Measurement of Ground Motion

Tile need for a greatly expanded network of strong-motion accelerographs throughout the seismic regions of the world is stressed. A summary of the characteristics of currently available strong-motion accelerographs is presented, and the design details are given for an instrument suitable for acquiring the basic data needed by structural engineers for earthquake resistant design. It is shown that for such an instrument, the natural period must be less than 0.1 seconds, and that the recording speed must be at least 1 cm/sec. The critical nature of the inertia starting device is discussed, and some information is given on the transient response of the standard pendulum starter used in the United States Coast and Geodetic Survey Strong-Motion Accelerograph. The use of simpler, non-time-recording instruments such as the U.S.C.G.S. Seismoscope to supplement the accelerograph network is described. INTRODUCTION The occurrence of a great ear thquake is a powerful stimulus to engineering researches in earthquake-resistant design, and structural engineers should always be prepared to make use of the general interest aroused by destructive earthquakes to further their studies in this field. At such times it is wise to evaluate anew the adequacy of our past knowledge, and to initiate new programs to supply deficiencies in it. The Chilean earthquakes of M a y 1960, with a main shock of magnitude 8.5, form one of the largest series of destructive earthquakes of history. Coming just before the Second World Conference on Ear thquake Engineering in Tokyo and Kyoto, Japan, a t a t ime when the at tention of a large number of ear thquake engineers and seismologists was foceussed on such problems, the Chilean earthquakes offered an unparalleled opportuni ty for field studies of destructive earthquakes. The importance of field studies of undamaged as well as damaged structures cannot be overemphasized. I t must be realized, however, tha t such studies must always be incomplete without an accurate knowledge of the actual ground motions a t each part icular site. In this respect we must regretfully report tha t we do not have available from the Chilean earthquakes a single measurement of ground motion at any point at which significant damage occurred. This situation is unfor tunate ly far from uncommon. In spite of the fact tha t ins t rumental seismology has now been in existence for nearly 100 years, we do not as yet have a single ground acceleration record from the central region of a truly major earthquake. The largest ground acceleration record obtained by a U. S. Coast and Geodetic Survey Strong-Motion Accelerograph was from the E1 Centro ear thquake of M a y 18, 1940, recorded at a point about 30 miles from the center of a magnitude 7 earthquake. The only accurate ground acceleration data available on potentially destructive ground motion anywhere in the world is tha t from the dozen or so earthquakes which have been recorded by the U. S. Coast and Geodetic Survey Network in California and the Pacific Coast States, and all studies of the dynamic response of structures to earthquakes have been based on this meager data. 419 420 BULLETIN OF THE SEISMOLOGICAL SOCIETY OF AMERICA The present lack of adequate strong motion accelerograph coverage cannot be attributed to a lack of knowledge as to how to proceed, nor to an absence of warnings as to the difficulties that would face earthquake engineers in the absence of such data. As early as the 1920's Dr. K. Suyehiro, the first director of the Earthquake Research Institute of the University of Tokyo, outlined dearly the type of aceelerographs that would be needed, and emphasized again and again the importance of instruments specially designed to record the data required by structural engineers. During his American lectures on Engineering Seismology in 1931, he said: "Engineering Seismologists must prepare suitable strong-motion siesmometers and accelerographs, and after distributing them in the seismic regions await with patience the useful data that must come in the future . . . . The amplitude and the period of the main principal motions taken from a seismogram have generally not much significance for the engineer. The requisite data for him are the acceleration and its period as recorded directly by an aeeelerograph suitable for engineering use."1* These ideas of Dr. Suyehiro were followed up in the United States by Mr. J. R. Freeman, who emphasized strongly in his important book Earthquake Damage and Earthquake Insurancd the need for ground acceleration measurements. These efforts culminated in 1931 in an allocation from the U. S. Congress to the U. S. Coast and Geodetic Survey for the development of suitable aeeelerographs. The first strongmotion earthquake recordings were obtained during the Long Beach earthquake of March 10, 1933, of magnitude 6.3, and since that time useful ground acceleration records have been obtained for about 20 moderately strong earthquakes occurring in the California-Pacific Coast region. There are at present about 60 of the standard U. S. Coast and Geodetic Survey Strong-Motion Aceelerographs in operation in the Western United States, and a few of these instruments have been sent to other countries to serve as prototypes for new networks. A similar network of strong-motion aceelerographs is now installed in Japan, and comparably useful records may be expected from that region. Two accelerometers suitable for the purpose were developed by the Japanese "Strong Motion Acceleration Committee" (S.M.A.C.) formed in 1951. 3 As of 1960, approximately 50 SMAC Type Accelerographs and 15 DC Type Accelerographs were in operation in Japan. 4 THE DESIGN OF A STRONG-MoTION ACCELEROMETER In order to understand the basic design requirements of a seismograph suitable for engineering applications, we must first consider the use to which the data are to be put, and second, the characteristics of the ground motions that are to be measured. The fundamental problem which confronts the structural engineer in earthquake-resistant design is the determination of the dynamic strains induced in structures by the base motion. The equations relating the earthquake ground motion to the resulting structural strains give the relative displacements of the structure in terms of the ground acceleration. It thus develops that, for the desired calculation of structural responses, the true acceleration of the ground as a function of time is required. * Numbers refer to references at the end of the paper. MEASUREMENT OF GROUND MOTION 421 There is, of course, a simple theoretical relationship between ground displacement, velocity, and acceleration, so that it might be thought that it is immaterial which of these quantities is originally measured. Considering the accuracy needed, however, it turns out that it is essential that the original measurement be of ground acceleration. This is a consequence of the fact that integration is inherently a more accurate process than differentiation. An acceleration-time curve can be integrated to give velocity and again to give displacement with an acceptable accuracy. It would not in general be possible to reverse this process and obtain a usable acceleration-time curve from a displacement curve by differentiation. The inevitable