The two‐dimensional scattering and diffraction of plane SH waves by a semi‐cylindrical canyon is analysed for a general angle of wave incidence. The closed‐form solution of the problem shows that the surface topography can have prominent effects on incident waves only when the wavelengths of incident motion are short compared to the radius of a canyon. The surface amplification of displacement amplitudes around and in the canyon changes rapidly from one point to another, but the amplification is always less than 2. The over‐all trends of amplification pattern are determined by two principal parameters: (1) γ, the angle of incidence of plane SH waves, and (2) η, the ratio of radius of the canyon to one‐half wave length of incident waves. The higher η leads to greater complexity of the pattern of surface displacement amplitudes characterized by more abrupt changes of amplification from one point to another, while γ mainly determines the over‐all trends of displacement amplitudes. For grazing and nearly grazing incidences, for example, a strong shadow zone is developed behind the canyon. The qualitative analysis of the topographic effects on the Pacoima Dam accelerogram,1 based on the semi‐cylindrical canyon, suggests that this strong‐motion record was not seriously affected by surface topography of the recording site.

Wave travel times of vertically propagating waves are measured in the Van Nuys 7‐storey hotel, located in the Los Angeles metropolitan area, from impulse response functions computed by deconvolution of the recorded earthquake response. East‐West response data from 11 earthquakes over a period of 24 years are analysed. Changes in wave travel times are used to infer about local (between sensors) and global changes of structural stiffness, from one event to another, and with time during the earthquakes that damaged the building (San Fernando and Northridge). Also, wave travel times are used to estimate the fundamental fixed‐base frequency of the building, f1, which is compared with independent estimates of the soil–structure system frequency fsys during the same earthquakes and during five ambient vibration tests. The analysis shows that monitoring only the changes of fsys can be misleading for structural health monitoring and can lead to erroneous alarms, while monitoring changes of f1 over suitably chosen time windows (before, during, and after excitation by strong earthquake motions) can be a powerful and robust tool for structural health monitoring. It is concluded that, under favourable conditions, this method can be used as a tool for global and local structural health monitoring. Copyright © 2007 John Wiley & Sons, Ltd.

The ambient and forced vibration techniques for testing full‐scale structures are critically compared. Both methods, based on small level excitation, may be used to determine many mode shapes and frequencies of vibration and the corresponding damping values, with adequate accuracy for most purposes. The two techniques give mutually consistent results. The mode amplitudes determined by ambient and forced vibration tests show systematic departure for high modes and near the top levels of buildings tested. This phenomenon is attributed to the participation of all mode shapes and is a consequence of excitation by a concentrated force near the top of a building and at a frequency differing by only a few per cent from a natural frequency of vibrations. A new way of showing the effect of unwanted modes on the response near resonance of the mode being sought is developed. It is particularly useful for the analysis of steady, forced vibration tests of structures using eccentric mass vibration generators.

No abstract

No abstract

No abstract

No abstract

SUMMARYForced vibration tests designed to isolate the effects of soil-structure interaction are described and the results obtained for the nine-storey reinforced concrete Millikan Library Building are analysed. It is shown that it is possible to determine experimentally the fixed-base natural frequencies and modal damping ratios of the superstructure. These values may be significantly different from the resonant frequencies and damping ratios of the complete structure-foundation-soil system. It is also shown that forced vibration tests can be used to obtain estimates of the foundation impedance functions.In the case of the Millikan Library it is found that during forced vibration tests the rigid-body motion associated with translation and rocking of the base accounts for more than 30 per cent of the total response on the roof and that the deformation of the superstructure at the fundamental frequencies of the system is almost entirely due to the inertial forces generated by translation and rocking of the base.