| Title |
Scaled RMSE and Shewhart Control Chart-based Abnormal Reference Day Detection Method to Improve the Forecasting Accuracy of Community-level Power Demand |
| Authors |
김치연(Chi-Yeon Kim) ; 김채린(Chae-Rin Kim) ; 김동근(Dong-Keun Kim) ; 최형진(Hyeong-Jin Choi) ; 박시삼(Si-Sam Park) ; 조수환(Soo-Hwan Cho) |
| DOI |
https://doi.org/10.5370/KIEE.2020.69.2.245 |
| Keywords |
Community level demand forecast; Shewhart Control Chart; Abnormal reference days; Similar days |
| Abstract |
Accurate short-term power demand forecasting is a key-technique for the optimal generator operation planning and the power market pricing. Especially, the accurate demand forecasting ability in city or community level micro-grids is more essential for local power utilities to obtain the optimal solutions for the economic energy mix and the daily energy supply planning. The forecasting error can increase the cost of purchasing additional power, cause economic losses by wasting energy, and exacerbate the system reliability. In order to minimize the forecasting error in city or community microgrid, the following two factors should be considered. One is how to determine the similar days whose time patterns of daily power demand are similar to each other and the other is how to detect and remove abnormal reference days of the similar days. Since the first issue applying the scaled RMSE method has been dealt in the previous related paper, we will summarize it in chapter 2 and focus at the second issue applying the control chart technique. In this paper, an abnormal reference days detection method based on the quality control charts such as , , and ? will be proposed and the detection performances will be compared and analyzed by simulations with actual annual demand data. |