| Title |
A Study on Input Variables of Distribution Line Load Forecasting Model for Distribution Planning |
| Authors |
김준오(Jun Oh Kim) ; 조진태(Jin Tae Cho) ; 김승완(Seung Wan Kim) |
| DOI |
https://doi.org/10.5370/KIEE.2022.71.8.1092 |
| Keywords |
Distribution planning; Load forecasting model; Ensemble learning; Input variable; Correlation analysis |
| Abstract |
The importance of the power distribution planning is increasing due to expansion of distributed generation in distribution system.
The distribution planning is based on accurate load forecasting. Recently, machine learning-based load forecasting models have been developed and applied in distribution planning. However, since the performance of the machine learning-based load forecasting model depends on the combination of input variables, it is important to select input variables. This paper proposed the process of selecting input variables for an ensemble based load forecasting model. The input variable of the distribution line load forecasting model was selected through correlation analysis of predictable input variables, then the improvement of forecasting performance with selected input variable was analyzed by comparing the forecasting and the actual value. In addition, the importance of the selected input variables and the validity of the proposed process were analyzed through XAI analysis. |