Construction of a Large Laminar Box for Experimental Investigation of Soil Effect on the Dynamic Response of SDOF Structures

Abstract

More and more emphasis has been laid on soil response and soil-foundation-structure interaction (SFSI) in recent research in earthquake engineering. These effects are mostly ignored in the traditional structural design procedure in which a structure is assumed to be fixed at its base. In order to have a better understanding of these effects, a large laminar box was developed. It was a flexible soil container which can reasonably simulate the boundary conditions observed in the field by allowing shear deformation of soil during excitations. Compared with the existing small laminar box at the University of Auckland, this newly-designed large box can be used to consider a large volume of soil specimen of 8 m3. In addition, it can be used to study soil liquefaction effect in large-scale experiments for future research. The large laminar box was a single axis shear box consisting of thirty-five horizontal frames. Three upright steel columns were installed on both sides parallel to the movement direction to support these frames. The frames were made of aluminium considering its light self-weight to reduce inertia effects on the soil response during earthquake excitations. Every frame had the same weight and an internal size of 2 m wide by 2 m long by 5 cm thick. With 7 mm gap between each frame, all frames can move freely with negligible friction. The overall internal dimension of the box was 2 m by 2 m by 2 m. The maximum allowable sliding displacement of the thirty-five frames was ±175 mm in the excitation direction. It was equivalent to a possible maximum shear strain of the soil of nearly 9%. Besides, inside the base of the box, a pipe system was installed to allow water to go inside the box. A waterproof membrane made of four flexible and durable PVC fabrics was used inside the box. Hence, the box can be utilised to test both dry and wet soil. In order to prepare soil specimen, a large wooden sand rainer matched with the laminar box was constructed. It had an internal dimension of 2.36 m wide by 2.36 m long by 0.38 m deep. In the central area (2 m by 2 m) of the sand rainer, nearly two thousand holes were drilled allowing the soil to pluviate into the laminar box. This ensured that the soil specimen had similar relative density throughout. Considering the transportation of the filled sand rainer, a steel supporting frame was built. As an application example a series of shake table tests were performed. In order to investigate how a structure would respond when it was located on the shallow soil and underwent relatively small earthquake excitations, the laminar box was filled with 250 mm depth of dry sand with a single degree-of-freedom (SDOF) structure located on the sand surface. The fundamental frequency of this structure was 2.77 Hz. The soil density was almost constant during three shakings. It was found that the structure responded linearly. The second case of this study was about the investigation of soil density effect on the soil and structure response with SFSI. The laminar box was filled with around 1 m depth of dry sand with four different SDOF structures located on the sand surface together. The fundamental frequency of four structural models was 0.59 Hz (M1), 1.18 Hz (M2), 2.53 Hz (M3) and 2.77 Hz (M4), respectively. There were ten shakings conducted in total. The soil density gradually increased from the first to the eight shaking, while it was almost the same during the last three shakings. The ground motions were different in last three shakings. The experimental results showed that an increase in soil density had insignificant effect on the top response of flexible structures (M1 and M2) but considerable effect on that of rigid structures (M3 and M4). The increase in soil density can also affect the footing response of a structure and soil response beneath the footing of the structure. This effect was more significant on rigid structures. Similar structural response and soil behaviour can be observed when a structure was subject to excitations with the same dominant frequency content and close predominant frequency (details presented in chapter 6). This study also aimed to understand how SFSI affects the behaviour of the soil and structure. Four models mentioned above were placed on the shake table with a fixed base condition and subject to three different shakings. The fixed-base results were used as a reference. The comparison results indicated that soil would filter high frequencies of earthquake excitations, thus reducing the top response of the structure. The soil response can be influenced by SFSI and in turn affected the response of the structure. These results suggested that it was inaccurate to estimate the earthquake response of a structure using the ground motions measured in the free field.