electromotive force, flux linkage, developed torque, torque ripple, cogging torque, torque speed characteristics and d and q-axis inductances variation over respective axis currents are verified by fabricated prototype.</P>

Individual Loss Reduction Technique for Each Phase in Three-Phase Voltage Source Rectifier Based on Carrier-Based Pulse-Width Modulation

Nguyen Minh-Hoang,Kwak Sangshin,Choi Seungdeog 대한전기학회 2024 Journal of Electrical Engineering & Technology Vol.19 No.3

This paper presents a carrier-based pulse-width modulation method for voltage source rectifers for individual phase loss reduction. Due to the reduction of switching frequency in the three-phase legs of the converter might signifcantly deteriorate the output performance, the proposed strategy aims at the most aged leg of the converter. The proposed individual phase loss reduction method modifes the creation of the ofset voltage and chooses an appropriate ofset voltage to generate the modulation voltage that only lowers the switching frequency and switching loss of the most aged leg. Hence, the lifetime of the most aged leg and entire converter is extended, whereas the output performance of the converter is not degraded too much. To verify the accuracy and efciency of the proposed approach in reducing the switching loss and providing satisfactory performance of the input/output waveforms, both simulation and experiment are conducted.

Indirect MPC method with improved output voltage and current waveforms for MMCs

Minh-Hoang Nguyen,Sangshin Kwak,Seungdeog Choi 전력전자학회 2022 JOURNAL OF POWER ELECTRONICS Vol.22 No.4

This paper focuses on a nonrequired weighting factor tuning task model predictive control (MPC) approach with improved steady-state performance, resulting in better harmonic distortion in output current and voltage for a single-phase modular multilevel converter (MMC) system. For the proposed indirect MPC technique, two independent control stages are designed to achieve various control objectives at once without the use of weighting factors. The first stage includes a single cost function, only related to the output current to generate a temporary quantity of inserted submodules (SMs). A circulating current regulation using predicted value is proposed to obtain the total quantity of inserted SMs regarding the circulating current suppression. The final quantity of inserted SMs is adjusted and generated from the temporary quantity to be used in a voltage sorting algorithm. Without the use of weighting factors, the control objectives in the MMC system can be achieved using the proposed indirect MPC. The output current and voltage have a considerable improvement over that of the conventional indirect MPC. The simulation and experimental outcomes validate that the proposed approach is correct and feasible.

Standard deviation based series DC arc detection method for voltage source converters

Jae-chang Kim,Sangshin Kwak,Seungdeog Choi 전력전자학회 2022 JOURNAL OF POWER ELECTRONICS Vol.22 No.7

A series DC arc detection technique using standard deviation is proposed in this paper. In the occurrence of a series DC arc, the standard deviation of the average of the frequency band of the load current increases. This standard deviation is set as the indicator for arc detection in the proposed algorithm. The average of the specific frequency band of the load current, which is in a range of 5–40 kHz, is calculated for arc detection. After that, the standard deviation of the average value is obtained and used for arc detection. Load currents collected under various load conditions were used to verify the proposed technique. By comparing the proposed method with the conventional method using frequency components, it was found that the proposed method has a lower algorithmic complexity, and is effective in both centralized frequency and spread frequency fluctuations. In addition, the consistency of the proposed arc detection technique was superior to that of the conventional technique.

Deep learning-based series AC arc detection algorithms

Park, Chang-Ju,Dang, Hoang-Long,Kwak, Sangshin,Choi, Seungdeog The Korean Institute of Power Electronics 2021 JOURNAL OF POWER ELECTRONICS Vol.21 No.10

Various studies on arc detection methods are described. Series AC arc is detected based on the characteristics extracted from arc voltage, frequency, and time domain of the current. Methods of arc detection using artificial intelligence have been studied previously. In the present study, the performance of multiple methods is analyzed by comparing different input parameters and artificial neural networks. In addition to the input parameters presented in the literature, the performance is compared and analyzed using the following parameters: zero-crossing period, frequency average, instantaneous frequency, entropy, combination of fast Fourier transform (FFT) and maximum slip difference, and combination of FFT and frequency average. These parameters and different neural networks are studied in the bounded and unbounded case, and the performance is compared. For different combinations of neural networks and input parameters, another research question is to identify the input parameters to be used if the number of training data is limited. Moreover, this study investigates the change in detection rate depending on the number of training samples. As a result, the minimum dataset size required to obtain the final detection rate is identified.

Analysis and Diagnosis Scheme of Parallel Arc Failure in DC Power Lines

Dang Hoang-Long,Kim Jae-Chang,Kwak Sangshin,Choi Seungdeog 대한전기학회 2023 Journal of Electrical Engineering & Technology Vol.18 No.3

The arc phenomena usually relate to the undesired ejection of two electric conductors. The emission power discharge from an arc event might wrath the electrical lines and cause a fire. Numerous studies were proposed to detect arc events and isolate them in time. The DC arc faults are sorted into two common types: series and parallel arcs. Due to corrupting contamination or insulation, a parallel arc occurs between two electrical wiring. The parallel arc currents of the system can be considerably amplified compared with the series type. In this research, the characteristic behaviors of the system in both time and frequency domains are studied during DC parallel arc failures, and arc energy was also discussed and analyzed. The unique behaviors are adopted to identify parallel arcs in different conditions. Sorting electrical arcs that are beneficial and trustworthy for the judicial procedure and deciding the protection schemes. The analyzed process is based on different domains of load current, source current, and arc voltage.

Evaluation of Single-Phase DC–AC Converters with Condition Monitoring Algorithm of Aluminum Electrolytic Capacitors Using Artificial Learnings with Various Circuit Signals and Filtering Combinations

Dang Hoang-Long,Park Hye-Jin,Kwak Sangshin,Choi Seungdeog 대한전기학회 2023 Journal of Electrical Engineering & Technology Vol.18 No.4

Capacitors are essential parts of power converters since the cost, size, and performance of converters are mainly dependent on them. Nevertheless, the capacitor is the most degeneration device among all converter parts owing to its aging failures and little lifetime. Thus, the monitoring process is an essential route for valuing health status and gives predictive maintenance to ensure steadiness in electric converter. The equivalent series resistance and the capacitance are commonly indexes employed for estimating the condition grade of capacitors. In this research, six artificial intelligence (AI) algorithms are adopted to estimate the aluminum capacitor (Al-Cap) parameters in the single-phase inverter system. Various circuit signals, such as load voltage and current, capacitor voltage and current, are examined by utilizing the discrete wavelet transform (DWT) analysis and the combinations of fast Fourier transform with various filters. The considered signals are handled as AI model’s inputs to guesstimate the health status of the Al-cap. In addition, the root-mean-square value is employed as an index to compare the accuracy with the analyzed signals. Furthermore, several indicators are mixed to acquire the best recipes for capacitor health evaluation.