Mobile QR Code QR CODE : The Transactions of the Korean Institute of Electrical Engineers

인공신경회로망에 의한 변전소의 기반 고장유형 식별 SOP based Fault Type Identification of Substation using ANN

https://doi.org/10.5370/KIEE.2019.68.9.1039

이경민(Kyung-Min Lee) ; 박철원(Chul-Won Park)

After AlphaGo, there has been a move to apply AI advancements and self-learning to substation fault determination system to transition to an automatic fault recovery system. In order to fault recovery, fault type identification and fault location determination must be preceded. Artificial Neural Network (ANN) with smart advantage has recently been increasing interest due to the advancement of computer hardware and software platform. In this study, ANN is used to identify fault type in substation. First, we made the structure of ANN using the components of substation and the fault types of Standard Operation Procedure (SOP). Then the learning pattern was included considering the steady state and the 15 fault types specified in SOP. After learning through Back Propagation (BP), the ANN for identifying fault type of substation presented as a test pattern was tested. Finally, the proposed technique was evaluated under various simulation conditions.

시간별 기온 민감도를 이용한 하절기 평일 단기 전력수요 예측 Short-Term Load Forecasting Using Hourly Temperature Sensitivity on Summer Weekdays

https://doi.org/10.5370/KIEE.2019.68.9.1045

김정환(Jung-Hwan Kim) ; 김규한(Kyu-Han Kim) ; 이흥석(Heung-Seok Lee) ; 박준호(June Ho Park)

Load forecasting is important to determine the market price and the supply reserve. The electric load in summer is influenced by meteorological elements, especially most affected by temperature. Therefore, the temperature directly related to the cooling loads must be precisely considered to improve the accuracy of the load forecasting. In this paper, we propose the load forecasting model for 24 hours during summer weekdays based on the artificial neural network. To improve the forecasting accuracy, we classify the weekdays into two groups of Monday and Tuesday-Friday, where electric load pattern is similar within each group. Furthermore, the hourly temperature sensitivity was calculated and used as an input variable. The simulation results show that this proposed approach can be applied to forecast the electric load in summer weekdays accurately.

메타 휴리스틱 기법을 이용한 최대수요관리의 정량적 효과가 반영된 ESS 최적운영전략 An Optimal ESS Operation Strategy Considering the Quantitative Effect of Peak Management with Meta-Heuristic Optimization

https://doi.org/10.5370/KIEE.2019.68.9.1052

주형준(Hyeong-Jun Ju) ; 손진만(Jin-Man Sohn)

Electricity charges consist of into electrical energy charges and basic charges. Electricity charges are charged according to time of use(TOU), not according to the total daily usage in household sector. Therefore, using ESS for peak management can decrease electricity charges. The optimal charging / discharging schedule of ESS considering the electrical energy charges can be established by linear programming. The basic charge reduction method through minimization of deviation in the previous researches does not reflect the exact basic charges. To cope with this problem, we proposes peak demand method for an optimal ESS charge / discharge schedule that minimizes total charges. The effectiveness of the proposed method is shown through case studies using various heuristic algorithms.

1기 무한모선 시스템의 선로 고장판별을 위한 강화학습 기반 외란관측기 설계 Design of a Reinforcement Learning-Based Disturbance Observer for Line Fault Detection of a Single Machine Infinite Bus System

https://doi.org/10.5370/KIEE.2019.68.9.1060

장수영(Su Young Jang) ; 강상희(Young Ik Son) ; 손영익(Sang Hee Kang)

According to the increase of electric power demand in the modern society the power system is gradually expanding. This results in a growing need for an intelligent method of fast determination and protection against various failures in the power system. As the computer platform is improved, the system fault detection and reliable protection devices have been trying to enhance their performances using artificial intelligence techniques. If a failure occurs in the single-machine infinite bus(SMIB) system. the electrical output of the generator changes, which can be regarded as a result of an external disturbance input. This paper presents a line fault detection method by using a reinforcement learning-based disturbance observer that estimates the magnitude of the equivalent disturbance. Reinforcement learning is an algorithm that models the relationship between the behavior of an agent and the reward from environment. This paper has adopted the Deep Q-Network for training of the proposed disturbance observer. The performance of the proposed reinforcement learning-based disturbance observer is verified by computer simulations. The results show that the disturbance can be estimated successfully and the estimate can be used to detect the line fault.

인공신경망에 기반한 전력시장에서의 균형가격과 혼잡현상 해석 Artificial Neural Network based Analysis on Equilibrium Price and Transmission Congestion in Electricity Market

https://doi.org/10.5370/KIEE.2019.68.9.1067

이광호(Kwang-Ho Lee)

This paper proposes an application of artificial neural networks for analyzing electricity market that has insufficient information for calculating equilibrium. Neural networks are constructed and trained on two representative cases in the electricity market. One is for calculating equilibrium price in perfect competition market and the other is for determining whether the transmission congestion occurs. The neural network uses a multilayer structure and learns with backpropagation algorithms for training. The neural networks trained in the case studies calculate the market price with a high probability and also determines an occurrence of the transmission congestion accurately

XGBoost 기법을 이용한 단기 전력 수요 예측 및 하이퍼파라미터 변화에 따른 영향 분석 Short-term Load Forecasting Using XGBoost and the Analysis of Hyperparameters

https://doi.org/10.5370/KIEE.2019.68.9.1073

오재영(Jae-Young Oh) ; 함도현(Do-Hyeon Ham) ; 이용건(Yong-Geon Lee) ; 김기백(Gibak Kim)

Accurate load forecasting is getting vital with social and economic development to secure electricity supply and minimize redundant electricity generation. The load forecasting is also essential for efficient power system operation. As machine learning techniques become popular due to the breakthroughs in the application of intelligent systems such as speech or image recognition, variety of machine learning algorithms have also been applied to predict electricity demand. For load forecasting, this paper employs XGBoost algorithm that has recently been receiving attention. To yield the maximum performance of the XGBoost model, we performed grid search method to find optimal hyperparameters of XGBoost. The effects of the XGBoost model's hyperparameters on the model are assessed and visualized.

PMU 빅데이터를 활용한 계통고장분류 모델 개발 Development of Classification Model of Power System Fault by Using PMU Big-Data

https://doi.org/10.5370/KIEE.2019.68.9.1079

강성범(Sung-Bum Kang) ; 고백경(Baek-Kyeong Ko) ; 남수철(Su-Chul NAM) ; 최영도(Yong-Do Choi) ; 김용학(Yong-Hak Kim) ; 전동훈(Dong-Hoon Jeon)

Recently, innovative techniques in artificial intelligence such as machine learning have emerged to efficiently process huge amounts of big data delivered from PMUs to WAMS. Through processing raw data and analyzing big data, It delivers highly useful and valuable system status information to system operators. The types of machine learning vary depending on the usage, but the CNN (Convolution Neural Network) model is mainly used for the post analysis and fault detection(classification) in the power system. In this paper, based on PMU big data, we study the power system fault classification model by using CNN Model. Using Convolution neural network model based on KERAS, the database for each fault type was built and supervised learning was conducted for the model. The constructed model was verified with test data and the validity of the model was verified by inputting the actual power system fault data for the trained model. As a result, developed model classified correctly for the actual fault.

자연재난에 의한 전력설비 피해 예측을 위한 인공신경망(ANN) 알고리즘 개발 Development of Artificial Neural Network Algorithm for the Prediction of Power Failures by Natural Disaster

https://doi.org/10.5370/KIEE.2019.68.9.1085

최민희(Min-Hee Choi) ; 정남준(Nam-Joon Jung) ; 이규철(Kyu-Chul Lee) ; 정재성(Jae-Sung Jeong) ; 서인용(In-Young Seo)

Damage to the power system caused by natural disasters, including typhoons, is gradually increasing. The amount of the power outage caused by major typhoons shows 1.25 million households by “Rusa” in 2002, 1.44 million by “Maemi” in 2003, 1.68 million by “Kompasu” in 2010, 1.93 million by “Bolaven” in 2012 and 0.25 million by “Chaba” in 2016. Power companies are striving to establish an integrated system and simulators to predict power facility damage by natural disasters in advance and to establish a rapid response system in case of damage. In this paper, we developed the power facility damage prediction algorithm applied artificial neural network (ANN) for 6 kinds of natural disasters such as typhoon, strong wind, heavy rain, heavy snow, cold wave and heat wave. The algorithm consists of three phases: ① the establishment of big data by extracting meteorological data from the Automatic Weather System from 2007 to 2018, ② the analysis of the correlation between the power failures and the weather conditions(such as wind speed, rainfall, etc.) and ③ the evaluation of damage prediction algorithms using the ANN. In particular, comparisons and analyses with the Linear Regression(REG) algorithm were performed to assess the accuracy of the ANN algorithm. This algorithm was applied to Typhoon “Chaba” in 2016 to predict the failure of electric wires and Cut Out Switch (COS) in Seogwipo. The prediction error(MAE) of the ANN is 0.127, which is better than the performance of the REG.

딥러닝 모델 기반 단기 전력수요 예측 Short-Term Load Forecasting Based on Deep Learning Model

https://doi.org/10.5370/KIEE.2019.68.9.1094

김도현(Dohyun Kim) ; 조호진(Ho Jin-Jo) ; 김명수(Myung Su Kim) ; 노재형(Jae Hyung Roh) ; 박종배(Jong-Bae Park)

This paper presents a Short-Term Long-short term memory Convolutional neural network(STLC) Model that is combined with Convolutional Neural Network(CNN) and Long-Short Term Memory(LSTM). CNN model predicts load pattern using past load profile, LSTM model forecasts load variation depending on temperature and time index. STLC model’s output is hourly load data to combine two model’s outputs. The input parameters of STLC model are composed of time index, weighted weather data, past load data. Weights are calculated based on electricity consumption by main region in South Korea and reflects in the weather data. STLC model is trained with data from 2013 through 2017 and is verified with data from 2018. The STLC model forecasts 1-day hourly load data. Simulation results obtained show the comparison of actual and forecasted load data and also compare with other methods in MAPE(Mean Absolute Percentage Error) to prove accuracy of the proposed model.

과도안정도와 미소신호안정도 관점에서 국내 전력계통의 임계관성계수 평가 Estimation of Critical Inertia of Korean Electric Power Systems Based on Transient and Small-Signal Stabilities

https://doi.org/10.5370/KIEE.2019.68.9.1100

백종오(Jong-Oh Baek) ; 정인주(In-Joo Jeong) ; 하정민(Jung-Min Ha) ; 맹창엽(Chang-Yeop Maeng) ; 권오근(Oh-Geun Kwon) ; 이배근(Bae-Geun Lee) ; 김수배(Soobae Kim)

The increase of inverter-based renewable energy sources reduces inertia constant of power systems and the loss of the system inertia may cause power system stability problems. To ensure a secure energy delivery system, these problems should be analyzed thoroughly from various perspectives about power system stability. This paper presents the estimation of the critical inertia of Korean power systems, which is the minimum amount of inertia required to maintain power system dynamic securities. The estimation has been made by gradually reducing the system inertia of Korean power systems and then by analyzing the change of dynamic securities in terms of transient and small signal stabilities. In the transient stability aspect, the decrease of critical clearing time(CCT) has been analyzed with four different reduced inertial models of Korean power systems. With the small signal stability perspective, the change of the system damping ratio has been studied. Based on the results from the stability studies, the critical inertia has been determined and it can be considered as the indirect limit of renewable energy shares in Korean power systems.

지역별 기상자료를 고려한 태양광발전출력 모형 연구 PV Generation Modeling using Regional Weather Factors for Dispatch Planning

https://doi.org/10.5370/KIEE.2019.68.9.1109

김완수(Wan-Soo Kim) ; 조하현(Ha-Hyun Jo)

Korea Power Exchange has responsibility for real-time electricity system balance. Recently the phenomenon of expanding renewable generators especially PV is a big challenge for the system operator. To make robust day-ahead dispatch plan, building the forecasting model for PV is one of the most important issue in electricity industry. In this study we compared the accuracy of statistical models with the physical model’s based on advanced research. We tried to apply the models to korean electricity market and national system operarion circumstance.

Back-to-Back 컨버터를 이용한 수차 발전 시스템의 직류단 전압 변동 저감 Minimization of DC-Link Voltage Variation in a Hydraulic Turbine Generation System Using Back-to-Back Converters

https://doi.org/10.5370/KIEE.2019.68.9.1118

전성수(Sung-Soo Jeon) ; 박영수(Yeongsu Bak) ; 이교범(Kyo-Beum Lee)

This paper proposes minimization method of DC-link voltage variation in a hydraulic turbine generation system using back-to-back (BTB) converters. When the generation power is drastically changed, it causes the variation of the DC-link voltage. It deteriorates stability and efficiency of the hydraulic turbine generation system. This paper describes how the compensation current is calculated to reduce the variation of the DC-link voltage. The stability and efficiency of the hydraulic turbine generation system can be improved through the proposed compensation current. The proposed method was verified through the simulation and experimental results.

팽윤에 의한 실리콘 고무 특성 변화에 관한 연구 A Study on Variation of Properties of Silicone Rubber by Swelling

https://doi.org/10.5370/KIEE.2019.68.9.1124

최진욱(Jin-Wook Choe) ; 남석현(Seok-Hyun Nam) ; 김진규(Jin-Gyu Kim)

In order to build underground the transmission line, it is necessary to install outdoor terminations that connect overhead transmission lines and underground power cables. The inside of the outdoor termination is filled with insulating oil and a stress relief cone made of silicone rubber is assembled at the end of outer semiconducting layer. Insulating oil is heated to lower viscosity prior to injection, which requires additional equipments and more time. It is necessary to examine the influence of impregnated silicone rubber on the use of low-viscosity insulating oil that can be injected without heating. In this paper, we investigated how much silicone rubber swells depending on the viscosity of three different insulating oils. The mechanical and electrical properties such as tensile strength, elongation, volume resistivity, relative permittivity, etc., of the fully swollen silicone rubber were measured. In addition, the chemical analyzes of the fully swollen silicone rubber through FTIR and TGA were performed.

자기 에너지를 이용한 지중 SAW 온도센서 구동 및 무선 측정 시스템 개발 Driving of Underground SAW Temperature Sensor System Using Magnetic Energy and Its Wireless Measurement System Development

https://doi.org/10.5370/KIEE.2019.68.9.1129

김시혁(SiHyeok Kim) ; 이기근(Keekeun Lee)

A wireless underground sensor system was developed based on magnetic antennas and surface acoustic wave (SAW) resonator to monitor temperature variations around buried utilities. A~250 MHz magnetic antenna generates a SAW along the piezoelectric substrate, and the returned SAW energy owing to the reflection bars on the sensor is reconverted to magnetic flux by the sensor’s interdigital transducer (IDT) and subsequently transmitted to a reader via magnetic antenna. By observing changes in the center frequency of the SAW sensor with temperature, we were able to monitor the underground temperature variations in real time. Temperature sensor was fabricated on a 128o YX LiNbO3. In soil testing, a long readout distance was observed. The temperature sensors provided stable performance in terms of underground temperature changes, soil type, and soil compactness. The sensitivity and linearity for the sensor was 0.3 MHz/℃ and 0.96, respectively.

폭염기간 인공지능 센서를 이용한 아파트 정전예방 방안 Prevention of Power Outage using AI Sensors during Summer Heatwave

https://doi.org/10.5370/KIEE.2019.68.9.1137

최태일(Tae-Il Choi)

A record-breaking heatwave of long-term disasters last summer was a serious threat to public safety. Since the power equipment of an apartment is owned by the customer, there is a limit to the power company providing the service in case of power outage. In order to overcome this problem and continuously supply stable electric power to the people, we developed a remote monitoring and control system of apartment power equipment using artificial intelligence (AI) sensors. AI sensors can prevent power outages by notifying power companies and apartments managers in real time when an overload or an unbalance occurs in the transformers of an apartment power equipment. A pilot system was installed in two apartments in the Seoul area, and after evaluating the effect of the operation, the system will be expanded to other areas.

구간 2형 퍼지 시스템의 고장 허용 슬라이딩 모드 카오스 동기화 제어기 설계 Fault Tolerant Sliding Mode Chaotic Synchronization Controller Design for an Interval Type-2 Fuzzy System

https://doi.org/10.5370/KIEE.2019.68.9.1143

김한솔(Han Sol Kim) ; 주영훈(Young Hoon Joo)

This paper deals with the fault tolerant interval type-2 fuzzy integral sliding mode controller design for time-varying uncertain chaotic systems including actuator faults. The overall control system consists of driving and response chaotic systems, and the control objective is to synchronize the state variables of these systems. To this end, the chaotic system including uncertain terms is expressed by the interval type-2 fuzzy model. Also, the failure occurring in the actuator is represented by a fault matrix having time-varying unknown parameters. In this paper, the fault matrix is divided into the time-varying and time-invariant terms, which makes the sufficient condition have only time-invariant part of the fault matrix. Finally, we validate the proposed method through simulation examples.

원자층 증착 ZnO/은나노와이어를 이용한 투명 UV 광 검출기 Transparent UV Photodetector by Atomic-layered ZnO and Silver Nanowires

https://doi.org/10.5370/KIEE.2019.68.9.1151

노민수(Min-Soo Roh) ; 반동균(Dong-Kyun Ban) ; 박주연(Ju-Yeon Park) ; 김준동(Joondong Kim)

The highly transparent UV Photodetector was realized by metal oxide layers by using magnetron sputtering system and ALD system. Device is consisted of p-n junction by p-NiO and n-TiO2. In addition, transmittance reaches near by 50% of value to guarantee the optical view to the human eyes. In order to improve the performance of UV Photodetector, the ALD ZnO layer was applied between NiO and TiO2. By embedding the thin ZnO layer by ALD process, the surface of defects of TiO2 can be relieved, resulting in the significant suppression of the leakage current. The electrically conductive and optically transparent silver nanowires (AgNWs) were coated onto the top layer, working to the hole transport layer (HTL), which is definitely advantageous for improving photocurrent value. The functional uses, ZnO layer of leakage current suppression and AgNWs of photocurrent enhancement, induce the great improve of the transparent UV photodetector performance for quick photo-responses (rise time: 0.98 ms, fall time: 1.59 ms) with high responsivity. This finding of functional use of ALD ZnO and AgNWs may provide a route for high-efficient photoelectric devices, including solar cells and photodetectors

고속운행을 위한 활차식 장력조정장치 성능향상 개발 Development for a Performance Improvement of a Pulley Type Tensioning Device for a Speed-Up

https://doi.org/10.5370/KIEE.2019.68.9.1158

서동훈(Dong-Hoon Seo) ; 서기범(Ki-Bum Seo)

It is designed the tension ratio of 1 : 3 to maintain tension variation within 3% for performance improvement of pulley tensioning device for high speed operation. The plant performance test and certification test of qualified office were performed for oilless bearing with a low friction coefficient and flexible wire rope. Also, it has been performed the running test, visual inspection and certification of qualified office to secure the reliability for high speed operation for 1 year. It is proved that applied the developed tensioning device for high speed operation section.

탄소섬유 차폐케이블용 접속재의 환경열화에 따른 절연특성 Electrical Insulation Characteristics according to Environmental Degradation of a Connector for Carbon Fiber Screened Cable

https://doi.org/10.5370/KIEE.2019.68.9.1163

곽동순(Dong-Soon Kwag)

Carbon fiber, because of difficulty in maintaining its shape, is required extra care in connections, and its problem is that fiber can easily break due to the binding post-tensioning force. So, a new type of screened cable connector which can reduce carbon fiber's breaking or cutting and also minimize contact resistance in connections needs to be developed. In this paper, a conductive spring-applied, gland-type cable connector was designed and manufactured, and its applicability was reviewed through measurement of contact resistance of the connector.