|nisee||National Information Service for Earthquake Engineering
University of California, Berkeley
Note:The following paper and references are taken from part of the lecture notes for "Seismic Loading: Code Versus Site Specific" presented at a "Portland Regional Seminar on Seismic Engineering Issues" in September, 1995. The notes are used by permission of the author.
One of the most important decisions in carrying out proper design is to select a design earthquake that adequately represents the ground motion expected at a particular site and in particular the motion that would drive the structure to its critical response, resulting in the highest damage potential. The quantification of such ground motion is not easy. It requires a good understanding of the ground motion parameters that characterize the severity and the damage potential of the earthquake ground motion and the seismological, geological, and topographic factors that affect them.
Early reliance on peak ground acceleration seemed very natural to engineers since Newton's Second Law clearly sets acceleration and the resulting inertial forces in direct proportion. However, more often than not, the peak acceleration corresponds to high frequencies which are out of the range of the natural frequencies of most structures. Therefore, large values of peak ground acceleration alone can seldom initiate either resonance in the elastic range or be responsible for large scale damage in the inelastic range.
Therefore, different parameters are required to characterize the severity and the damage potential of the earthquake ground motion. In general, these parameters can be grouped into three categories.
Peak ground motions include peak ground acceleration, peak ground velocity, and peak ground displacement. Of these, peak ground acceleration is the parameter most often associated with severity of ground motion. However, it has generally come to be recognized that this is a poor parameter for evaluating the damage potential. For example, a large recorded peak acceleration may be associated with a short duration impulse of high frequency (acceleration spike) or with a long duration impulse of low frequency (acceleration pulse). In the first case, most of the impulse is absorbed by the inertia of the structure with little deformation. However, a more moderate acceleration in the second case can result in a significant deformation of the structure. For the second case, Anderson and Bertero (1987) have suggested the use of maximum incremental velocity (IV) and maximum incremental displacement (ID) for characterizing the damage potential of earthquake motion. Incremental velocity represents the area under an acceleration pulse and the area under the velocity pulse equals the incremental displacement.
Anderson and Naeim (1984), Anderson and Bertero (1987), Uang and Bertero (1988), and Bertero et al. (1991) have shown that earthquake ground motion attributes such as frequency content, duration, velocity, displacement, incremental velocity, and incremental displacement can have profound effects on the structural response than the peak ground acceleration, particularly in the inelastic range. Naeim and Anderson (1993) have evaluated and classified the range and relative importance of these attributes.
Response spectra analysis is currently the most popular method of dynamic response analysis. Response spectra representation can either be elastic or inelastic; whereas inelastic spectra can either be constant strength or constant ductility. One of the more significant shortcomings of the current elastic as well as inelastic design spectra is the fact that they do not account for the duration of input ground motion. Energy spectra which affects the possibility of high energy dissipation demand with long duration addresses this drawback. Near-fault impulse type ground motion results in a sudden burst of energy into the structure which must be dissipated immediately. This is usually characterized by a single unidirectional large yield excursion. On the other hand, the sinusoidal type ground motion with longer duration requires a more steady dissipation of energy over a longer period of time with numerous yield reversals. The current earthquake response spectra do not provide this type of information.
Hysteretic energy spectra give a more reliable indication of the damage potential and the degree of inelastic deformation. Development of input energy spectra and hysteretic energy spectra for a large database of recorded ground motion by Naeim and Anderson (1993) is a major step for providing information for improving current earthquake-resistant design procedures.
During the past 15 years an ever-increasing database of recorded earthquakes has indicated that the dynamic characteristics of the ground motion can vary significantly between recording stations that are located in the same general area. This is particularly true for recording stations located in the epicentral region. Some of these motions have distinguishable acceleration pulses. In some cases the motions contain high peak acceleration, short duration pulses that are known as acceleration spikes. In other cases, the pulses are of longer duration but lower peak acceleration. In the near-fault region, seismologists described this phenomenon as being the result of shear waves moving in the direction of the fault rupture being crowded together to produce long duration pulse or "fling" (Singh 1985). The critical condition with respect to the damage potential to structures occurs when a long duration pulse has an average acceleration that is of the same order as the yield resistance seismic coefficient of the structure. The yield resistance of a structure can be expressed in terms of the yield resistance seismic coefficient by dividing the base shear capacity with its effective weight. Acceleration amplitudes in the range of 2035 percent of gravity are sufficient to produce serious structural damage if the acceleration pulse duration is long relative to the period of the structure. What makes these motions particularly damaging is the area under the acceleration pulse that represents the incremental velocity and when multiplied by the mass represents what is called the impulse.
Perhaps the first recorded motion with a recognized acceleration pulse (having a maximum amplitude of 50 percent gravity) was that obtained at a distance of approximately 200 feet from the fault trace during the Parkfield, California, earthquake of June 1966. The maximum incremental velocity for this pulse was only 35 inches per second, which was not large enough to damage structures by forcing them into the inelastic range. For this reason, this record received limited attention. The first recorded data in the United States that contained damaging acceleration pulses was the record obtained at Pacoima Dam during the San Fernando, California earthquake of February 1971. Even in this case, the characteristic of the record that initially gained attention was not the acceleration pulse that occurred about three seconds after the instrument was triggered but the large acceleration spikes that were recorded at about eight seconds and had an amplitude in excess of 100 percent of gravity. Only after an extensive study of the heavily damaged Olive View Hospital, was it concluded (Bertero et al., 1978) that the damage was caused by the long duration pulse rather than the high frequency acceleration spike. The maximum incremental velocity that occurred in the Pacoima Dam record was 64 inches per second, which is nearly twice that of the spike in the Parkfield record.
The Imperial Valley, California, earthquake of October 1979 produced approximately 25 records of ground motion data in close proximity to an active fault. The stations located in the direction of propagation, such as the recording station at the Imperial Valley College (Station No. 7) recorded a peak acceleration of 45 percent of gravity with a maximum incremental velocity of 63 inches per second. A comparison of this record with the record of the May 1940 El Centro earthquake shows that the peak ground acceleration of that record was 34 percent of gravity with a maximum incremental velocity of only about 19 inches per second. The ground motion in the near field can vary significantly depending upon the direction of rupture propagation.
Although the short duration acceleration spike that has a high peak acceleration is typical of the near-fault region, long duration pulses can occur at sites located at large distances from the epicenter. For example, this was the case during the earthquake at Bucharest, Romania, in March 1977. Here the acceleration amplitude was a very modest 20 percent of gravity but the maximum incremental velocity was approximately 50 inches per second.
Because the nonlinear dynamic response of a structure, and in particular the displacement response, is very sensitive to the dynamic characteristics of the ground motion, it is important that earthquake-resistant design be given special considerations in the near-fault region and other regions where long duration acceleration pulses can occur. The near-fault impulse type ground motion can be particularly damaging to structures and can even cause collapse. Under an impulse-type load the interaction effects between lateral displacements and axial load (P-Delta) can be exaggerated. Studies have shown that a structure in the near-fault region may experience a dynamic response that is twice or more of a similar structure located at some distance from the fault, even though the peak ground accelerations are about the same. In order to obtain a more meaningful classification, the use of other parameters such as incremental velocity in peak ground displacement may be required.
Near-fault motions are dominated by the source characteristics. For example, near a strike-slip fault, the properties of the extended "double-couple" source tend to dominate the ground motion. First, the seismic source directivity can overshadow the effects of soils on ground motion as a result of propagation of the rupture along the fault, and, second, ground motion perpendicular to the fault can be several times stronger than that parallel to the fault.
Anderson, J. C., and V. V. Bertero. 1987. "Uncertainties in Establishing Design Earthquakes." J. Structural Engineering. Vol. 113. No. 8. August: 1709-24.
Bertero, V.V., J. C. Anderson, H. Krawinkler, E. Miranda. The CUREe and Kajima Research Teams. 1991. Design Guidelines for Ductility and Drift Limits: Review of State-of-the-Practice and State-of-the-Art in Ductility and Drift-Based Earthquake-Resistant Design of Tall Buildings. Berkeley, California: Earthquake Engineering Research Center, University of California at Berkeley. UCB/EERC-91/15.
Chopra, A. K. 1995. Dynamics of Structures, Theory and Applications to Earthquake Engineering. Englewood Cliffs, N.J.: Prentice Hall.
Heaton, T. H., J. F. Hall, D. J. Wald, M. W. Halling. 1995. "Response of High Rise and Base Isolated Buildings to Hypothetical MW 7.0 Blind Thrust Earthquake." Science Magazine. Vol. 267: 206-211.
Iwan, W. D. 1994. "Near Field Considerations in Specification of Seismic Design Motions for Structures." Proceedings of the 10th European Conference on Earthquake Engineering. Vienna: A. A. Balkema, Rotterdam. Vol. 1: 257-267.
Lilhanand, K., and W. S.Tseng. 1989. "Development and Application of Realistic Earthquake Time Histories Comparable with Multiple-Damping Design Spectra." Proceedings of the Ninth World Conference on Earthquake Engineering. Tokyo-Kyoto, Japan. Tokyo: 9 WCEE Organizing Committee, Japan Association for Earthquake Disaster Prevention. Vol. 2: 819-824
Singh, J. P. 1981. The Influence of Seismic Source Directivity on Strong Ground Motion. Ph.D. dissertation. Berkeley, Calif: University of California. COMTEX Scientific Corporation. N.Y.
Singh, J. P. 1982. "Directivity Evidence from the August 6, 1979 Coyote Lake Earthquake." Proceedings, Conference on Earthquake Hazards in the Eastern San Francisco Bay Area, Hayward, California. California Division of Mines and Geology. Special Publication 62, March 24-27.
Singh, J. P. 1982. "Importance of Local Structure and Source Characteristics in Estimation of Near-field Strong Ground Motion." Proceedings, 3rd International Earthquake Microzonation Conference, University of Washington, Seattle, June 28-July 1. Vol. VII: 623-630.
Singh, J. P. 1983. "New Procedures for Estimating Seismic Design Input for Near-Field Conditions."Proceedings, Lawrence Livermore National Laboratory Seismic Risk and Heavy Industrial Facilities Conference. San Francisco, California, May 11-13.
Singh, J. P. 1984. "Characteristics of Near-Field Strong Ground Motion and Their Importance in Building Design."Proceedings, Applied Technology Council (ATC) Seminar on Earthquake Ground Motion and Building Damage Potential, San Francisco, March 27. ATC 10-1.
Singh, J. P. 1986. "A Simple Method for Generating Synthetic Time Histories for Design of Base Isolation Systems." Proceedings Applied Technology Council (ATC) Seminar on Base Isolation and Passive Energy Dissipation Devices, March 12&13, San Francisco, ATC-17. Redwood City, Calif.: ATC. 391-402.
Singh, J. P., M. Tabatabie, and L. D. Porter. 1991. "Strong Ground Motion Data - Application to Multiple Support Structures." Geotechnical News. Vol. 9. No. 1. March.
Uang, Chia-Ming and V. V. Bertero. 1988. "Implications of Recorded Earthquake Ground Motion on Seismic Design of Building Structures." Berkeley, Calif.: Earthquake Engineering, Research Center, University of California, Berkeley. UCB/EERC-88/13.