Vibration and Soil-structure Interaction Tests Nine-story Reinforced Concrete Buii,ding by Paul C. Jennings and Julio

The Millikan Library Building, a nine-story reinforced concrete shear-wall structure at the California Institute of Technology, was tested dynamically by means of two eccentric mass vibration generators located on the ninth floor. The response levels ranged up to a maximum acceleration of 0:02 g. The natural periods of vibration, the mode shapes and the energy dissipation were measured for the first and second E-W translational modes, the N-S fundamental mode and the first torsional mode. Soil-structure interaction was investigated by measuring foundation motion and nearby soil surface movements during resonent vibrations in the N-S and E-W fundamental modes. Other tests included "man-excited" vibrations at low stress levels and a measurement of resonance of an air handling unit on the roof, which was found to magnify the roof response by a factor of 8.5. The measured fundamental periods were short compared to typical framed structures of this height, 0.50 sec in the N-S direction, 0.66 in the E-W direction and 0.46 in torsion. These values increased roughly 3 per cent over the range of testing. The energy dissipation as measured by a viscous damping factor, varied between 0.70 and 2.00 per cent of critical. This large variation over the testing range indicates that tests at higher stresses are needed to determine the energy dissipation expected during the response to strong earthquake motions. The soil-structure interaction measurements showed that the building responded very nearly as if fixed at the foundation; rocking contributed less than 1 per cent to the total roof motions of the structure and foundation translation about 2 per cent. Although negligible as far as the building motion is concerned, the results demonstrate the possibility of performing full-scale soil-structure interaction experiments. INTRODUCTION Full-scale vibration tests using large shaking machines have proved to be one of the most effective methods for determining the dynamic characteristics of structures needed for understanding the effects of strong earthquake motions. With these vibration generators it is possible to determine experimentally the most important periods and mode shapes of many actual structures, and to measure their energy dissipation properties. Also, as reported below, these machines can be used to perform full-scale experiments on the phenomenon of soil-structure interaction. The principal limitation of the vibration generators is that the testing range is generally limited to small vibrations. An important advance in this field of dynamic testing was the development in 1961 of a vibration generating system with precise speed control and the capability 891 892 B U L L E T I N OF T H E S EIS M OLOGICAL SOCIETY OF A M E R I C A of phase control among separate units (Hudson, 1962). These vibration generators have been used to test a variety of actual structures, including dams (Keightley, 1963, 1964); a reservoir intake tower (Keightley et al, 1961); an atomic reactor (Matthiesen and Smith, 1966); a variety of buildings (Nielsen, 1964; Bouwkamp and Blohm, 1966; Englekirk and Matthiesen, 1967; Hanson, 1965a); and special struetures (Hanson, 1965b). In Japan also, a program of full-scale vibration tests has been carried out using similar equipment (Kawasumi and Kanai, 1956; Hisada and Nakagawa, 1956; Takeuchi, 1960; Nakagawa, 1960). The main purpose of the tests reported here was twofold. First, to study the dynamie characteristics of the Millikan Library Building, and second to study the amount of interaction between the structure and the surrounding soil. These results will b e useful in analytical studies being made of the structure and in the interpretation of future records from the strong-motion aecelerographs which are installed in the building. In this test it was possible to excite the structure in the fundamental modes in the N-S and E-W directions, and in torsion. I t was possible also to study the seeond E-W mode to a limited extent but other modes were out of the range of the equipment. Soil-structure interaction data were taken for the two fundamental modes in translation. A more detailed diseussion of test results and procedures, and additional data not reported here are available in a recent report (Kuroiwa, 1967). Building Data. The strueture tested was the Robert A. Millikan Library Building located on the campus of the California Institute of Technology. Built during 19661967, it is a nine-story reinforeed concrete building with a basement. The overall dimensions of the building are shown in Figures la, lb and le, and a general view in Figure ld. The floor system eonsists of 9 in. slabs of lightweight concrete reinforced in two directions and supported by reinforced concrete beams. The north side and south side facades, installed during preliminary testing, are preeast window wall panels weighing 11 tons each, connected to the main structure with steel angle clips. The lateral resistance in the N-S direction is supplied primarily by the two end walls; and in the E-W direction by the central core. The masses in kips at the time the tests were performed were assumed to be lumped at each floor level and were estimated to be: M10(roof) = 2,600; M9 Ma = 1950; M2 = 2433; and M1 = 2280. The mass of the roof ineludes the mechanical equipment, most of which is used to heat and cool the building. Instrumentation and Test Procedures. The vibration tests consist basically of determining the steady-state response of the building to a sinusoidal exciting force produced by the eccentric masses of the vibration generators. The building was excited with two generators, installed on the ninth floor at locations A and B of Figure lb. Each unit generates a unidirectional, sinusoidal force which can be varied in magnitude and direction. The maximum force level is about 5000 lbs, and the maximum rotational speed is about 10 cycles per second. Particulars of the vibrators, including operating instructions :for the machines, are available (Hudson, 1962). The response of the building and nearby ground at selected locations was recorded with a 6-channel aecelerometer-amplifier-reeorder system. Data were recorded only after transient structural vibrations had died out; this occurred quickly except near resonance where 20-30 cycles of response were often necessary. By taking a series of