https://doi.org/10.5370/KIEE.2026.75.9.2017

윤승진(Seung-jin Yoon) ; 김대진(Dae-jin Kim)

As the penetration of renewable energy resources increases, output variability and short-term forecasting errors have become critical factors affecting power system operation and market settlement outcomes. In Korea, the Jeju renewable energy market pilot scheme has introduced a dual settlement framework that combines day-ahead and real-time market settlements. The imbalance penalty(IMBP) mechanism of the pilot scheme has recently been revised from a one-sided structure, which mainly penalized over-generation deviations, to a bilateral structure that also penalizes under-generation deviations under specific system and dispatch conditions. However, the quantitative effect of this transition on operating behavior and settlement revenues under different bidding conditions has not been sufficiently examined. In addition, although performance-related capacity factors have been introduced, the current settlement scheme still lacks an explicit settlement signal that directly rewards dispatch instruction compliance accuracy at each trading interval. This paper first formulates the revised bilateral IMBP mechanism and quantitatively evaluates its effect through a VPP-based mixed-integer linear programming simulation. It then proposes a dispatch instruction compliance compensation scheme that rewards resources according to their dispatch-following accuracy. The previous one-sided IMBP scheme, the revised bilateral IMBP scheme, and the revised scheme with the proposed compensation mechanism are compared under predefined over-bidding and honest-bidding scenarios. The results show that the revised bilateral IMBP partially reduces the settlement advantage under the predefined over-bidding scenario, but the reduction is limited because most of the advantage arises from the day-ahead and real-time price difference in the dual-settlement structure. The proposed compensation scheme provides an additional economic signal for accurate dispatch following and encourages ESS operation for improved compliance performance within the allowable deviation band. Since the bidding quantity is specified exogenously, these results should be interpreted as a comparison of settlement outcomes under predefined bidding scenarios rather than as an analysis of participant bidding choices.

https://doi.org/10.5370/KIEE.2026.75.9.2032

최현태(Hyuntae Choi)

This study examines changes in agricultural electricity demand during Korea’s stepwise tariff increases between 2021 and 2024. We combine monthly electricity sales by municipality and customer class for 2017?2024 with statutory tariff schedules. Using a fixed municipal sample, total agricultural electricity use is exactly decomposed into customer accounts, contracted capacity per account, and use per unit of contracted capacity. Between December 2020 and December 2024, total use increased by 1.5%. Customer accounts and contracted capacity per account increased by 10.9% and 1.2%, respectively, whereas use per unit of contracted capacity decreased by 9.6%. Comparisons with residential and general-service electricity show a similar combination of account growth and declining use intensity. These results demonstrate that limited aggregate change can conceal substantial and offsetting changes in demand composition. These findings highlight the value of distinguishing changes in the contract base, contracted capacity, and utilization when interpreting agricultural electricity demand during tariff adjustments.

https://doi.org/10.5370/KIEE.2026.75.9.2040

임한별(Han-Byul Lim) ; 송성윤(Sung-Yoon Song)

The increasing penetration of variable renewable energy reduces the reactive power reserves and short-circuit capacity of power systems, raising concerns about transient voltage recovery following faults. Voltage Source Converter High Voltage Direct Current (VSC-HVDC) systems can mitigate these issues through fast reactive power support, but their voltage recovery effect strongly depends on installation location. While bus vulnerability indicators have traditionally been used for siting reactive power compensation devices, their applicability to VSC-HVDC site selection has not been quantitatively validated. This paper examines whether six conventional vulnerability indicators?two static (voltage sensitivity, short-circuit ratio) and four dynamic (comprehensive, self-recovery, victim, source types)?can effectively identify buses with large VSC-HVDC installation effects. Exhaustive dynamic simulations were performed on 109 buses of the IEEE 118-bus test system across 8 representative N-1 contingencies, and Spearman rank correlations and top-15 ranking comparisons were used to assess identification capability. The strongest indicator, voltage sensitivity R1, exhibited only a weak correlation (), and the top-15 most effective buses showed an average R1 rank of 51. The largest installation effects were concentrated in two regions (Bus 1?13 and Bus 49?58) located in close electrical proximity to the representative fault buses, indicating that installation effect is determined primarily by the topological relationship between fault and installation buses rather than by individual bus properties alone. These results suggest that single vulnerability indicators are insufficient for VSC-HVDC site selection, and that fault scenario sets and fault?bus topological information must be jointly considered.

https://doi.org/10.5370/KIEE.2026.75.9.2049

김주홍(JooHong Kim) ; 차준상(JunSang Cha) ; 송성윤(SungYoon Song)

This paper proposes an AC sensitivity-based mixed-integer linear programming (MILP) optimization method for minimum renewable energy curtailment to relieve transmission line overloads. While DC power transfer distribution factor (PTDF) methods offer fast computation, they introduce approximation errors by neglecting voltage variation, reactive power, and line losses. AC exhaustive search guarantees accuracy but suffers from combinatorial explosion. The proposed method computes sensitivities by perturbing each candidate generator output and re-solving the AC power flow, then formulates a MILP for minimum curtailment. A Drop-One Refinement post-processing step is introduced to remove unnecessary curtailments caused by linear superposition errors. Case studies on IEEE 9-, 59-, and 118-bus standard systems and a KPG 193-bus Korean power grid model demonstrate that the proposed method achieves overload relief with computation time comparable to DC PTDF and curtailment levels comparable to AC exhaustive search, providing a practical balance between accuracy and efficiency. N-1 contingency analysis on IEEE 118 and KPG 193 systems confirms that the proposed method provides higher resolution rates than DC PTDF under N-1 conditions while maintaining real-time feasibility.

https://doi.org/10.5370/KIEE.2026.75.9.2059

박성민(Sung-Min Park)

With the rapid expansion of renewable energy in the Jeju region, managing surplus power and ensuring grid stability have become critical challenges. This paper proposes an integrated approach combining Energy Storage System (ESS) operation, HVDC reverse power flow, and Dynamic Line Rating (DLR) to address renewable energy curtailment issues. Using the IEEE 30-bus test case, a curtailment threshold analysis was conducted by integrating photovoltaic (PV) generation, and the optimal ESS capacity was derived to mitigate voltage violations and curtailment compensation costs. AC power flow analysis on a synthetic 4-bus Jeju grid verified that coordinated ESS and HVDC operation successfully stabilized bus voltage within the allowable range while enabling surplus power transmission to the mainland. Furthermore, seasonal DLR analysis demonstrated that winter conditions significantly expanded HVDC line capacity, reducing line utilization from an overloaded state to a stable level without additional infrastructure. These results confirm that the proposed integrated strategy of ESS optimization and DLR application can effectively minimize renewable energy curtailment and reduce associated compensation costs in island power systems such as Jeju.

https://doi.org/10.5370/KIEE.2026.75.9.2067

김준혁(Jun-Hyeok Kim)

This paper investigates how load variability, PV penetration, and voltage operating conditions affect voltage-limit activation and weak-bus characteristics in distribution systems. The IEEE 13-node feeder is used as an unbalanced system, considering fixed-load and time-varying residential, commercial, and industrial load scenarios under the same rated load. PV penetration levels of 0%, 50%, and 100% are defined based on the total rated active load, and the source voltage is varied from 0.95 p.u. to 1.05 p.u. Simulation results show that the dominant voltage constraint shifts from low-voltage to high-voltage as the source voltage increases, with the transition region depending on load characteristics and PV penetration. Time-varying load conditions produce different weak-bus distributions from the fixed-load case, indicating that voltage-constrained weak buses are not fixed network properties but depend on the interaction among load profiles, PV penetration, and operating conditions.

https://doi.org/10.5370/KIEE.2026.75.9.2073

김대진(Dae-Jin Kim)

In low-inertia islanded microgrids, frequency stability is critically challenged by unpredictable power imbalances arising from renewable intermittency and load fluctuations. Droop- and virtual synchronous machine (VSM)-based grid-forming (GFM) controls inherently leave a steady-state frequency deviation, since the deviation itself serves as the power-sharing signal of primary control. This paper proposes a feed-forward disturbance-compensated GFM frequency control, in which a robust disturbance observer (DOB) estimates the aggregated supply?demand imbalance from the measured frequency deviation, and an offline model predictive frequency controller (MPFC) injects the compensating power into the active power reference, restoring the nominal frequency proactively without modifying the virtual inertia or damping of the underlying VSM. The design model embeds the diesel governor dynamics so that the observer estimates only the residual imbalance, structurally avoiding conflicts with the existing diesel controls, and an observability analysis justifies the reduced fourth-order design model. Both gains are obtained from Lyapunov-based linear matrix inequality optimization under polytopic parameter uncertainties and are computed entirely offline, so that the real-time computation reduces to a few matrix?vector products at a 1 ms control period. Simulations on a 300 kVA islanded microgrid show that, against five baseline controllers including VSM with secondary frequency restoration, the proposed method achieves disturbance-estimation convergence within about 0.3 s, reduces the frequency-nadir depth by up to 72% and the RoCoF by 32% for supply-side disturbances.

https://doi.org/10.5370/KIEE.2026.75.9.2083

이주석(Juseok Lee) ; 박천규(Cheonkyu Park) ; 김도인(Do-In Kim)

This study proposes a robust open-set classification framework for power system events using Graph Attention Networks (GAT). To address limited Phasor Measurement Unit (PMU) deployment, we implement a physics-informed graph reduction strategy for computational efficiency. Transient features are extracted via Discrete Wavelet Transform (DWT) to capture multi-resolution signatures from voltage and frequency measurements. The primary contribution is a multi-stage discrimination mechanism that overcomes the limitations of conventional closed-set models. By integrating synthetic steady-state data during training and utilizing L2-normalized latent embeddings with class-specific distance thresholds, the framework effectively identifies unseen disturbances. Simulation results on the IEEE 68-bus system demonstrate a reliable solution for situational awareness in evolving power grid environments.

https://doi.org/10.5370/KIEE.2026.75.9.2093

차정훈(Junghoon Cha) ; 유원아(Wona Ryoo) ; 이승훈(Seunghoon Lee) ; 김희민(Huimin Kim)

This paper presents the design and performance analysis of a vehicle power-coupled permanent magnet synchronous generator (PMSG) operating under variable rotational speed conditions. To satisfy the output requirements of a vehicle engine driven system, the generator was designed to achieve 20 kW or higher at 500 rpm and 30 kW or higher at 900 rpm, and a voltage 380 Vac at the generator shaft rotational speed of 2100 rpm. In the initial design stage, the loading distribution method was employed to determine the principal design parameters. Electromagnetic finite element analysis using Ansys Maxwell was subsequently conducted to evaluate the no load and load characteristics over the variable speed operating range. Based on the analysis results, a prototype was fabricated and experimentally validated from 500 rpm to 2100 rpm. The experimental results demonstrated that the proposed generator satisfied the required output power and voltage over the entire operating range. The proposed design methodology is expected to provide an effective design approach for vehicle power-coupled PMSG operating under variable speed conditions.

https://doi.org/10.5370/KIEE.2026.75.9.2107

윤예빈(Ye-Bin Yoon) ; 서현우(Hyun-Woo Hyun) ; 차명준(Myung-Jun Cha) ; 김래영(Rae-Young Kim)

This paper presents a novel common-mode(CM) choke structure that achieves impedance balancing by controlling Equivalent Parallel Capacitance(EPC) through an adjustable bobbin design. High-speed power converters generate significant high-frequency CM noise during rapid switching, requiring effective suppression for Electromagnetic Compatibility(EMC) and reliability. While conventional CM filters can attenuate CM noise, their high-frequency performance is strongly affected by the EPC of the inductors and tolerances of Y-capacitors. To address this limitation, a bobbin structure is proposed that enables precise adjustment of the inter-winding distance, thereby allowing control of the EPC without additional passive components. The proposed structure satisfies the Wheatstone bridge-based impedance balancing condition. The validation is performed using a 3.7 kW resonant inverter system, demonstrating up to 9.3 dBμV of CM noise reduction in the 1 MHz to 5 MHz frequency band.

https://doi.org/10.5370/KIEE.2026.75.9.2116

이재혁(Jae-Hyuk Lee) ; 김민혁(Min-Hyeok Kim) ; 이성준(Sung-Jun Lee) ; 이재아(Jae-A Lee) ; 김종훈(Jong-Hoon Kim)

This study proposes a unified methodology to consistently link electrochemical characteristics observed by electrochemical impedance spectroscopy (EIS) under normal and abnormal conditions to operating-condition voltage?current responses and the corresponding estimated impedance and distribution of relaxation times (DRT) features. Using current and voltage signals acquired under normal operation, an electrical equivalent circuit model (EECM) is developed in a Simulink environment, and its internal parameters are identified via recursive least squares (RLS). Parameter change rates derived from comparisons of EIS data measured under normal and abnormal conditions are then applied to the RLS-estimated parameters to simulate time-domain voltage responses representative of abnormal scenarios. Short-time Fourier transform (STFT) is performed on the simulated voltage?current signals to extract frequency components, and passive electrochemical impedance spectroscopy (PEIS)-based impedance is computed as the voltage-to-current ratio at identical frequency components. The resulting impedance is compared in Nyquist form to assess condition-dependent impedance-pattern differences. Furthermore, the PEIS results are transformed into the DRT domain to re-express impedance responses in the time-constant space and to quantitatively compare reaction characteristics under abnormal conditions relative to the normal reference. The proposed procedure bridges conventional EIS analysis and signal-based diagnostics applicable to real operating environments, providing a foundational framework for designing abnormal-state diagnostic indicators and conducting scenario-based evaluations.

https://doi.org/10.5370/KIEE.2026.75.9.2129

박건후(Geonhu Park) ; 이주연(Juyeon Lee) ; 이준석(June-Seok Lee)

The Highly Efficient and Reliable Inverter Concept(HERIC) inverter is a transformerless topology with high power conversion efficiency and power density in photovoltaic(PV) system. The HERIC inverter reduces leakage current by electrically decoupling the DC and AC sides during the freewheeling period. The modulation scheme applied to the HERIC inverter affects its operating range, loss characteristics, and DC-link current characteristics. In this paper, modulation schemes for the HERIC inverter are reviewed in terms of leakage current characteristics and operating modes. The conduction and switching losses characteristics of the modulation schemes are then analyzed. Based on this analysis, an appropriate modulation scheme is determined with respect to high efficiency operation. In addition, the DC-link capacitor current characteristics are investigated to reduce the thermal stress imposed on the capacitor. Based on this analysis, a ripple current reduction method is proposed for a two-stage single-phase photovoltaic (PV) system to improve system reliability and power density. Consequently, this study contributes to achieving high power density operation in two-stage single-phase PV systems. The validity of the analytical results and the proposed method is verified through experiments.

https://doi.org/10.5370/KIEE.2026.75.9.2139

성연서(Yeon-Seo Sung) ; 이윤지(Yun-Ji Lee) ; 김수길(Soo-Gil Kim) ; 홍선기(Sun-Ki Hong)

With growing interest in renewable energy utilization, artificial intelligence techniques have been widely applied to photovoltaic power forecasting. However, most previous studies have focused on fixed PV systems and do not reflect the characteristics of PV panels mounted on moving vehicles. This paper proposes a power generation forecasting method for vehicle-mounted PV systems using an autoregressive Long Short-Term Memory model. Driving route information was collected through a navigation API, while weather data were obtained from the Korea Meteorological Administration API. The proposed method predicts solar power generation according to vehicle routes and operating conditions. Simulations using electric bus operation data from Jeju Island demonstrate the feasibility of improving mobile PV power prediction and vehicle energy management efficiency.

https://doi.org/10.5370/KIEE.2026.75.9.2146

김경범(Kyungbeom Kim) ; 서민혁(Minhyuck Seo) ; 김종민(Jong-Min Kim) ; 신훈규(Hoon-kyu Shin) ; 이남석(Nam-Suk Lee)

This study investigated the structural and crystallographic properties and the local strain accommodation behavior by growing silicon carbide (SiC) epitaxial layers with a thickness of 12 μm on a SiC wafer. Because epitaxy involves the substrate and the epitaxial layer sharing the same crystal structure and lattice constant, the occurrence of defects due to lattice mismatch was significantly reduced compared to heterojunctions. The SiC epitaxial layers formed under appropriate growth conditions exhibited excellent surface morphology and the same crystal structure as the substrate, and the residual stress generated near the interface could be alleviated through local strain accommodation. Therefore, by comprehensively applying various analytical techniques?atomic force microscopy, scanning electron microscopy, energy dispersive X-ray spectroscopy, and high-resolution transmission electron microscopy (HR-TEM)?the SiC homo-epitaxial structure and local strain accommodation behavior of the epitaxial layer grown on the SiC substrate were structurally analyzed. Specifically, since studies directly analyzing atomic-scale local strain accommodation and the fine displacement behavior of Si atomic columns at interfaces where homo-epitaxial structures are formed are relatively limited, the structural stability of the epitaxial layer was confirmed using high-angle annular dark-field?scanning transmission electron microscopy images and fast Fourier transform pattern analysis.

https://doi.org/10.5370/KIEE.2026.75.9.2151

주혜린(Hyerin Ju) ; 박수민(Sumin Park) ; 정준기(Junki Jung) ; 석오균(Ogyun Seok)

A breakdown voltage of 9.2 kV with an improved JTE dose window was achieved in a 6.5 kV SiC PiN diode with a p-type epitaxial layer using an optimized edge termination. Since the etching profile and charge balance critically affect the electric field distribution in high voltage structures, the effects of single and double etched structures were first investigated, followed by the optimization of SZ-JTE and RA-JTE. To further improve the edge termination characteristics, a Hybrid-JTE incorporating MFZ-JTE was developed based on the optimized RA-JTE structure. The optimized Hybrid-JTE achieved a breakdown voltage of 9.2 kV and exhibited enhanced tolerance to JTE dose variation.

https://doi.org/10.5370/KIEE.2026.75.9.2158

김령은(Ryeong-Eun Kim) ; 윤주은(Ju-Eun Yun) ; 차호영(Ho-Young Cha) ; 김현섭(Hyun-Seop Kim)

In this work, Al0.85Ga0.15N/Al0.5Ga0.5N high-electron-mobility transistors (HEMTs) with TiN Schottky gates and source field plates were fabricated and evaluated for high-temperature operation and off-state voltage blocking. The devices maintained gate-controlled operation and an on/off ratio above 104 at 400 °C. Despite a negative threshold voltage shift from ?2.9 V at room temperature to ?3.7 V at 400 °C, the on-resistance increased by only 10% at the same gate-overdrive voltage, owing to the compensating effects of reduced channel electron mobility and decreased access resistance. The off-state breakdown voltage increased with increasing LGD and the application of a source field plate. A device with LGD = 5 μm and LFP = 2 μm exhibited a breakdown voltage of approximately 1000 V, whereas the maximum breakdown voltage of 1800 V was obtained with LGD = 20 μm and LFP = 2 μm. TCAD simulations showed that the source field plate reduced the electric-field peak at the drain-side gate edge and redistributed the field toward the field plate edge. These results demonstrate stable high-temperature operation and kilovolt-class voltage blocking for high-power applications.

https://doi.org/10.5370/KIEE.2026.75.9.2166

도건진(Geonjin Do) ; 김상엽(Sangyeob Kim) ; 김정태(Jeongtae Kim) ; 강규혁(Gyuhyeok Kang) ; 신원영(Wonyoung Shin) ; 석오균(Ogyun Seok)

This study proposes a low-concentration double-epitaxial structure for a 650-V planar SiC MOSFET to mitigate heavy-ion-induced electric-field crowding near the P-well/drift junction. TCAD simulations were conducted by varying the doping concentration and thickness of the upper epitaxial layer. The static characteristics varied with the upper-epi doping concentration and thickness, showing corresponding changes in breakdown voltage and specific on-resistance. Before the simulated heavy-ion strike, the proposed structure reduced the peak electric field near the P-well/drift junction by 9.6% compared with the conventional single-epi structure. Increasing the upper-epi thickness from 1 to 2 μm reduced the peak electric field by 8.9%. Under a simulated heavy-ion strike, the proposed structure reduced the peak electric field near the P-well/drift junction by 6.2% compared with the conventional single-epi structure. Furthermore, the electric field was distributed over a wider drift region, indicating reduced local electric-field crowding during the transient response following a simulated heavy-ion strike.

https://doi.org/10.5370/KIEE.2026.75.9.2172

김세영(Se-Yeong Kim) ; 안태풍(Tae-Pung An) ; 허재호(Jae-Ho Hur) ; 김영배(Yong Bae Kim) ; 이방욱(Bang-Wook Lee)

This paper presents an analysis of the floating potential characteristics of conductors that can occur during the inspection and replacement of distribution-class load break switches (LBSs) following equipment failure or aging. When an LBS is in the open state, the source or load-side conductors may remain in a floating state, thus causing the formation of a floating potential due to residual charges or applied voltage. This potential formation is closely associated with the structural characteristics of vacuum interrupters and must be critically addressed from an operational safety perspective. Recently, structures that incorporate disconnecting switches (DSs) have been introduced to improve this configuration. However, because the DS can serve as a floating conductor in the open state, a comprehensive analysis of the potential characteristics across configurations that include floating conductors is required. In this paper, a horizontal spherical electrode test jig was constructed to experimentally analyze variations in the floating potential according to changes in the gap distance between the electrodes and sphere diameter. The results of this study provide fundamental data for the insulation design of distribution-class power equipment containing floating conductors.

https://doi.org/10.5370/KIEE.2026.75.9.2179

박재준(Jae-Jun Park)

In this study, epoxy resin samples with four types of tree needle electrode curvature radii (1, 5, 40, and 100 μm) were fabricated. To investigate the electrical degradation process, tree image signals were captured using a microscope under an applied voltage of 18kV at 60Hz from tree initiation through progression to final breakdown. Concurrently, partial discharge (PD) occurrences and the corresponding acoustic emission (AE) pulse waveforms were measured using AEwin software. A larger radius of curvature (r) indicates a blunter tree needle tip, which theoretically decreases the maximum electric field intensity at the tip. Experimental results using the fabricated samples confirmed that as the radius of curvature increased (1 540100μm), the time to tree breakdown significantly increased. Furthermore, a larger radius of curvature prolonged the initial tree inception time, whereas a smaller, sharper radius shortened it. The AE signal data, generated by the propagation of AE pulses from the PD source, was analyzed. By utilizing the AE waveform amplitude and pulse counts, distinctive signal features were extracted for each degradation stage: tree initiation, progression, pre-breakdown, and breakdown. These features effectively enable the characterization and analysis of the insulation degradation state and its underlying mechanisms.

https://doi.org/10.5370/KIEE.2026.75.9.2192

홍창현(Chang Hyeon Hong) ; 홍동석(Dong-Suk Hong) ; 정재철(Jaecheol Jung) ; 권구영(Gu-Young Kwon) ; 이춘권(Chun-Kwon Lee) ; 장승진(Seung Jin Chang)

This study investigates whether variations in dielectric loss inside a gas-insulated cable termination can be detected through temperature changes at the lower copper tube and PVC band. A three-dimensional model of a 154 kV, 2500 mm² XLPE cable termination was developed for the grid connection of a submarine cable system. Conductor heating was applied equally to all cases, while dielectric heat generation in the XLPE insulation and stress cone was calculated using the electric field, relative permittivity, and loss tangent. Three cases with different dielectric properties were evaluated. Electric field distributions were obtained using ANSYS Maxwell 3D, and the calculated volumetric heat sources were applied to a transient thermal model. Under identical load and ambient conditions, Case 3 showed temperature increases of approximately 1.0 °C at the lower copper tube and 1.4 °C at the PVC band compared with Case 1. These results provide a physical basis for detecting changes in internal insulation conditions through continuous external temperature monitoring.

https://doi.org/10.5370/KIEE.2026.75.9.2199

조성진(Sungjin Cho) ; 김대석(Daesuk Kim) ; 조중석(Jungseok Cho)

This paper presents an analytic solution for the nozzle-extension thickness of a 3-DOF nozzle. Since measuring nozzle-extension thickness is difficult due to high temperature and pressure, we propose a scheme of estimation for nozzle-extension thickness by using 3-DOF nozzle motion. We establish a kinematic model of 3-DOF nozzle motion. The kinematic model is analytically solved from joint and length constraints of a 3-DOF nozzle driven by two actuators with fixed-pivot points. Moreover, we combine a nozzle-extension thickness model and the kinematic model in order to obtain an analytic solution of nozzle-extension thickness. The proposed scheme of nozzle-extension thickness is verified by numerical simulations.

https://doi.org/10.5370/KIEE.2026.75.9.2206

부준석(Junseok Boo) ; 전지훈(Ji Hun Jeon) ; 안유선(Yusun Ahn) ; 이강복(Kang Bok Lee)

This paper proposes a neural network-based adaptive integral sliding mode control method for bidirectional vehicle platooning under unknown dynamic parameters and matched and mismatched disturbances. The unknown nonlinear dynamics induced by disturbances and parameter uncertainties are approximated and compensated using a neural network, while the uncertain vehicle mass directly affecting the control input is estimated through an adaptive law. To improve transient performance and robustness, a modified constant time-headway policy and a chattering-suppression robust control term are further incorporated. Theoretical analysis proves that the proposed controller guarantees asymptotic spacing error stability and string stability of the overall platoon system. Numerical simulations verify the effectiveness of the proposed method under disturbances and dynamic parameter uncertainties.

https://doi.org/10.5370/KIEE.2026.75.9.2216

김나현(Na-Hyeon Kim) ; 김희서(Hee-Seo Kim) ; 이규아(Kyu-A Lee) ; 홍채령(Chae-Ryoung Hong) ; 정경용(Kyoung-Yong Chung)

This paper proposes a hybrid Dual-Path architecture for acoustic anomaly detection in poultry environments, integrating an EfficientNet- B0-based CNN multi-label classifier with an unsupervised contrastive autoencoder (CAE). The CNN classifier performs multi-label classification of three abnormal vocalization types?distress, respiratory disease, and stress?while the CAE suppresses false alarms via reconstruction error, with only partial sensitivity to unseen anomalies. A seven-stage preprocessing pipeline was applied to improve the effective energy ratio from 32.6% to 58.3%. Trained on 70,001 samples, the system achieves a Composite Score of 0.8417, Hamming Accuracy of 0.9459, and Healthy Rejection Rate of 0.8154, with F1-Scores of 0.9969 and 0.9436 for distress and respiratory categories, respectively.

https://doi.org/10.5370/KIEE.2026.75.9.2223

남의석(Eui-Seok Nahm)

This study proposes an artificial intelligence-based framework for reducing energy consumption in wastewater treatment plants while maintaining effluent quality. Three years of hourly data were used, including influent flow and quality, process operating variables, blower power consumption, and effluent organic matter, ammonium nitrogen, total nitrogen, and total phosphorus. After preprocessing, the data were chronologically divided into training, validation, and test sets at 70%, 15%, and 15%, respectively. An uncertainty-weighted multi-task TCN-LSTM model was developed to simultaneously predict blower power consumption and multiple effluent quality variables. The TCN extracted short- and medium-term temporal patterns, while the LSTM learned long-term dependencies and process delays. The trained model was then used as a surrogate model for constrained economic model predictive control. The dissolved oxygen setpoint, airflow rate, internal recycle ratio, and external recycle ratio were optimized under effluent quality, operating-range, and rate-of-change constraints. The NRMSE-based fit scores were 96.5%, 95.3%, and 95.2% for the training, validation, and test datasets, respectively. were 96.5%, 95.3%, and 95.2% for the training, validation, and test datasets, respectively. The proposed control strategy reduced relative energy consumption from 100% to 79.8%, indicating an estimated energy-saving potential of 20.2%, while satisfying all effluent quality constraints. These results indicate that the proposed framework can effectively support energy-efficient and stable wastewater treatment operation.

https://doi.org/10.5370/KIEE.2026.75.9.2236

박찬묵(Chan-Muk Park) ; 윤민호(Min-Ho Yoon) ; 조유정(Yoo jung Cho) ; 임성훈(Sung-Hun Lim)

When a DC series arc fault occurs, it generates high-temperature plasma?often exceeding several thousand degrees Celsius?which can destroy insulation materials and cause electrical fires. To address this risk, this study proposes an arc fault detection method based on the frequency band energy characteristics of the current signal. The current signal is acquired at a sampling rate of 100 kS/s and transformed into the frequency domain using the Fast Fourier Transform (FFT). The 5?50 kHz range is divided into six sub-bands to extract relative frequency band energy ratios. These ratios are then subjected to pattern analysis by mapping the energy order across bands to all possible permutations (6! = 720). A sliding window approach is used to compute the temporal occurrence of dominant patterns. Experimental results demonstrate that changes in the distribution and frequency of these patterns between normal and arc states serve as effective indicators for DC series arc conditions. The proposed pattern analysis method enables precise arc fault detection, and it can serve as a robust feature extraction technique for machine learning models or be integrated into real-time arc monitoring systems in various electrical environments.

https://doi.org/10.5370/KIEE.2026.75.9.2246

이현구(Hyun-Goo Lee)

This paper comparatively investigates the DC-link pre-charging and voltage formation characteristics of 8200-series and 8500-series electric locomotives using field data acquired under actual operating conditions. DC-link voltage and contactor operating signals were simultaneously analyzed to identify the relationship between the pre-charging sequence and voltage rise characteristics. The results show that the 8200-series locomotive exhibits a relatively gradual DC-link voltage rise, whereas the 8500-series locomotive shows a rapid change in voltage slope near the main contactor closing point. These differences may be associated not only with the operating characteristics of the contactors but also with the parameters and operating sequence of the pre-charging circuit, including pre-charge resistance and DC-link capacitance. Since the charging current was not directly measured in this study, the magnitude of inrush current and its quantitative effects on component lifetime were not evaluated. The results provide a comparative characterization of the DC-link pre-charging and voltage formation behavior of the two locomotive series and can serve as a basis for further studies involving simultaneous voltage and current measurements.

https://doi.org/10.5370/KIEE.2026.75.9.2254

박철민(Chulmin Park) ; 이기원(Kiwon Lee) ; 손상진(Sangjin Son) ; 김용석(Yongseok Kim)

Insulators mounted on rigid overhead conductors (T-bar) in underground urban-railway tunnels are continuously exposed to metallic particulate matter from the wheel?rail?brake interface and arc discharge at the catenary?collector interface, forming a conductive contamination layer that can degrade insulation performance. Domestic operators currently manage this contamination by periodic manual cleaning, which is slow, hazardous to workers, and limited in effectiveness. This study developed an unmanned automated cleaning system for such insulators. It comprises position detection and tracking that follows the conductor's stagger and height, active uplift-force control that maintains a constant uplift force at the cleaning head, a high-pressure water-jet cleaning module, a compressed-air drying module, a wastewater recovery and purification unit, and remote monitoring and control. In field demonstrations on an underground revenue line of Seoul Metro, the mean accumulated-dust mass per unit length on the insulator surface decreased from 11.7 g/m to 0.527 g/m, an average removal rate of 95.5% verified by an accredited testing institution. Compressed-air drying followed by natural drying reached about 99% dryness within two hours, and the cleaning distance reached 5.1 km per hour. By combining stagger tracking with active uplift-force control, the system enables unmanned removal of the conductive contamination layer and safer maintenance.

https://doi.org/10.5370/KIEE.2026.75.9.2263

이기원(Kiwon Lee) ; 김용석(Yongseok Kim) ; 박철민(Chulmin Park) ; 박영(Young Park)

Overhead contact system (OCS), which supply traction power to trains through sliding contact with pantograph, is core fixed facilities in electric railway systems, and the level of their design, construction, and verification directly determines the safety of train operation and current collection performance. In Korea, the design, construction, and performance verification of OCS are based on the Railway Construction Rules, the Railway Design Standards, the Korean Railway Standards (KRS), and various specifications managed by the Korea National Railway. Many of these domestic standards are known to correspond substantially to international and European standards established by the International Electrotechnical Commission (IEC) and the European Committee for Electrotechnical Standardization (CENELEC); however, a systematic review of such correspondence has yet to be conducted, and recent revisions and corrigenda to European standards over the past several years have not been promptly reflected in domestic standards. This study comprehensively examines (1) the overall classification system and interrelationships among major international and European standards related to electric railway OCS, (2) the core technical content of each standard, and (3) recent revision trends, with particular focus on organizing recently revised content to identify considerations for future incorporation into domestic standards. By providing a comprehensive overview of the international and European standard framework while identifying the latest revisions that domestic standards urgently need to review and reflect, this study offers practical foundational data to support future amendment of domestic standards and enhancement of international conformity. This review reflects standard editions confirmed as of August 2026 and additionally provides item-by-item comparison tables and defining formulas for selected interoperability and dynamic-behaviour criteria.

https://doi.org/10.5370/KIEE.2026.75.9.2274

채은경(Eunkyung Chae) ; 오용국(Yongkuk Oh) ; 류준형(Joonhyoung Ryu)

With the global push toward carbon neutrality, the development of hydrogen-fueled rail vehicles is actively progressing to replace aging diesel trains. While hydrogen trains offer zero-emission operation and long-range capabilities, they incorporate high-risk technologies such as high-pressure hydrogen storage systems, fuel cell power systems, and high-voltage batteries. Hydrogen exhibits extremely low ignition energy and high flammability. Combined with railway-specific operating environments?such as tunnels, underground sections, and continuous vibration?rigorous safety verification is essential for various accident scenarios, including storage tank ruptures, valve failures, overpressure during refueling, and battery thermal runaway. However, due to the lack of operational experience and accident data, applying conventional railway practices or safety standards from other industries presents clear limitations. Therefore, this study proposes a methodology for hazard analysis and safety requirements derivation tailored to hydrogen rail vehicles by integrating the railway RAMS standard (IEC 62278) risk management process with hydrogen-related international standards (ISO, IEC). Potential hazards were identified by conducting a Hazard and Operability (HAZOP) analysis based on system definition, operating modes, functional analysis, and relevant standards review. Mitigation measures were established for each identified hazard to derive finalized safety requirements. Furthermore, by establishing verification plans for the derived safety requirements, this study provides technical safety assessment criteria applicable to the demonstration and commercialization stages of hydrogen rail vehicles.

https://doi.org/10.5370/KIEE.2026.75.9.2282

박종영(Jong-young Park) ; 홍수민(Sumin Hong) ; 허재행(Jae-Haeng Heo) ; 정호성(Hosung Jung)

Previous studies have demonstrated the technical effectiveness of artificial intelligence (AI)-based energy management systems (EMS) for railway stations using an artificial neural network (ANN) and a Deep Q-Network (DQN) to optimize energy operation while utilizing photovoltaic (PV) and regenerative braking energy. This study evaluates the techno-economic feasibility of the AI-based EMS by considering capital expenditure (CAPEX), operating expenditure (OPEX), and annual economic benefits. Economic performance is assessed using the payback period, net present value (NPV), internal rate of return (IRR), benefit-cost ratio (B/C), and sensitivity analyses of ESS price, electricity-price escalation, and regenerative energy utilization. The results indicate that the proposed EMS is economically feasible and provides a practical basis for evaluating AI-based railway station energy management systems.

https://doi.org/10.5370/KIEE.2026.75.9.2289

송민섭(Min-Sup Song) ; 김재원(Jaewon Kim) ; 최동호(Dongho Choi) ; 김형철(Hyungchul Kim) ; 정호성(Hosung Jung)

The phase-leg stacked PWM inverter generates multiphase voltages from a single stacked leg using a segmented dc link and ac-side dc-blocking capacitors. This paper compares six two-level PWM methods: SPWM, SVPWM, 60°&30° DPWM, +30°&?30° DPWM, DPWMMAX, and DPWMMIN. They are classified as continuous, asymmetric-clamping, and symmetric-clamping PWM and evaluated for current harmonics, switching frequency, semiconductor loss, dc-link ripple, and balancing dynamics. Results from a 450-V three-stack three-phase prototype show that continuous PWM gives lower current distortion and better balancing at low modulation indices. DPWM reduces the average switching frequency by about one-third and improves light- and medium-load efficiency. DPWMMAX and DPWMMIN offer the best compromise between loss reduction and stable dc-link regulation without clamp-type alternation. PWM selection guidelines are presented according to modulation index, load power factor, and operating objective.

https://doi.org/10.5370/KIEE.2026.75.9.2303

이기원(Kiwon Lee) ; 김용석(Yongseok Kim) ; 박철민(Chulmin Park) ; 노경수(Kyeongsu Noh) ; 이형진(Hyeongjin Lee)

In electric railway catenary systems, the span refers to the distance between two adjacent support structures, and span length directly affects current-collection performance; catenary design must therefore specify a standard span appropriate to the operating speed. Maximizing span length is a key means of reducing the number of support poles, foundations, and cantilever assemblies, thereby directly lowering construction and maintenance costs. However, the maximum span length must be determined in accordance with the Remaining Working Width (Mw ≥ 0) criterion of UIC 606-1 OR, which is jointly governed by various catenary design parameters such as stagger, wind speed, and curve radius. This study first performs a one-dimensional sensitivity analysis, followed by a two-dimensional sensitivity analysis with derivation of safe design bands, based on the catenary design conditions of the Honam High-Speed Railway, in order to identify which of these design parameters has the primary influence on the calculation of maximum span length. The results show that the rear stagger is the most dominant parameter directly controllable by the designer, and on this basis a stagger optimization algorithm is proposed. Finally, the algorithm is validated by applying it to a general-section catenary case on the Honam High-Speed Railway under Zone 2 (Jeongeup-Gwangju), Site 2 (exposed site), and curve radius R ≥ 5,000 m. The algorithm was implemented in Python and cross-validated against the calculation program used in the original Honam High-Speed Railway detailed design; relative to the current stagger standard of -50 to -200 mm, radius-specific optimized stagger values ranging from -140 to -60 mm were obtained, securing an additional 0.5 - 4.5 m of span length across the curve radii examined. The resulting maximum span is defined strictly as the value permitted under the lateral geometric and wind-load safety condition (Mw ≥ 0) of UIC 606-1 OR. All optimized spans remain within the 65 m span limit adopted in the Honam High-Speed Railway detailed design and fall within the span range (40-65 m) presented by the dynamic-characteristics simulation results performed during the detailed design for the same catenary system.

https://doi.org/10.5370/KIEE.2026.75.9.2313

이지원(Ji-Won Lee) ; 조윤진(Yun-Jin Cho) ; 문원식(Won-Sik Moon)

This paper applies a Bayesian optimization-based parameter calibration procedure to a Type 4 wind turbine generator generic model to reproduce responses that comply with the Korean grid code performance verification criteria. A time-weighted squared-error objective function was formulated as the evaluation criterion for parameter calibration. After setting the calibration target parameters, a sequential calibration procedure was applied considering the control structure of the model. The proposed procedure was applied to a case system, and the calibrated model showed output responses closer to the reference responses than the initial model.

https://doi.org/10.5370/KIEE.2026.75.9.2320

홍성준(Seong-Joon Hong) ; 손진근(Jin-Geun Shon)

This paper investigates partial discharge(PD) classification in eco-friendly gas-insulated power equipment using g³. Four representative PD defects were simulated, and nine time-domain, frequency-domain, and statistical parameters were extracted from UHF signals. Artificial neural network(ANN), support vector machine(SVM), and k-nearest neighbor(KNN) algorithms were evaluated using 1,000 PD pulses with five-fold cross-validation. The ANN showed the highest average accuracy of 92.7%, followed by the SVM and KNN. These results demonstrate the feasibility of machine learning-based PD defect classification in eco-friendly gas-insulated power equipment.

https://doi.org/10.5370/KIEE.2026.75.9.2328

안한웅(Hanwoong Ahn)

This paper presents dynamic modeling and stability analysis of a hybrid magnetic bearing (HMB) for a solar array drive assembly (SADA). A non-contact HMB using permanent-magnet bias and electromagnet control is considered to avoid friction, wear, and contamination in harsh environments. From an equivalent magnetic circuit, the magnetic force is derived and linearized at the nominal operating point using current and position stiffness. The results show that the HMB is open-loop unstable due to negative magnetic stiffness. A displacement-feedback stabilization condition is derived, and the closed-loop behavior is expressed with normalized control-authority and cross-coupling parameters. Simulations verify the instability, feedback stabilization, and tolerance bound for decoupled per-axis control.

https://doi.org/10.5370/KIEE.2026.75.9.2335

손석금(Seok-Geum Son)

This paper focuses on the development of a highly secure power cable for hair dryers to ensure user safety under varied operating conditions. In hand-held appliances like hair dryers, power line disconnection represents the most cause of product failure. The study identifies the cause of cable disconnection and identifies the arc generation mechanism. Based on these findings, two practical safety conditions?the allowable bending radius and the maximum allowable current?were derived. To solve the issue, we developed parts with improved power line safety and evaluated their safety performance. The result provides practical safety conditions for hair dryer operation and contributes to improving the overall safety standard of power cables for home appliances.

https://doi.org/10.5370/KIEE.2026.75.9.2342

임준수(Joonsu Im) ; 김진성(Jin Sung Kim) ; 강모세(Moses Kang) ; 김청훈(Chunghun Kim)

This paper proposes a photovoltaic-battery energy storage system (PV?BESS) coordinated control strategy for voltage regulation in a weak low-voltage distribution network under topology-changing fault conditions. The proposed method is formulated as a multi-objective model predictive control (MPC) problem that simultaneously determines the active power command of the BESS and the output command of PV. The objective function considers voltage regulation, battery degradation cost, price-based energy transaction cost, PV curtailment cost, and control smoothing. To reflect the voltage response of the target node under network topology changes, a Generalized LinDistFlow (GLDF)-based voltage sensitivity model is incorporated into the voltage prediction model. The control performance is evaluated through a CYME-Python co-simulation platform using a two-stage load-flow procedure at each simulation step. Three operating cases are compared: BESS-only control with the original BESS specification, PV?BESS coordinated control with the original BESS specification, and PV?BESS coordinated control with half BESS capacity. Simulation results show that the proposed coordinated control reduces voltage variations during fault conditions by using PV output control as an additional voltage regulation resource. In particular, the coordinated strategy maintains voltage regulation performance even when the BESS capacity is reduced, indicating that PV output adjustment can partially compensate for limited BESS capacity. These results demonstrate that PV?BESS coordinated control can improve voltage stability under fault conditions and may also be considered as a factor in BESS capacity planning in low-voltage distribution networks.