Strong motion accelerograph evaluation

Abstract

A brief study is made of the effect of common instrument errors on the accuracy of data obtained from strong motion earthquake accelerographs. Error sources considered include zero drift, tilts, nonlinearities, cross-axis sensitivity, lack of initial conditions, noise, and time base errors. It is concluded that most data of current engineering interest are not critically affected by the level of errors found in existing accelerographs. Techniques are suggested for reducing or eliminating many of these errors by instrument design changes. An experimental study is made of a new strong motion accelerograph during its engineering development. This new accelerograph is designed to record an FM analog of ground acceleration on magnetic tape, providing a record which may be rapidly and automatically converted to digital form. The accuracy limits of the accelerograph are explored and the design reasons for these limits investigated. The more significant findings maybe briefly summarized: (1) Static accuracy. The sensitivity and linearity of the instrument are found to depend critically on a series of interdependent adjustments. Reasonable care will bring errors in both of these quantities to within ±2% of 1/2 g full scale. Higher accuracies are possible, but require much more time and care, primarily due to the limiting effect of mechanical drift in the accelerometers. (2) Zero point drift. Uncertainties in the accelerometer zero point arise from both mechanical and electronic drifts. Long term drifts may be related to temperature or relative humidity, or may be entirely random. Short term drifts of up to 2% of full scale may occur during the course of a typical record. The total variation may be as much as ± 30% of full scale for a 100 [degrees] F range of temperatures. These variations require adjustment of the data before processing, but are not sufficient to interfere with operation of the accelerograph. (3) Noise. Random noise in the system as tested amounted to 1.4% of full scale, RMS, and was mostly due to the tape recording system. By comparison with optical accelerographs, this noise figure is marginal, but acceptable, and can be improved by changes to the compensation system. (4) Timing. The advantages of an effectively continuous time base over discrete time marks were discovered and means devised to obtain such a base from the test accelerograph. This method of timing is a qualitative improvement over the best system which is practical on optical recorders. The overall performance of the test accelerograph is adequate to yield acceptably accurate acceleration vs. time records and Fourier spectra within the range of frequencies which are of current engineering interest. It is able to produce useful displacement records only for periods shorter than several seconds. The reasons for this latter limitation are sufficiently fundamental that markedly superior instruments are not expected to be available within the next ten years.

Files

Additional details