應用希爾伯特-黃轉換之受震結構安全性實驗研究

Abstract

近年來發生許多重大地震事件，使得民眾與工程人員意識到結構健康診斷之重要性，希望於災害來臨前強化結構較脆弱之部分，也希望於災害來臨時即時追蹤結構健康狀態，在結構進入不穩定狀態前發出警報，使民眾能即時做出應變措施，降低人員傷亡與經濟損失。本研究針對結構物受地震之安全狀態進行分析並提出即時預警系統，以希爾伯特-黃轉換(Hilbert-Huang Transform, HHT)對於時頻分析之強大能力，分析結構受外力振動干擾下所測得訊號資料，利用固有模態函數(Intrinsic Mode Function, IMF)的瞬時頻率(Instantaneous Frequency, IF)與能量加權得到加權瞬時頻率(Weighted Instantaneous Frequency, WIF)，接著透過定義最大最小值(Maximum and Minimum value)、信心區間(Confidence interval)、勁度比例(Stiffness ratio)三種方法界定結構健康狀態的頻率範圍，當結構頻率超出此範圍即發出警報。為了驗證所提出的三種判定方法，本研究使用中部科學工業園區管理局之縮尺結構模型進行振動台試驗，使用不同強度之地震紀錄逐步震動至破壞並紀錄其上部結構震動歷時，同時進行即時數據處理。分析結果顯示，本研究可看出結構物行為上的改變，並準確地在結構破壞前提供預警之功能。

In recent years, many catastrophic seismic events have occurred. It is of importance that real-time structural health diagnosis and alert can be done before the structure becomes unstable, so that people can react immediately to reduce casualties and economic losses. In this study, the Hilbert-Huang Transform (HHT) is used to analyze the measured signal from the superstructure under the seismic experiment and the warning alarm system is developed. Firstly, the instantaneous Frequency (IF) of Intrinsic Mode Function (IMF) weighted with energy is used to get the Weighted Instantaneous Frequency (WIF). By defining the range of Maximum and Minimum value, Confidence Interval (CI), Stiffness Ratio (SR), the structure’s health condition can be determined. In order to verify the proposed method, shaking table tests are conducted on the benchmark structure and are examined afterwards. According to the analysis of the vibration signal recorded up to the time of collapse, it is shown that the change of structure behavior can be rapidly reflected, and warning alarm before structure collapse can be accurately provided.