• 전기전자학회논문지
• /
• 제23권1호
• /
• pp.134-142
• /
• 2019

에너지인터넷 기술의 발전과 다양한 전자기기의 보급으로 에너지소비량이 패턴이 다양해짐에 따라 수요예측에 대한 신뢰도가 감소하고 있어 발전량 최적화 및 전력공급 안정화에 문제를 야기하고 있다. 본 연구에서는 고신뢰성을 갖는 수요예측을 위해 딥러닝 기법인 Convolution neural network(CNN)과 Bidirectional Long Short-Term Memory(BLSTM)을 융합한 1Dimention-Convolution and Bidirectional LSTM(1D-ConvBLSTM)을 제안하고, 제안한 기법을 활용하여 시계열 에너지소비량대한 소비패턴을 효과적으로 추출한다. 실험 결과에서는 다양한 반복학습 횟수와 feature map에 대해서 수요를 예측하고 적은 반복학습 횟수로도 테스트 데이터의 그래프 개형을 예측하는 것을 검증한다.

• 한국태양에너지학회 논문집
• /
• 제40권4호
• /
• pp.35-44
• /
• 2020

Currently, various policies regarding ecofriendly vehicles are being proposed to reduce carbon emissions. In this study, the required areas for charging electric vehicle (EV) batteries using electricity produced by photovoltaic (PV) power plants were estimated. First, approximately 2.4 million battery EVs, which represented 10% of the total number of vehicles, consume approximately 404 GWh. Second, the power required for charging batteries is approximately 0.3 GW, and the site area of the PV power plant is 4.62 ㎢, which accounts for 0.005% of the national territory. Third, from the available sites of buildings based on the region, Jeju alone consumes approximately 0.2%, while the rest of the region requires approximately 0.1%. Fourth, Seoul, which has the smallest available area of mountains and farmlands, utilizes 0.34% of the site for PV power plants, while the other parts of the region use less than 0.1%. The results of this study confirmed that the area of the PV power plant site for producing battery-charging power generated through the supply of EVs is very small. Therefore, it is desirable to analyze and implement more specific plans, such as efficient land use, forest damage minimization, and safe maintenance, to expand renewable energy, including PV power.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 2004년도 하계학술대회 논문집 A
• /
• pp.698-700
• /
• 2004

Electric power restructuring offers a major change to the vertically integrated utility monopoly. Deregulation has had a great impact on the electric power industry in various countries. Bidding competition is one of the main transaction approaches after deregulation. The energy trading levels between market participants is largely dependent on the short-term price forecasts. This paper presents the short-term System Marginal Price (SMP) forecasting implementation using backpropagation Neural Network in competitive electricity market. Demand and SMP that supplied from Korea Power Exchange (KPX) are used by a input data and then predict SMP. It needs to analysis the input data for accurate prediction.

• 제어로봇시스템학회:학술대회논문집
• /
• 제어로봇시스템학회 2003년도 ICCAS
• /
• pp.2283-2286
• /
• 2003

This paper presents conservation of electrical energy in building with harmonics analysis and compensation which occur in electrical system. We use load controlling and management system in order to adjust load factor of system.The maximum demand limiting and controlling are used ,then the system can acquire the prediction and compare it to the maximum demand set point.The electrical signal analysis based on FFT technique. The harmonics are compensated by using harmonic filters.This system consists computer which works as controller, processor , analysis and database unit together with digital power meter in form of multidrop network through serial communication via RS-485.The load control system uses PLC to control load via serial communication RS-485. The A/D converter is used for sampling the electrical signals via parallel port of computer.The harmonic filters are controlled by a computer.The data of measurement such as voltage, current, power, power factor, total harmonic distortion, energy, etc., can be saved as database and analysis. The load factor is adjusted by limiting and controlling maximum demand. The load factor adjustment can reduce the cost of electric consumption and energy generation together with harmonics compensation in order to increase high efficiency of electrical system.

• 대한전기학회:학술대회논문집
• /
• 대한전기학회 1999년도 추계학술대회 논문집 학회본부 A
• /
• pp.210-212
• /
• 1999

This paper proposes a novel wavelet transform based technique for prediction of System Marginal Price(SMP). In this paper, Daubechies D1(haar), D2, D4 wavelet transforms are adopted to predict SMP and the numerical results reveal that certain wavelet components can effectively be used to identify the SMP characteristics with relation to the system demand in electric power systems. The wavelet coefficients associated with certain frequency and time localisation are adjusted using the conventional multiple regression method and then reconstructed in order to predict the SMP on the next scheduling day through a five-scale synthesis technique. The outcome of the study clearly indicates that the proposed wavelet transform approach can be used as an attractive and effective means for the SMP forecasting.

• 대한전기학회논문지:전력기술부문A
• /
• 제53권2호
• /
• pp.87-96
• /
• 2004

This paper summarizes the research results of the load management for pole transformers done in 1997-1998 and 2000-2002. The purpose of the research is to enhance the accuracy of peak load estimation in pole transformers. We concentrated our effort on the acquisition of massive actual load data for modifying the load regression coefficients, which related to the peak load estimation of lamp-use customers, and adjusting the demand-factor coefficients, which used for the peak load prediction of motor-use customers. To enhance the load regression equations, the 264 load data acquisition devices are equipped to the sample pole transformers. For the modification of demand factor coefficients, the peak load currents are measured in each customer and pole transformer for 13 KEPCO (Korea Electric Power Corporation) distribution branch offices. Case studies for 50 sample pole transformers show that the proposed coefficients could reduce estimating error of the peak load for pole transformers, compared with the conventional one.

• 디지털융복합연구
• /
• 제18권12호
• /
• pp.267-278
• /
• 2020

에너지 패러다임이 격변하는 시점에서 ESS는 전력부족 및 전력수요관리의 해소와 재생에너지의 증진에 필수적인 요건이다. 이에 본 논문에서는 부하 및 태양광 발전 예측량을 통하여 비용효과적인 ESS Peak-Shaving 운영방안을 제안한다. ESS 운영방안을 위해 통계적 척도인 RMS을 통해 부하 및 태양광 발전 예측하였으며 예측된 부하 및 태양광 발전량을 통해 한 시간 단위의 목표 부하 절감량 Guide-line을 설정하였다. 대상 수용가의 1년 실데이터를 활용한 부하 및 태양광 발전 예측 시뮬레이션으로 2019년 5월 6일 ~ 10일의 부하 및 태양광 발전량을 예측 하였으며 시간별 Guide-line을 설정하였다. 부하 예측 평균오차율은 7.12%였으며, 태양광 발전량 예측 평균오차율은 10.57%를 나타냈다. ESS 운영방안을 통한 시간별 Guide-line 제시를 통해 수용가의 Peak-shaving 최대화에 기여하였음을 확인하였다. 본 논문의 결과를 통해 태양광과 연계하여 화석에너지로 발생하는 환경적인 영향을 최소화하며 신재생에너지를 최대 활용하여 에너지 문제를 줄일 수 있다고 기대한다.

• 한국조명전기설비학회:학술대회논문집
• /
• 한국조명전기설비학회 2007년도 춘계학술대회 논문집
• /
• pp.393-396
• /
• 2007

Recently one level higher intelligent switching board than established one by the application of intelligent building and digital system are being constructed. Therefore facility's high efficiency, high degree satisfaction, miniaturaization, standardization through application of communication technology and monitoring and controlling by computer system utilized by web-basis power control system and electric IT are practiced. Especially network must be constructed through unified IBS server that monitors every educational institute's switching boards in real time control system. And I intend to create methods to save energy and raise electricity quality by power demand prediction and remote-controled management and operation. In this thesis I intend to suggest measures of forming unified system through researching educational institute's ways of constructing switching board and status analysis and overcoming technical difficulties in user's side and saving and maintenance expense.

• 설비공학논문집
• /
• 제30권3호
• /
• pp.130-142
• /
• 2018

We use heat pumps with thermal storage system to reduce peak usage of electric power during winters and summers. A heat pump stores thermal energy in a thermal storage tank during the night, to meet load requirements during the day. This system stabilizes the supply and demand of electric power; moreover by utilizing the inexpensive midnight electric power, thus making it cost effective. In this study, we propose a system wherein the thermal storage tank and heat pump are modeled using the TRNSYS, whereas the control simulations are performed by (i) conventional control methods (i.e., thermal storage priority method and heat pump priority method); (ii) region control method, which operates at the optimal part load ratio of the heat pump; (iii) load response control method, which minimizes operating cost responding to load; and (iv) dynamic programming method, which runs the system by following the minimum cost path. We observed that the electricity cost using the region control method, load response control approach, and dynamic programing method was lower compared to using conventional control techniques. According to the annual simulation results, the electricity cost utilizing the load response control method is 43% and 4.4% lower than those obtained by the conventional techniques. We can note that the result related to the power cost was similar to that obtained by the dynamic programming method based on the load prediction. We can, therefore, conclude that the load response control method turned out to be more advantageous when compared to the conventional techniques regarding power consumption and electricity costs.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2008년도 춘계학술발표대회
• /
• pp.231-234
• /
• 2008

Currently an automated methodology based on data mining techniques is presented for the prediction of customer load patterns in load demand data. The main aim of our work is to forecast customers' contract information from capacity of daily power consumption patterns. According to the result, we try to evaluate the contract information's suitability. The proposed our approach consists of three stages: (i) data preprocessing: noise or outlier is detected and removed (ii) cluster analysis: SOMs clustering is used to create load patterns and the representative load profiles and (iii) classification: we applied the K-NNs classifier in order to predict the customers' contract information base on power consumption patterns. According to the our proposed methodology, power load measured from AMR(automatic meter reading) system, as well as customer indexes, were used as inputs. The output was the classification of representative load profiles (or classes). Lastly, in order to evaluate KNN classification technique, the proposed methodology was applied on a set of high voltage customers of the Korea power system and the results of our experiments was presented.