Bender element (BE) testing has become a widely used technique to evaluate S-wave velocities and to derive shear moduli from them in laboratory tests. Results of BE tests depend heavily on the proper identification of the S-wave travel time from waveforms often affected detrimentally by factors such as the near-field effect, crosstalk, noise, or drift. The authors have performed BE tests on soil specimens of Texcoco Clay over the last few years in triaxial cells and oedometers and have sometimes dealt with waveforms influenced by those undesirable factors thus, it was observed that the using of different available techniques to evaluate the S-wave travel time are often prone to errors. This paper shows that the S-wave travel time can be assessed unambiguously from the distance between the inflection points present in both the source and the received signals when waveforms generated with different input frequencies are superimposed, even in signals affected by the near-field effect, crosstalk, noise, or drift. This approach, named here the Wave Inflection Method (WI method), yields good results, at least when applied to data obtained from BE tests on soft Texcoco Clays. The method was validated by applying it to synthetic signals and experimental waveforms obtained from triaxial and oedometer tests equipped with BE. When applied to previously published waveforms on a wide range of soil types, our results suggest that the WI method significantly reduces subjectivity and produces improved reliability in evaluating the S-wave travel time.
The determination of shear wave velocity (Vs) using bender elements can be affected by the near-field effect. This phenomenon hinders the identification of the S-wave arrival time of an output signal. Multiple alternatives have been presented to counteract these effects, including the use of wavelet transforms during the last decade. In this article, the Haar wavelet is used because of the simplicity and speed of calculation. The determination of S-wave arrival time in signals influenced by the near-field effect is performed by setting an explicit event with a high value of energy at a specific time, in relation to the scale average energy in different levels of multiresolution analysis. The results show that regardless of the input frequency, the S-wave arrival time values are very similar, and the near-field effect is not observed when wavelet approach is used to determine the S-wave arrival times, unlike when applying other traditional interpretation approaches.
5 April 2024 in Soil Dynamics and Earthquake Engineering