Á¤º¸°úÇÐȸ ÄÄÇ»ÆÃÀÇ ½ÇÁ¦ ³í¹®Áö (KIISE Transactions on Computing Practices)
ÇѱÛÁ¦¸ñ(Korean Title) |
ºñÀüÇü ÀÜ¿© ±â»ó Á¤º¸¸¦ È°¿ëÇÑ ÇÏÀ̺긮µåÇü Àü·Â ¼ö¿ä ¿¹Ãø ¸ðµ¨ ¼³°è |
¿µ¹®Á¦¸ñ(English Title) |
Hybrid Load Forecasting Model based on Atypical Residual of Meteorological Information |
ÀúÀÚ(Author) |
¹Ú°±¸
¼ÛÁØÈ£
ȲÀǼ®
Kanggu Park
Junho Song
Euiseok Hwang
|
¿ø¹®¼ö·Ïó(Citation) |
VOL 25 NO. 01 PP. 0052 ~ 0057 (2019. 01) |
Çѱ۳»¿ë (Korean Abstract) |
Àα¸Áõ°¡¿Í ´õºÒ¾î Àü±â ±â±âÀÇ º¸±Þ È®´ë¿¡ µû¶ó ¿¡³ÊÁö ¼Òºñ°¡ ²ÙÁØÈ÷ Áõ°¡ÇÏ°í È¿À²ÀûÀÎ Àü·Â °ø±Þ ¹× °ü¸®¸¦ À§ÇÑ ½º¸¶Æ® ±×¸®µå ±â¼ú¿¡ ´ëÇÑ ¿ä±¸ ¶ÇÇÑ Áõ°¡ÇÏ°í ÀÖ´Ù. ÀÌ¿¡ µû¶ó¼, ½º¸¶Æ® ¿¡ ³ÊÁö ±â¼úÀÇ ÁÖ¿ä ¿øµ¿·ÂÀ¸·Î¼ ºÎÇÏ ¿¹Ãø ±â¼úÀÌ È°¹ßÈ÷ ¿¬±¸µÇ°í ÀÖ´Ù. º» ¿¬±¸¿¡¼´Â ÇÏ·ç Àü Àü·Â »ç ¿ë·® ¿¹Ãø ¹æ¹ý¿¡ ÃÊÁ¡À» ¸ÂÃß°í ÀÖÀ¸¸ç, ±â»óÁ¤º¸, »ç¶÷ÀÇ È°µ¿ µîÀÇ ¿µÇâÀ¸·Î º¹ÀâÇÑ ÆÐÅÏ Æ¯¼ºÀ» °®´Â Àü·Â µ¥ÀÌÅÍ¿¡ È¿°úÀûÀÎ ¿¹Ãø ±â¹ýÀÎ ¼±Çü ¿¹Ãø°ú ºñ¼±Çü ¿¹ÃøÀ» °áÇÕÇÑ ÇÏÀ̺긮µå ¿¹Ãø ¸ðµ¨À» ¼³°èÇÑ ´Ù. ¶ÇÇÑ, Áß¿äÇÑ ÀÔ·Â º¯¼ö¸¦ ã¾Æ³»±â À§Çؼ ·£´ý Æ÷·¹½ºÆ® ±â¹ÝÀ¸·Î ºñ¼±Çü ¸ðµ¨ÀÇ ´ÙÁß ÀÔ·Â °£ÀÇ ÀÇ ¹Ì ÀÖ´Â º¯¼ö¸¦ ½Äº°ÇÏ´Â ½Ã½ºÅÛ Àû¿ë ¹æ¹ýÀ» Á¦¾ÈÇÑ´Ù. ¸¶Áö¸·À¸·Î ºñÀüÇü ÀÜ¿© ±â»ó Á¤º¸¸¦ °í·ÁÇÑ ¿¹ Ãø ¸ðµ¨ÀÌ ±×·¸Áö ¾ÊÀº ¸ðµ¨°ú ºñ±³ÇÏ¿© ¿¹ÃøÀÇ ½Å·Úµµ°¡ ³ôÀ½À» º¸ÀδÙ.
|
¿µ¹®³»¿ë (English Abstract) |
The steady increase in energy consumption with the population growth and spread of electricity driven devices, has led to increasing demand for efficient power supply and management. Therefore, load prediction schemes are actively studied as key enablers of the smart energy coordination. This study proposed a day-ahead load forecasting method based on hybrid load prediction model combining the linear and nonlinear prediction approaches for forecasting electricity usage data, which is impacted by complex pattern characteristic such as meteorological information and human factor. Also, it proposed a systemic method to discriminate the meaningful variables among the multiple inputs of nonlinear models based on random forest scheme. In conclusion, this study established that the prediction model considering atypical residue of meteorological information improves prediction performance.
|
Å°¿öµå(Keyword) |
½Ã°è¿ µ¥ÀÌÅÍ
ÇÏ·ç Àü Àü·Â ¼ö¿ä ¿¹Ãø
ÀÚ±âȸ±Í¸ðµ¨
ÇÏÀ̺긮µåÇü Àü·Â ¼ö¿ä ¿¹Ãø ¸ðµ¨
ºñÀüÇü ÀÜ¿© ±â»ó Á¤º¸
Àΰø½Å°æ¸Á
time-series data
day-ahead load forecasting
autoregressive model
artificial neural network
hybrid load prediction model
atypical residual of meteorological information
|
ÆÄÀÏ÷ºÎ |
PDF ´Ù¿î·Îµå
|