Analysis of the earthquake response of a nine-story steel frame building during the San Fernando earthquake

Abstract

A study has been made of the earthquake response of the ninestory steel frame Building 180, located at the C.I.T. Jet Propulsion Laboratory, Pasadena, during the San Fernando earthquake of February 9, 1971. The study was motivated by the likelihood that an earthquake similar to the February 9, 1971, shock could occur close to the JPL grounds with consequent very strong ground shaking. It was, therefore, judged desirable to make a thorough study of the response of the building to the moderately strong ground shaking during the February 9, 1971 event. The analysis throws light on the actual dynamical properties of the building during the earthquake, and also demonstrates that it is possible to make accurate calculations of building motions during earthquakes when the ground motion is specified. Methods of evaluating the lower mode periods and damping ratios from the earthquake records are described and these values are compared with values obtained by dynamic testing before and after the earthquake and with the periods computed from structural models of the building. Although no structural damage, as a result of the earthquake motions, was detected, and computed stresses in the frame were less than yield stresses, the ambient test periods after the earthquake were approximately 1016 higher than the pre-earthquake test values. The maximum periods during the earthquake were found to be about 30% higher than the post-earthquake ambient test values, and it is believed that this significant increase was mainly the result of the non-linear behavior of the concrete encased steel columns. The periods computed from the structural models based on two different assumptions regarding the influence of the composite concrete agreed within 5% of the postearthquake ambient test results and within 10% of the earthquake periods. The model roof acceleration responses were computed using the recorded base motions as inputs. Good agreement was obtained between the recorded and computed roof responses confirming that the elastic response computation method currently used in the earthquake design of buildings gives satisfactory stress predictions. A stress analysis of the lower stories of the model gave a maximum total steel stress during the earthquake of 34 Kips/in2. This maximum value occurred in the second story columns and was computed by adding to the gravity load stresses the root- mean- square sum of the two horizontal earthquake stress components. The Fourier amplitude spectra of the recorded horizontal base motions were found to contain peaks corresponding closely with a number of the lower mode natural frequencies. A study was undertaken to determine whether these peaks could have resulted from soil-structure interaction. It was concluded that soil-structure interaction would be unlikely to have produced significant changes to the horizontal freefield motions. The influence of soil- structure interaction in the vertical direction was found to be rather more significant, producing a peak and a dip in the Fourier amplitude spectrum close to the fundamental vertical frequency.

Files

Additional details