抄録
It is significant to accurately evaluate how a building is damaged soon after a seismic event from the building use continuation judgement point of view. As an effective tool for such a judgement, structural health monitoring (SHM) schemes have appealed the attentions of structural engineers. In conducting SHM for a building structure, it would be desirable to install sensors into all the stories. At the present stage, however, such a sensor installation scheme would be expensive, and thus it would be realistic to construct the SHM strategy based on the use of a limited number of sensors. From the point of view of what kind of responses are appropriate for effective SHM, on the other hand, inter story drift deformation could be a powerful index. The second authors' research group has developed a direct measurement devise sensing drift displacements. Accounting for the fact that accelerometers have been much more practically utilized, this paper proposes a methodology which would estimate the drift deformations for all the stories with accelerometer implementation on a limited number of floors.
The beneficial point of the proposed method is to derive all of the necessary information with respect to a building from the measured response data. In other word, the estimation of all the stories' responses is conducted based on only the accelerometer data. The method proposed in this paper utilizes cubic spline interpolation (CSI). In applying CSI to building's response interpolation, how to set the end conditions at the top floor and base floor locations is very significant. This paper proposes how those end conditions should be set. As the boundary condition with respect to the building top floor vicinity, the location over the top floor at which the first derivative of the response distribution is always zero is introduced. This location is to be determined from the first mode shape of a building. As above-mentioned, however, this method does not use other information than the measured acceleration data. The first mode shape is derived from the acceleration responses. For the boundary condition at the base floor location, the degree of fixation at the base floor is accounted for. The semi-fixed boundary condition is introduced by mixing the fixed and pin base conditions. The mixing ratio of these two is also determined from the derived first mode shape. Employing the two newly-employed boundary conditions, all the story responses are estimated accurately.
The validity of the proposed method is demonstrated by using the experimental data from the E-Defense shake table tests of a 1/3 scale model steel building with eighteen stories in December 2013. In this experiment, the model building was repeatedly shaken with simulated long-period ground motions with different magnitudes. The ground motion became larger and larger until the model building was collapsed. Employing the data measured by five accelerometers set on the first, second, seventh, thirteenth and roof top floors for some cases of the experiment, the paper demonstrates how accurately the acceleration response time histories of all the stories are estimated and then the drift deformation time histories are estimated.
The interpolation method proposed in this paper provides an accurate estimation of all the story responses. It would lead to an accurate judgement framework of story-based damage states only with the installation of a limited number of sensor and without any of pre-information on the motion characteristics of a building, such as mode shape function.
