Lessons learned from VELACS project

Abstract

The VELACS project was conceived to examine the ability of current numerical codes with pore pressure development/liquefaction capability, to function as real-life design tools, with some simplifications. In common practice, a site is selected by a building owner, a geotechnical investigation is invited, boreholes are drilled, tests are made, and geophysics is pondered. Eventually, the soil profile and properties are known, the active faults are distinguished and earthquakes hypothesized (this is an ideal example), the structure postulated and the final, awful question remains: how will the site and structure respond to the putative input motion? The magical computer is invoked to model site and structure, which are then tweaked with the input ground motion (a lot of problems are side-stepped in this imaginary construction) to give by computation, the soil and structure motion and stress history. On this basis the design is modified and proceeds. But how correct is the result obtained thus, particularly if liquefaction is an issue? For most codes, at present, we do not know the answer to this, because the real-life, full-scale data are lacking. VELACS proposed to give at least a partial response to this question by substituting centrifuge for real-life, but requiring the code calculations to represent the real situation by being performed in advance of the test event. This is known in the trade as a "Class A prediction." VELACS project has now run its course; the predictions were done, the centrifuge tests performed, and the comparisons made. The result of the exercise is not uncomplicated, and the purpose of this paper is to describe some of the things we have learned from the project.

Additional details