強震即時警報主要目的為能在強震波來臨前，準確於當地提出警報，以避免人員的傷亡與財產的損失，同時避免因為過度預測的誤報，所造成社會成本的損失。因此，準確的分辨該次地震是否會對於人員傷亡與財產損失，為強震即時警報技術研發的首要目標。然而，實際上造成傷亡與損失的，並不是地震本身，主要為地震導致結構物破壞(例如建築物倒塌、橋梁斷裂、交通運輸脫軌、水壩潰提等，為主要的傷亡來源)、結構物附屬設施掉落或損壞(例如天花板掉落、儲存櫃倒塌砸傷人員、瓦斯管線斷裂造成火災等)、重要設備物的損傷或功能性喪失(例如重要歷史文物受損、高科技面板廠房Stocker內產品破裂、高科技晶圓廠房爐管內產品破裂、高科技廠房內Scanner當機導致無預警長期停工等)等等。考慮結構型態的不同所呈現的結構特性亦不同，且各類結構物的受震後反應皆不同，較難以用統一方式概括。本研究擬結合人工智慧與結構動力特性分析技術，發展結構受震反應快速評估技術，期能助於估計對於各類結構物、附屬設施及設備物的可能損害程度，以及所引致的人員損傷程度，更為準確的分辨來臨之地震是否可能造成傷亡與損失，期能藉由自動化控制的輔助，獲得更佳的警報效果。本研究擬採用初達P波前三秒內資訊，配合人工智慧技術建立類神經模型，並推估即將來臨之地震訊號在頻譜振幅與頻譜相位中的含量，進而推估地表地震訊號。將所得地表訊號結合目標結構物各樓層之轉換函數，即可大致推估出各樓層的最大受震加速度反應。最後，利用雙塔試驗模型、中科管理局大樓及兩棟建築物等案例探討此研究在實務上所能達到的成果。試驗結果顯示，本研究所提出之方法大致上皆可有效預測的最大樓板反應。其中，在實驗室之雙塔試驗模型，以及兩棟建築物的案例，其預測結果較為準確；相對的，在中科管理局大樓案例預測結果較為不準確，且不論是使用量測訊號所得之轉換函數，或是有限元素模型所得之轉換函數，結果均較為不準確，其原因可能為該結構系統較為複雜所導致。最後，目前本研究所探討的案例仍不足，未來應增加案例，並探討應用於各種不同形式結構物之可行性。

The main purpose of real-time early warning of strong earthquakes is to accurately issue local warnings before the arrival of strong earthquake waves to avoid casualties and property losses, and to avoid excessive prediction of social cost losses caused by false positives. Therefore, the main purpose of developing strong earthquake real-time alarm technology is to accurately distinguish whether the earthquake will cause casualties and property losses. However, the actual casualties and losses are not the earthquake, but mainly the structural damage caused by the earthquake (such as building collapse, bridge breakage, traffic derailment, dam collapse, etc., which is the main cause of casualties). Falling or damaged facilities (such as ceiling drops, lockers collapse, fires caused by gas pipe rupture, etc.), damage to important equipment or loss of function (such as the destruction of important historical sites, the rupture of high-tech panel factory Stocker products, the breakdown of high-tech wafer tubes, the collapse of high-tech factory scanners, resulting in no warnings, long-term downtime, etc. etc.). Considering different structural types and different structural features, it is difficult to summarize them after various structures after the earthquake. The purpose of this study was to use artificial intelligence and structural dynamics analysis techniques to quickly assess structural seismic responses to help estimate the extent to which various structures, auxiliary facilities and equipment may be damaged, and the extent of personal injury. A more accurate distinction between impending earthquakes will result in casualties and losses, and it is hoped that better results will be obtained through automatic control. The purpose of this study was to establish an artificial intelligence neural model using the first three seconds of the initial P-wave information, and estimate the content of the upcoming seismic signal in the spectral amplitude and spectral phase, and then estimate the ground earthquake signal. By combining the resulting surface signal with the transfer function of each layer of the target structure, the maximum seismic acceleration response of each layer can be roughly estimated. Finally, using the two-tower test model, the Central Science Park Administration Building and two buildings, the results of this research can be explored. The test results show that the method proposed in this study is usually effective for predicting the maximum floor response. Among them, the laboratory's two-tower test model and the situation of the two buildings have more accurate prediction results; on the contrary, by using the measurement signals to obtain the transmission signals, the case prediction results of the Central Science Park Administration Building are relatively inaccurate. The conversion functions obtained from the finite element model are all inaccurate, which may be due to the structural system. Finally, the cases discussed in this study are still insufficient. Cases should be added in the future and the feasibility of applying them to various forms of structures should be explored.

參考文獻：

Andrews, D. (1986). Objective determination of source parameters and similarities of
earthquakes of different size.
Convertito, V., Iervolino, I., Zollo, A., & Manfredi, G. (2008). Prediction of response spectra via real-time earthquake measurements. Soil Dynamics and Earthquake Engineering, 28(6), 492-505. doi:10.1016/j.soildyn.2007.07.006
Fujinawa, Y., & Noda, Y. (2013). Japan's Earthquake Early Warning System on 11 March 2011: Performance, Shortcomings, and Changes. Earthquake Spectra, 29, S341-S368. doi:10.1193/1.4000127
Hsu, T. Y., Lin, P. Y., Wang, H. H., Chiang, H. W., Chang, Y. W., Kuo, C. H., . . . Wen, K. L. (2018). Comparing the Performance of the NEEWS Earthquake Early Warning System Against the CWB System During the 6 February 2018 M(w)6.2 Hualien Earthquake. Geophysical Research Letters, 45(12), 6001-6007. doi:10.1029/2018gl078079
Kanda K, Kobori T, Ikeda Y, & H., K. (1994). The Development of a" Pre-arrival Transmission System for Earthquake Information" Applied to Seismic Response Controlled Structures.
Kodera, Y., Saitou, J., Hayashimoto, N., Adachi, S., Morimoto, M., Nishimae, Y., & Hoshiba, M. (2016). Earthquake early warning for the 2016 Kumamoto earthquake: performance evaluation of the current system and the next-generation methods of the Japan Meteorological Agency. Earth Planets and Space, 68. doi:ARTN 202
10.1186/s40623-016-0567-1
Kuyuk, H. S., & Motosaka, M. (2008). Spectral Forecasting of Earthquake Ground Motion Using Regional and National Earthquake Early Warning Systems for Advanced Engineering Application Against Approaching Miyagi-Ken Oki Earthquakes.
Maddaloni, G., Caterino, N., & Occhiuzzi, A. (2011). Semi-active control of the benchmark highway bridge based on seismic early warning systems. Bulletin of Earthquake Engineering, 9(5), 1703-1715. doi:10.1007/s10518-011-9259-1
Nuzzo, I., Caterino, N., Maddaloni, G., & Occhiuzzi, A. (2017). Smart hybrid isolation of a case study highway bridge exploiting seismic early warning information. Engineering Structures, 147, 134-147. doi:10.1016/j.engstruct.2017.05.057
Picozzi, M. (2012). An attempt of real-time structural response assessment by an interferometric approach: A tailor-made earthquake early warning for buildings. Soil Dynamics and Earthquake Engineering, 38, 109-118. doi:10.1016/j.soildyn.2012.02.003
Pnevmatikos, N. G., & Gantes, C. J. (2007). Pole selection for structural control using the complex Fourier characteristics of the incoming earthquake. Structural Control & Health Monitoring, 14(3), 428-447. doi:10.1002/stc.165
Pnevmatikos, N. G., Kallivokas, L. F., & Gantes, C. J. (2004). Feed-forward control of active variable stiffness systems for mitigating seismic hazard in structures. Engineering Structures, 26(4), 471-483. doi:10.1016/j.engstruct.2003.11.003
Richard M. Allen, Paolo Gasparini, Osamu Kamigaichi, & and Maren Böse. (2009). The Status of Earthquake Early Warning around the World: An Introductory Overview.
Salinas, V., Luzon, F., Garcia-Jerez, A., Sanchez-Sesma, F. J., Kawase, H., Matsushima, S., . . . Campillo, M. (2014). Using Diffuse Field Theory to Interpret the H/V Spectral Ratio from Earthquake Records in Cibeles Seismic Station, Mexico City. Bulletin of the Seismological Society of America, 104(2), 995-1001. doi:10.1785/0120130202
Satriano, C., Wu, Y. M., Zollo, A., & Kanamori, H. (2011). Earthquake early warning: Concepts, methods and physical grounds. Soil Dynamics and Earthquake Engineering, 31(2), 106-118. doi:10.1016/j.soildyn.2010.07.007
林子剛. (2016). 結構受震反應快速評估方法.
凃雅瀞. (2017). 考量震波之群延時及相位譜特性之強地動模擬. 114.