| Title |
Customer Classification Method Using Customer Attribute Information to Generate the Virtual Load Profile of non-Automatic Meter Reading Customer |
| Authors |
김영일(Kim, Young-Il) ; 고종민(Ko, Jong-Min) ; 송재주(Song, Jae-Ju) ; 최훈(Choi, Hoon) |
| Keywords |
Customer classification ; Virtual load profile ; K-means ; PNN ; C5.0 |
| Abstract |
To analyze the load of distribution line, real LPs (Load Profile) of AMR (Automatic Meter Reading) customers and VLPs (Virtual Load Profile) of non-AMR customers are required. Accuracy of VLP is an important factor to improve the analysis performance. There are 2 kinds of methods to generate the VLP; one is using ALP (Average Load Profile) per each industrial code and PNN (Probability neural networks) algorithm; the other is using LSI (Load Shape Index) and C5.0 algorithm. In this paper, existing researches are studied, and new method is suggested. Each methods are compared the performance with same LP data of real high voltage customers. |