Overview of Research

In this study, a complete measure for both the vibration-sensitive equipment and the ultra-sensitive equipment against earthquakes is developed and verified. First, the equipment isolation

doi:10.6342/NTU201901242 6

systems with adequate periods are still necessary. Then, for the vibration-sensitive equipment, the active and semi-active control systems integrated with the advanced control algorithms and the signal processing techniques can be used to compensate for inertial loading when subjected to earthquakes, even near-fault earthquakes. For the ultra-sensitive equipment, a well-designed isolation system is not enough at all; alternatively, the ground motion features identified from the feature extraction techniques can be used to issue the early warning and turn off the ultra-sensitive equipment before far distant earthquakes bring impact. With the advanced control algorithm and the informative early warning, the seismic vibration of vibration-sensitive equipment can be mitigated, the functionality of critical equipment can be secured, and the overall seismic hazard can be consequently reduced. The detailed organization of this dissertation is briefly described as follows:

Chapter 2: The useful signal processing and feature extraction techniques are first introduced and reviewed. Considering that seismic excitations are highly non-stationary in both amplitude and frequency, the introduced techniques can be implemented not only in the time domain but also in the frequency domain to reveal the important information hiding in the non-stationary time series. With the features extracted from these techniques, the essential problem that causes the damage to the vibration-sensitive equipment and the shutdown to the ultra-sensitive equipment can be successfully solved. Finally, the advantages of these techniques compared with other traditional techniques are briefly summarized.

Chapter 3: To avoid the shutdown caused by the far distant earthquakes, the strong motion networks in Taiwan are introduced and the seismic waveforms of these far distant earthquakes collected from the broadband seismometers are investigated. The dynamic characteristics of the ultra-sensitive equipment are then numerically exanimated. The key features hiding in the seismic waveforms generated by the far distant earthquakes are efficiently extracted through the above-mentioned techniques and also connected with the unique characteristics of the ultra-sensitive equipment. Subsequently, the possibility that uses the extracted ground motion features to issue an early warning and turn off the equipment is discussed.

Chapter 4: Following the last chapter, the far distant earthquakes happened around the circum-Pacific seismic belt were sifted from the database of the United States Geological Survey (USGS).

The seismic waveforms of the one-fourth of these earthquakes are collected from the strong motion networks in Taiwan for further examination, including the ones that produced an impact on the ultra-sensitive equipment. The cardinal and principal direction of these seismic waves are also compared to provide the fast and correct information. Finally, for the ultra-sensitive equipment, the early

doi:10.6342/NTU201901242 7

warning resulted from the proposed ground motion features is verified and this chapter concludes with a discussion.

Chapter 5: To enhance the active and semi-active isolation system, the traditional control algorithms are first reviewed for the vibration-sensitive equipment against strong earthquakes. Two advanced control algorithms individually corresponding to the active and semi-active control systems are developed by integrating the traditional control algorithms with the real-time signal processing techniques. Then, several performance indices are proposed to evaluate the control performance of not only the vibration-sensitive equipment but also the structures that house the equipment.

Subsequently, the numerical simulation is conducted to verify these control algorithms and the results are compared under different earthquakes and control strategies.

Chapter 6: Following the numerical study, the full-scale experiment is conducted through a 3‐

story steel frame equipped with an equipment isolation system and a semi-active control device. In this experimental validation, the specimen is subjected to one‐dimensional earthquake excitations using a shaking table system. The structural system and the control system, integrating with sensors, damper, and processors, are described and the results of isolated equipment, passive control, as well as semi-active control are sequentially presented. The control performance and efficiency are then demonstrated and this chapter concludes with a discussion, even it may be slightly varied for the multi-dimensional excitations.

Chapter 7: Last of all, this chapter summarizes the study presented in this dissertation and provides recommendations and possible directions for future work on the earthquake protection of critical equipment, particularly the vibration-sensitive equipment and the ultra-sensitive equipment.

The original contributions of this dissertation will be:

1. One of these signal processing techniques is interpreted as a matrix multiplication from a linear algebra point of view, updated to allow the real-time processing, and verified by the shaking table test.

2. The ground motion features are linked to the impact brought by the far distant earthquakes through the numerical study on the dynamic characteristics of the isolated equipment; moreover, the possibility of issuing the timely warning, turning off the equipment, and avoiding the impact before the arrival of the main shock is demonstrated.

3. The traditional control algorithms are improved by combining the real-time signal processing

doi:10.6342/NTU201901242 8

techniques for mitigating the seismic vibration produced by the local ground motions, even those generated by near-fault earthquakes.

4. This study integrates the signal processing and features techniques individually with early warning systems and control algorithms; therefore, it forms the comprehensive measure for vibration-sensitive equipment against earthquakes.

doi:10.6342/NTU201901242 9