| ³í¹®¸í |
Áý´Ü¿¡³ÊÁö Àǹ«»ç¿ë °í½ÃÁö¿ª ³» µ¥ÀÌÅͼ¾ÅÍÀÇ ÇÕ¸®Àû ³Ã°¢¿¿ø ½Ã½ºÅÛ ¼±Á¤¿¡ °üÇÑ ¿¬±¸ / Optimization of Cooling System Selection for Data Centers within Mandatory District Energy Supply Zones |
| ÀúÀÚ¸í |
º¯»ó¿ì(Sangwoo Byeon) ; ¹Ú¹è±Õ(Baekyun Park) ; Á¶Áø±Õ(Jinkyun Cho) |
| ¼ö·Ï»çÇ× |
¼³ºñ°øÇÐ³í¹®Áý, Vol.37 No.12 (2025-12) |
| ÆäÀÌÁö |
½ÃÀÛÆäÀÌÁö(616) ÃÑÆäÀÌÁö(10) |
| ÁÖÁ¦¾î |
Èí¼ö½Ä ³Ãµ¿±â; Ãʱâ ÅõÀÚºñ; µ¥ÀÌÅͼ¾ÅÍ; Áö¿ª¿¿ø; Àü±â½Ä ³Ãµ¿±â; ¿î¿µºñ ; Absorption chiller; Capital expenditure; Data center; District heating and cooling; Electrical chiller; Operating expenditure |
| ¿ä¾à1 |
AI¿Í HPC È®»êÀ¸·Î µ¥ÀÌÅͼ¾ÅÍ Àü·Â ¼ö¿ä°¡ ±ÞÁõÇϰí ÀÖÀ¸¸ç, µ¥ÀÌÅͼ¾ÅÍÀÇ À§Ä¡¿¡ µû¸¥ Áý´Ü¿¡³ÊÁö Àǹ«Á¤Ã¥°úÀÇ Ãæµ¹ÀÌ ÁÖ¿ä ÀïÁ¡À¸·Î ºÎ»óÇϰí ÀÖ´Ù. ÀÌ¿¡ º» ¿¬±¸´Â 40 MW±Þ µ¥ÀÌÅͼ¾Å͸¦ ´ë»óÀ¸·Î Àü±â½Ä ³Ãµ¿±â¿Í Èí¼ö½Ä ³Ãµ¿±â¸¦ µ¿ÀÏÇÑ ITºÎÇÏ¿Í ³Ã°¢ ¿ë·® Á¶°Ç¿¡¼ °æÁ¦¼º ¹× ±â¼úÀû ÀûÇÕ¼º Ãø¸é¿¡¼ ÃʱâÅõÀÚºñ, ¿¬°£ ¿î¿µºñ, ½Ã½ºÅÛ ½Å·Ú¼º, ¹Ì·¡ ±â¼ú ȣȯ¼ºÀ» ºñ±³?ºÐ¼®ÇÏ¿´´Ù. ÁÖ¿ä °á·ÐÀº ´ÙÀ½°ú °°´Ù. - Á߿¼ö Èí¼ö½Ä ³Ãµ¿±â ½Ã½ºÅÛÀº Àü±â½Ä ³Ãµ¿±â ½Ã½ºÅÛ ´ëºñ ÃʱâÅõÀÚºñ ¾à 232%, ¿î¿µºñ ¾à 539% Áõ°¡·Î ºÒ¸®ÇÔÀ» º¸¿´´Ù. ³·Àº COP·Î ÀÎÇÑ ³Ã°¢Å¾, ÆßÇÁ ¿ë·® Áõ°¡´Â °ø°£°ú ÀÎÇÁ¶ó ºÎ´ãÀ» °¡Áß½ÃÄ×À¸¸ç, Á߿¼ö ¿ä±Ý°ú »ó¼öµµ ºñ¿ëÀº Àü·Âºñ Àý°¨À» »óȸÇÏ¿´´Ù. - Á߿¼ö Èí¼ö½Ä ³Ãµ¿±â ½Ã½ºÅÛÀº ¿ÜºÎ ¿¿ø ÀÇÁ¸¼º°ú °ø±Þ ºÒ¾ÈÁ¤¼ºÀ¸·Î ÀÎÇØ 100% ¹é¾÷½Ã½ºÅÛÀÌ ÇʼöÀûÀ̸ç, ¹è°ü ³ëÈÄÈ ¹× À¯·® º¯µ¿ µîÀº ¹«Áß´Ü ¿î¿µÀÌ ¿ä±¸µÇ´Â µ¥ÀÌÅͼ¾ÅÍ È¯°æ¿¡¼ Ä¡¸íÀûÀÎ ¾àÁ¡À¸·Î ÀÛ¿ëÇÑ´Ù. - ³Ã°¢ ±â¼ú ¹ßÀü°úÀÇ È£È¯¼º Ãø¸é¿¡¼ AI µ¥ÀÌÅͼ¾ÅÍÀÇ °í¹Ðµµ ¹ß¿ ÇØ¼Ò¸¦ À§ÇÑ ¼ö³Ã½Ä ³Ã°¢ ±â¼ú È®»ê¿¡ µû¶ó 40¡É ÀÌ»ó °í¿Â ³Ã¼ö Ȱ¿ëÀÌ Áõ°¡Çϰí ÀÖÀ¸³ª, Á߿¼ö Èí¼ö½Ä ³Ãµ¿±â´Â ÀÌ·¯ÇÑ ±â¼úÀû ÆÐ·¯´ÙÀÓ º¯È¿¡ ´ëÀÀÇϱ⠾î·Á¿î ±¸Á¶Àû ÇѰ踦 °¡Áø´Ù. Á¾ÇÕÀûÀ¸·Î, º» ¿¬±¸´Â 40 MW±Þ µ¥ÀÌÅͼ¾ÅÍÀÇ »ç·Ê ºÐ¼®À» ÅëÇÏ¿© µ¥ÀÌÅͼ¾ÅÍÀÇ ³Ã°¢ ½Ã½ºÅÛÀ¸·Î´Â Àü±â½Ä ³Ãµ¿±â¸¦ ±â¹ÝÀ¸·Î ÇÏ´Â ±¸¼ºÀÌ À¯Áö°ü¸®ºñ¿Í ÃʱâÅõÀÚºñ¿ë Ãø¸é¿¡¼ ÇÕ¸®ÀûÀÎ ´ë¾ÈÀÎ °ÍÀ¸·Î È®ÀÎÇÏ¿´´Ù. º» ¿¬±¸ °á°ú´Â Áý´Ü¿¡³ÊÁö »ç¿ë °í½ÃÁö¿ª ³» ÇâÈÄ ´ë±Ô¸ð µ¥ÀÌÅͼ¾ÅÍ ³Ã°¢ ½Ã½ºÅÛ ¼³°è¿Í ¿¡³ÊÁö Á¤Ã¥ ¼ö¸³ °úÁ¤¿¡¼ Âü°í ÀÚ·á·Î Ȱ¿ëµÉ ¼ö ÀÖÀ» °ÍÀ¸·Î ÆÇ´ÜµÇ¸ç, Áö¿ª³¹æ Àǹ«»ç¿ë Á¤Ã¥°ú ½ÇÁ¦ ¿î¿µ Çö½Ç °£ÀÇ ¹®Á¦Á¡À» ¼±Á¦ÀûÀ¸·Î °ËÅäÇÏ¿© ¿ÏÈÇÏ´Â µ¥ ÀÏÁ¤ ºÎºÐ ±â¿©ÇÒ ¼ö ÀÖÀ» °ÍÀ¸·Î »ç·áµÈ´Ù. |
| ¿ä¾à2 |
The rapid growth of artificial intelligence (AI) and high-performance computing (HPC) has led to a sharp increase in electricity demand at large-scale data centers. In Korea, recent extensions of mandatory district heating and cooling (DHC) policies for these facilities have raised questions about the best cooling system options. This study evaluates and compares an electrical chiller system (ALT-1) and a medium-temperature absorption chiller system (ALT-2) for a 40 MW data center in Incheon. The analysis examines capital expenditure (CAPEX), annual operating expenditure (OPEX), system reliability, and compatibility with emerging cooling technologies. The results indicate that ALT-2, with a coefficient of performance (COP) of 0.736, requires significantly larger auxiliary equipment than ALT-1, which has a COP of 8.600. This difference results in a 232% higher initial investment for ALT-2. Additionally, the annual OPEX for ALT-2 is 539% greater due to the high cost of medium-temperature hot water. Its reliance on external thermal sources, the need for full backup capacity, and incompatibility with high-temperature liquid cooling further limit its applicability. Therefore, electrical chillers are the more cost-effective and reliable choice for large-scale data centers. These findings offer practical insights for future system design and policy development in regions with mandatory DHC supply. |