Multilayer Perceptron(MLP), Nonlinear Autoregressive exogenous (NARX) 모형을 이용한 계통한계가격(SMP) 예측
Forecasting System Marginal Price Using Multilayer Perceptron and Nonlinear Autoregressive exogenous model
0
조회
0
다운로드
2
인용
0
참고문헌
키워드
초록
우리나라 전력시장 발전부문에 경쟁을 도입함으로써 각발전사들이 얼마나 발전해야 할지는 중요한 문제가 되었고, 안정적인 계통한계가격(System Marginal Price, SMP)유지는 중요한 문제다. 본 연구에서는 인공신경망 중 MLP(Multi-Layer Perceptron), NARX(Nonlinear Autoregressive exogenous)를 이용하여 서로 비교 분석했고 외부변수로 전력수요를 설정하여 SMP를 예측했으며, ARIMA(Autoregressive Integrated Moving Average)와도 비교를 실시했다. 모두 30일, 40일, 50일을 예측해 봤으며 MLP의 경우 은닉층의 개수가 90개, 85개, 80개 였을 때 최적의 결과를 보여준 반면, NARX의 경우 95개, 80개, 70개 였을 때 최적의 결과를 보여주었다. 전반적으로 모든 예측에서 ARIMA, MLP, NARX 순으로 예측 오차가 작게 나타났다.
Abstract
With the introduction of competition in the power generation sector in Korea, determining how much each electricity company should develop and maintain a stable system marginal price (SMP) has become an important issue. In this study, SMP was predicted by comparing the values for each company using a multilayer perceptron (MLP) and nonlinear autoregressive exogenous (NARX) model among artificial neural networks (ANN), and setting electricity demand as an external variable. The Autoregressive Integrated Moving Average (ARIMA) was also compared. Each method predicted 30, 40, and 50 days. In the case of MLP, optimal results were obtained when the number of hidden layer units were 90, 85, and 80, whereas in the case of NARX, optimal results were obtained when the number of hidden layer units were 95, 80, and 70. Overall, the forecast errors were small in the order of ARIMA, MLP, and NARX for all predictions.
참고문헌 (14)
- 정상윤, "신경회로망을 이용한 전일 계통한계가격 예측", 석사, 전기공학, pp.2006
- Jeong, S. Y., "A hybrid neural network framework for hour-ahead system marginal price forecasting", The Transactions of the Korean Institute of Electrical Engineers (KIEE) Autumn Conference 2005, pp.162~164 (2005)
- 김대욱, "우리나라 전력시장의 용량요금제도 개선방안에 관한 연구", 에너지경제연구, 14(3), pp.293~318 (2015), 재단법인 에너지경제연구원[DOI]
- 김대용, "ARIMA 모형을 이용한 계통한계가격 예측방법론 개발", 전기학회논문지 A권, 55(2(A)), pp.85~93 (2006), 대한전기학회
- Kim, D. Y., "A day-ahead system marginal price forecasting using ARIMA model", The Transactions of the Korean Institute of Electrical Engineers (KIEE) Summer Conference 2005, pp.819~821 (2005)
- Kim, N. Y., "Study on Power Price Prediction Model Based on Artificial Neural Networks", MS, pp.2019
- A Study on the Behavior of Market Participants in the Power Wholesale Market: Game Theoretical Approach, "Korea Energy Economics Institute", KEEI, 91~, pp.KEEI Report 2003-03 (2003)
- A Study on the Stabilization of Power Market Price, "Korea Energy Economics Institute", KEEI, 23~, pp.KEEI Issue Paper 13-03 (2013)
- Major Issues and Problems of the Current Power Market Operating System and Direction of Improvement, "Korea Institute Industrial Economics & Trade", KIET, 53~68, pp.KIET Industrial Economic Review Report 2007-07 (2007)
- Lee, J. K., "A system marginal price forecasting method based on an artificial neural network using time and day information", International Federation of Automatic Control(IFAC) Journal 16th Triennial World Congress, pp.122~127 (2005)
- 이상헌, "평일 계통한계가격의 변동성 예측 – 다중주기(multi-frequency) 모형을 중심으로", 에너지경제연구, 15(2), pp.89~119 (2016), 재단법인 에너지경제연구원[DOI]
- Shin, D., "System marginal price time series data forecasting model development", The Korean Society of Mechanical Engineers, pp.3172~3176 (2018)
- 신동현, "전력 발전시장의 충격식별과 계통한계가격 변동성 분석: 전력수요 예측오차 충격을 중심으로", 응용경제, 17(2), pp.121~166 (2015), 한국응용경제학회
- Deep Learning Tutorial-Understanding Neural Network & Deep Learning, "Software Policy & Research Institute", SPRI, 71~, pp.SPRI Report 2016-007 (2017)