Of the three testing methods used to determine V and V , it is the opinion of the S P authors that DST is the preferred technique. Compared to CST, DST is more cost effective, does not require two boreholes and in-situ disturbance is significantly lower. SASW has some limitations that should be carefully considered before this method is used: • SASW requires surface measurements that are significantly more susceptible to measurement noise compared to downhole receivers utilized with DST. • Limited depth of penetration (one to two wavelengths) of Rayleigh waves implies a limited depth of in-situ site characterization. • SASW requires and depends on accuracy of a near surface earth model or structure. • Pertinent data (e.g., phase measure- ments for generating dispersion curves) are not easy to obtain. • Thin layers that are either much stiffer or much softer than the surrounding material can be missed. • A portion of each layer adjacent to a large velocity contrast is difficult to resolve. It should be noted that the commonly stated disadvantage of DST is that near surface estimates are difficult to obtain and subsequent interval velocity estimates are inaccurate. However, it has been shown that this disadvantage is resolved by using relatively large radial offsets for the seismic source and by applying ray path refraction analysis when estimating interval velocities. Seismic Cone Penetration Testing SCPT has proven a very cost-effective DST method for ground that can be penetrated, which results in minimal soil disturbance, does not require multiple boreholes for site characterization, has excellent probe-soil coupling and requires a comparatively short set-up time. SCPT was devised to measure seismic velocities directly through data obtained by seismic sensors installed in the cone, in addition to the standard bearing pressure, sleeve friction, and pore pressure sensors. As the cone is advanced through the ground, the advance is halted at regular intervals (commonly 1 m or 3.3 ft), at which time a seismic event is generated at the surface (e.g., a hammer blow onto a steel plate) causing seismic waves to propagate from the surface through the soil to the seismic sensors. This event is recorded, after which the cone is advanced another increment and the process is repeated. By determining the seismic arrival times, the seismic velocities can then be calculated for each interval. For the test equipment set up, relatively large sensor-source radial offsets (SSRO) are recommended because a seismic event creates a displacement field that contains both near-field and (desired) far-field components. The near-field particle motions are quite complex and tend not to adhere to Hooke’s law; therefore, it is common practice to use only the far- field terms. Since a pro- pagation distance of two to four wavelengths is necessary before the far-field term is the dominant component in the source wave displacement field, the travel distance should be at least 5 m (about 16.5 ft) considering that a source wave with a dominant frequency of 80 Hz traveling within a soil medium having a velocity of 200 m/s (656 ft/s) will have a corresponding wavelength of 2.5 m (8.2 ft). A larger SSRO will signifi- cantly decrease near-field amplitudes. Another advantage of must be considered when analyzing the acquired seismic SCPT data. As an illustrative example, seismic data with very similar source wave arrival times were recorded at each depth increment. However, when processing this data set using the straight ray assumption (SRA), it appears that nonsensical SH-source waves were recorded, with interval velocities exceeding 5800 m/s (19030 f t /s). Conversely, if ray path refraction and Fermat’s Principle are considered using the Forward Modeling / Downhill Simplex Method (FMDSM), the resulting interval velocities become very plausible and the ray paths are anything but straight lines. Triaxial SCPT data acquired at different depths large SSROs is that noise from nearby CPT rods interfere with the source wa v e i s s i gni f i c ant l y reduced. Large SSROs increase the near surface characterization of the layer or depth under analysis because the source wave travels within stratigraphic layers for a longer period. When applying larger SSROs, the implementation of Fermat’s Principle, which states that a wave will take that ray path with minimal travel time, Graphical results for interval velocity estimates using FMDSM A challenging aspect is to characterize a recorded data set to determine the analysis method that will result in the most accurate seismic interval velocities, even when refraction and Fermat’s Principle are applied. This analysis is referred to as seismic trace characterization, and, in recent DEEP FOUNDATIONS • MAY/JUNE 2018 • 105
