| Title |
Development of Machine Learning Data Management System for Improving Electric Distribution Asset Condition Prediction Model |
| Authors |
이해성(Hae-Sung Lee) ; 이병성(Byung-Sung Lee) |
| DOI |
https://doi.org/10.5370/KIEE.2021.70.9.1345 |
| Keywords |
Intelligence electric distribution grid; Big data; Machine learning; Distribution asset condition prediction model |
| Abstract |
The predictive information provided through the electric distribution asset failure prediction model is used as key information for transmission and distribution investment plans in consideration of new and renewable energy, distribution line capacity analysis, or distributed power location analysis. The predictive performance of the distribution asset failure prediction model is affected by the quality of the training data. Also, due to the characteristics of the highly volatile electrical distribution line, the electric distribution asset failure prediction model should maintain the predictive performance at a certain level by performing periodical machine learning on newly generated data. In this paper, we develop an automated system for machine learning data quality management in order to continuously manage machine learning data and efficiently improve electric distribution asset condition prediction model. As a result, our system make it possible to effectively shorten the development time and the cost of the electric distribution asset failure prediction model through the provision of high-quality data for training the prediction model. And the system developed in this paper will be able to improve the accuracy of electric distribution asset condition prediction model. |