A STUDY OF MID-TO LONG-TERM DISTRIBUTION PLANNING BASED ON PV INSTALLATION FORECASTING

Recently, distributed generations connected to the power distribution system is rapidly increasing due to issue of carbon neutral. For this reason, uncertainty and complexity are increasing in distribution system planning. However, it is difficult to plan the distribution system using the existing method with the distributed generations such as PVs are rapidly connected in Korea. Due to the continuous demand for PV connection, the construction of distribution lines and power substations is required, and it is difficult to plan the size and timing of them. Therefore, the capacity of distribution system should be evaluated through mid- to long-term load and distributed generation forecasting. This paper proposed the PV installation forecasting method using machine learning. It is described in the optimization algorithm that plans the substations and feeders for distribution system so that the cost of the distribution system is minimized by applying it to KEPCO's distribution planning system. In addition, this paper presented the development of KEPCO's power distribution planning system using PV installation forecasting scenario and the result of testing on a KEPCO's branch office.