°ÇÃ൵½Ã°ø°£¿¬±¸¼Ò

Architecture & Urban Research Institute

pdf¿ø¹®º¸±â ¿¡·¯ ÇØ°á¹æ¹ý ¹Ù·Î°¡±â



¹®ÇåȨ > ¿¬±¸³í¹® > »ó¼¼

[¿ø¹®º¸±â½Ã ¼ÒºñµÇ´Â Æ÷ÀÎÆ® : 100 Æ÷ÀÎÆ®] ¹Ì¸®º¸±â Àοë

´ëÇѰÇÃàÇÐȸ|³í¹®Áý 2021³â 2¿ù

³í¹®¸í º£ÀÌÁö¾È ÇÊÅ͸¦ ÀÌ¿ëÇÑ Ç³Á÷°¢¹æÇâ ÇÏÁ߸𵨠ÆÄ¶ó¹ÌÅÍ ÃßÁ¤ / Model Parameter Estimation of Across-Wind Load Based on Bayesian Approach
ÀúÀÚ¸í ȲÀç½Â(Hwang, Jae-Seung)
¹ßÇà»ç ´ëÇѰÇÃàÇÐȸ
¼ö·Ï»çÇ× ´ëÇѰÇÃàÇÐȸ³í¹®Áý, Vol.37 No.02 (2021-02)
ÆäÀÌÁö ½ÃÀÛÆäÀÌÁö(181) ÃÑÆäÀÌÁö(8)
ISSN 2733-6247
ÁÖÁ¦ºÐ·ù ±¸Á¶ / Àü»ê
ÁÖÁ¦¾î dzÁ÷°¢¹æÇâ ÇÏÁß; º£ÀÌÁö¾È ÇÊÅ͸µ; ¹«ÇâÄ®¸¸ÇÊÅÍ; ÇÏÁߺи®; ¸ðµ¨ ÆÄ¶ó¹ÌÅÍ ÃßÁ¤ ; Across-wind load; Bayesian filtering; Unscented Kalman Filtering; Load separation; Model parameter estimation
¿ä¾à1 dzÁ÷°¢¹æÇâÀÇ Ç³ÇÏÁßÀº ¹öÇÍÆÃ, ½ºÆ®·ÎÇÒ, Çǵå¹éÇÏÁß ¼ººÐÀÌ ÇÕ¼ºµÈ °ÍÀ¸·Î ¾Ë·ÁÁ® ÀÖ´Ù. Çǵå¹éÇÏÁßÀº ±¸Á¶¹°ÀÇ ¼Óµµ¿Í °ø·Â°¨¼è¸¦ ÀÌ¿ëÇÏ¿© ¿ëÀÌÇÏ°Ô ºÐ¸®ÇÒ ¼ö ÀÖÁö¸¸, ¹öÇÍÆÃ, ½ºÆ®·ÎÇÒ ÇÏÁß¼ººÐÀº °áÇÕµÈ ÇüÅ·Π³ªÅ¸³ª±â ¶§¹®¿¡ °¢ ÇÏÁß¼ººÐÀÌ Áøµ¿ÀÀ´ä¿¡ ±â¿©Çϴ Ư¼ºÀ» Á¤·®ÀûÀ¸·Î Æò°¡Çϰųª °¢±â ÇÏÁßÀÇ ¼öÇÐÀû¸ðµ¨ÀÏ ¾ò´Â °ÍÀÌ ¾î·Á¿î ½ÇÁ¤À̾ú´Ù. º» ¿¬±¸¿¡¼­´Â º£ÀÌÁö¾È ÇÊÅ͸µÀ» ÀÌ¿ëÇÏ¿© µÎ°¡Áö ÇÏÁß¼ººÐÀ» ºÐ¸®ÇÏ°í °¢ ÇÏÁ߸ðµ¨ÀÇ ÆÄ¶ó¹ÌÅÍ ÃßÁ¤ÇÏ´Â »õ·Î¿î ±â¹ýÀ» Á¦¾ÈÇÑ´Ù. º» ±â¹ýÀº Àü´ÞÇÔ¼ö ÇüÅ·Π³ªÅ¸³¾ ¼ö ÀÖ´Â ÇÏÁ߸ðµ¨ÀÇ ÆÄ¶ó¹ÌÅ͸¦ ¹«ÇâÄ®¸¸ÇÊÅ͸¦ ÀÌ¿ëÇÏ¿© Áøµ¿¼ö¿µ¿ª¿¡¼­ ÃßÁ¤Çϰí, ¼øÂ÷ÀûÀ¸·Î ÇÏÁߺи®¸¦ ¼öÇàÇÏ´Â ÇÁ·¯¼¼½º·Î ÀÌ·ç¾îÁø´Ù. °øÅº¼º½ÇÇèÀ¸·ÎºÎÅÍ ¾ò¾îÁø dzÁ÷°¢ ¹æÇâÀÇ ÀÀ´äÀ¸·ÎºÎÅÍ ¿ªÇؼ®À» ¼öÇàÇÏ¿© dzÇÏÁßÀ» µµÃâÇÏ°í º» ¿¬±¸¿¡¼­ Á¦¾ÈµÈ ±â¹ýÀ» Àû¿ëÇÑ °á°ú ¹öÇÍÆÃ, ½ºÆ®·ÎÇÒ °¢ ÇÏÁ߸ðµ¨ÀÇ ÆÄ¶ó¹ÌÅÍ ÃßÁ¤ÀÌ ¾ÈÁ¤ÀûÀ¸·Î ÀÌ·ç¾îÁö´Â °ÍÀ» ¾Ë ¼ö ÀÖ¾ú´Ù.
¿ä¾à2 It is known that the across-wind load is mainly consist of three components: the buffeting, Strouhal, and feedback load components. The
feedback load can be easily separated from the across-wind load using the structural velocity response and aerodynamic damping ratio.
However, it is difficult to quantitatively and independently evaluate the contribution of the buffeting and Strouhal load since the two load
components are strongly combined in the across-wind load. In this study, it is proposed a new approach to separate the two load components
using Bayesian filtering and to identify the parameters of each mathematical model. This technique consists of a process of estimating the
parameters of the mathematical load model, which can be expressed in the form of a transfer function, in the frequency domain using the
unscented Kalman filter, and sequentially performing load separation. For the validation of the proposed approach, the aeroelastic model test
was performed and wind load was inversely identified from the acceleration response obtained from the wind tunnel test. And then, the
proposed technique is applied on the wind load, which shows that the stable parameter estimation of the buffing and Strouhal load,
respectively, was reliably peformed.
¼ÒÀåó ´ëÇѰÇÃàÇÐȸ
¾ð¾î Çѱ¹¾î
DOI https://doi.org/10.5659/JAIK.2021.37.2.181
ºÐ¼®¼­Áö
°ÇÃ౸Á¶ > ±âŸ ¿¬±¸ºÐ¾ß > ±âŸ¿¬±¸ÁÖÁ¦