(Click for larger image [60KB])


Revolutionizing Surface Wave Methods for Engineering Analyses - from Deterministic and Incoherent to Probabilistic and Standardized

Principal Investigator: Brady R. Cox, University of Texas at Austin

Surface wave methods (SWMs) have become fully entrenched as powerful tools in geotechnical site investigation over the past decade, and their end result—a subsurface profile of small-strain shear modulus/shear wave velocity (Vs)—is used as a key input parameter in many engineering analyses. The expanding use of SWMs is driven by the desire to "reach" within the earth and retrieve accurate and meaningful engineering parameters without the need for borings. Traditionally, SWMs have been used to provide a single, deterministic Vs profile for each site tested, without consideration given to measurement/dispersion uncertainty and how it propagates forward through the inversion process used to estimate Vs. However, as the profession moves toward probabilistic design and performance-based engineering, the inability to quantify uncertainty in Vs from SWMs has been exposed as a major impediment to future progress. An ever increasing number of researchers and practitioners are using SWMs without understanding how acquisition parameters such as spatial sampling interval, array aperture, source proximity, and signal-to-noise ratio influence the uncertainty of their results. The PI will address these issues in his career by revolutionizing SWMs from Deterministic and Incoherent to Probabilistic and Standardized (DIPS). The DIPS plan (aimed at "smoothing-out the dips" in SWMs) involves: (1) quantifying measurement/dispersion uncertainty in SWMs so that Monte Carlo-based inversions can be used to propagate this uncertainty forward into a suite of acceptable Vs profiles with confidence intervals on layer thickness and velocity (i.e., advancing from deterministic to probabilistic), and (2) developing standards for SWMs applied to solving engineering problems (i.e., advancing from incoherent recommendations to coherent standards). The DIPS plan is guided by the vision to collect and analyze a unique, large and freely-shared set of experimental data at key benchmark sites across the country using the four main types of SWMs with systematically varied acquisition parameters. Meaningful dispersion uncertainty will be evaluated for each set of acquisition parameters using newly-proposed methods. Intra-method variability in dispersion estimates will then be examined and the set of parameters with the lowest uncertainty selected to anchor the development of standards. With meaningful estimates of dispersion uncertainty, Monte Carlo inversions will be used to establish confidence intervals for Vs layer thickness and velocity, resulting in fully probabilistic results that can be incorporated into subsequent performance-based analyses. Following this step, inter-method variability will be examined in order to evaluate bias between various SWMs and borehole Vs measurements.