센싱데이터와 기상기후 빅데이터를 융합한 태양광발전량의 회귀모델 예측분석 (The Prediction of Photovoltaic Power Using Regression Models Based on Weather Big-data and Sensing Data)

서지정보

· 발행기관 : 대한전기학회
· 수록지 정보 : 전기학회논문지 / 68권 / 12호 / 1662 ~ 1668페이지
· 저자명 : 박소연, 방준호, 유인호, 김태형

초록

In this paper, a model for predicting photovoltaic power based on collecting weather big-data and data from photovoltaic power plants is proposed with linear regression models of machining technique. The temperature, humidity, illumination, and fine dust data of photovoltaic power plants were collected and the values fused with the weather big-data were utilized as the regression model learning data. The three regression models of LR, SVR and DNN were compared and the results and accuracy of the error function were predicted by applying photovoltaic power data to each model. When using the DNN model, it was confirmed that it would have the highest accuracy from the data for predicting photovoltaic power generation. Using the designed DNN model, photovoltaic power can be predicted in any area, and accuracy can be improved according to the seasonal climate and standards of the area and the quality of comparative data.

영어초록

In this paper, a model for predicting photovoltaic power based on collecting weather big-data and data from photovoltaic power plants is proposed with linear regression models of machining technique. The temperature, humidity, illumination, and fine dust data of photovoltaic power plants were collected and the values fused with the weather big-data were utilized as the regression model learning data. The three regression models of LR, SVR and DNN were compared and the results and accuracy of the error function were predicted by applying photovoltaic power data to each model. When using the DNN model, it was confirmed that it would have the highest accuracy from the data for predicting photovoltaic power generation. Using the designed DNN model, photovoltaic power can be predicted in any area, and accuracy can be improved according to the seasonal climate and standards of the area and the quality of comparative data.

참고자료

· 없음