³í¹®¸í |
ÆÐ½Ãºê ¹× ¾×Ƽºê ¿ä¼Ò ±â¼úÀÇ µ¿Àû ±³È£ ÀÛ¿ëÀ» °í·ÁÇÑ ±×¸° ¸®¸ðµ¨¸µ ÀÇ»ç °áÁ¤ / Green Remodeling Decision-Making Considering Dynamic Interactions of Passive and Active Retrofit Variables |
ÀúÀÚ¸í |
Çã¼±¿µ(Heo, Seon-Young) ; ¶ó¼±Áß(Ra, Seon-Jung) ; ±è¿µ¼·(Kin, Young-Sub) ; ÃßÇÑ°æ(Chu, Han-Gyeong) ; ¹Úö¼ö(Park, Cheol-Soo) |
¼ö·Ï»çÇ× |
´ëÇÑ°ÇÃàÇÐȸ³í¹®Áý, Vol.39 No.8 (2023-08) |
ÆäÀÌÁö |
½ÃÀÛÆäÀÌÁö(187) ÃÑÆäÀÌÁö(8) |
ÁÖÁ¦¾î |
¹Î°¨µµ ºÐ¼®; SHAP ¹æ¹ý; ¿ä¼Ò±â¼ú; ±³È£ ÀÛ¿ë; ¼º´É Áß½É; ÀÇ»ç °áÁ¤ ; Sensitivity analysis; SHAP method; Retrofit variables; Interaction; Performance-based; decision making |
¿ä¾à1 |
°Ç¹°ÀÇ ¿¡³ÊÁö ¼º´ÉÀº ¿ä¼Ò ±â¼ú °£ ±³È£ ÀÛ¿ë¿¡ µû¶ó °³º° ¿ä¼Ò ±â¼úÀÇ ¼º´ÉÀÌ °Ç¹°¿¡ ¹ÌÄ¡´Â ¿µÇâ·ÂÀÌ ´Þ¶óÁú ¼ö ÀÖ´Ù. ¶ÇÇÑ, ¿ä¼Ò ±â¼ú °£ ±³È£ ÀÛ¿ëÀº °Ç¹°¿¡ ³»ÀçµÈ ºÒÈ®½Ç¼º (³»ºÎ ¹ß¿, ħ±â, ȯ±â, ³Ã¡¤³¹æ ¼³Á¤¿Âµµ)¿¡ ÀÇÇØ ¿µÇâÀ» ¹ÞÀ» ¼ö ÀÖ´Ù. µû¶ó¼, ±âÁ¸ °Ç¹°ÀÇ ¸®¸ðµ¨¸µ ÀÇ»ç °áÁ¤ ¹æ¹ýÀº °Ç¹°ÀÇ µ¿Æ¯¼º°ú ¿ä¼Ò ±â¼ú °£ ±³È£ÀÛ¿ëÀ» °í·ÁÇÑ °Ç¹°ÀÇ ¡®¼º´É Á᫐ Á¢±Ù¡¯ÀÌ ¹Ý¿µµÇ¾î¾ß ÇÑ´Ù. µû¶ó¼, º» ¿¬±¸¿¡¼´Â »ó±âÇÑ ºÒÈ®½Ç¼ºÀ» °í·ÁÇÑ ¹Î°¨µµ ºÐ¼®À» ½Ç½ÃÇÏ°í ¸®¸ðµ¨¸µ ´Ü°è¿¡¼ Àû¿ëµÇ´Â ÆÐ½Ãºê ¹× ¾×Ƽºê ¿ä¼Ò ±â¼úÀÇ ±³È£ÀÛ¿ëÀ» ºÐ¼®ÇÏ°íÀÚ ÇÑ´Ù. ´ë»ó °Ç¹°Àº µ¿Àû ¿¡³ÊÁö ½Ã¹Ä·¹ÀÌ¼Ç ¸ðµ¨ÀÎ EnergyPlusÀÇ ¼Ò±Ô¸ð »ç¹«¼Ò °Ç¹°·Î ¼±Á¤ÇÏ¿´´Ù. ±³È£ ÀÛ¿ëÀ» °í·ÁÇÑ ¹Î°¨µµ ºÐ¼®À» À§ÇØ, ¼³¸í °¡´ÉÇÑ AI ±â¹ý (Explainable AI)À¸·Î ¾Ë·ÁÁø SHAP ¹æ¹ýÀ» »ç¿ëÇÏ¿´´Ù. SHAP value Çؼ®À» ÅëÇØ °Ç¹°¿¡ Àû¿ëµÇ´Â ÆÐ½Ãºê ¹× ¾×Ƽºê ¿ä¼Ò ±â¼ú »çÀÌÀÇ ±³È£ ÀÛ¿ëÀ» ÆľÇÇÏ°í, À̸¦ ¹ÙÅÁÀ¸·Î °Ç¹° ¿¡³ÊÁö Àý°¨¿¡ È¿°úÀûÀÎ ¼º´É Á᫐ Á¢±ÙÀ» ±â¹ÝÇÑ ÀÇ»ç °áÁ¤ÀÇ Çʿ伺À» Á¦¾ÈÇÏ°íÀÚ ÇÑ´Ù. |
¿ä¾à2 |
With the urgency of global warming, the building sector is trying to curtail building energy for sustainable development. The current building codes relies on a prescriptive approach that specifies thermal properties of building components such as wall U-value, window U-value, fenestration SHGC, heat pump COP, lighting density. However, this prescriptive falls short because it does not consider the holistic energy performance of a building as well as sometimes results in biased or wrong decision making. To overcome this, we selected a target building (US DOE small office building), conducted global sensitivity analysis and analyzed interwoven interactions between retrofit variables and building energy use through the SHAP method, one of the explainable AIs. Conclusively, it was found that if the interwoven interactions were not properly taken into account, the prescriptive approach could lead to a wrong decision making. In addition, it emphasized that the performance based thinking must be employed in building energy decision making. |