Source near-field effects and pulse tests in soil samples

Abstract

Shear modulus measurement using bender elements in laboratory samples has become very popular. However, the test results are hard to interpret. Simple plane wave test models are too coarse and result in substantial measurement uncertainty: near 100% in G0. A possible refinement of the test model is based on Stokes's fundamental solution for an isolated source. Near-field effects appear, and they have been regularly quoted as a major source of uncertainty in tests. There are some criteria in use for avoidance of near-field distortions, but they are signal dependent and not always successful. After some consideration of Stokes's fundamental solution the authors propose a new frequency domain signal-independent criterion to avoid near-field effects. Time-domain criteria are also given, but they are shown to be signal dependent. Applying this criterion to some experimental results the authors then also show how near-field induced errors are not responsible for much of the observed signal distortion. Uncertainty in signal arrival lasts well beyond the end of the Stokes source near field.