VTI:s nationella bibliotekskatalog

Normalvy MARC-vy

Real-time evaluation of site characterization to direct and conclude exploration Graettinger, Andrew J ; Supriyasilp, Thanaporn ; Durrans, S Rocky

Av: Medverkande(n): Utgivningsinformation: Transportation Research Record, 2001Beskrivning: nr 1755, s. 51-9Ämnen: Bibl.nr: VTI P8167:1755Location: Abstrakt: A computationally efficient method to direct and evaluate the sufficiency of site characterization activities associated with roadway construction is demonstrated. The method, reliability-based exploration (RBE), combines three-dimensional subsurface analysis with the uncertainty associated with subsurface data to produce a quantifiable measure of exploration sufficiency, which is directly related to the probability of project success. RBE employs a first-order Taylor series expansion to combine the performance model sensitivity and input data uncertainty to calculate the variance in project performance in real time, which, given enough time and computer resources, could be calculated by conventional Monte Carlo (MC) simulations. The study shows that the Taylor series together with direct derivative coding (DDC), can generate the performance variance much faster than MC simulations. For the example presented, the Taylor series with DDC calculates results 160 times as fast as MC. The RBE method begins with a previous model of the subsurface that can include engineering judgment. As new information is collected at a site, the prior subsurface model is updated with new site data. The robustness and efficiency of RBE are illustrated with a practical example of exploration associated with two designs of a highway interchange. The case study illustrates the RBE approach by identifying the next location for exploration and evaluating characterization sufficiency. In addition, the effects of different designs on exploration are described. The example shows that design affects surface settlement, location of the next boring, and termination of exploration (number of required borings).
Exemplartyp: Rapport, konferenser, monografier
Bestånd
Aktuellt bibliotek Hyllsignatur Status Förfallodatum Streckkod
Statens väg- och transportforskningsinstitut Tillgänglig

A computationally efficient method to direct and evaluate the sufficiency of site characterization activities associated with roadway construction is demonstrated. The method, reliability-based exploration (RBE), combines three-dimensional subsurface analysis with the uncertainty associated with subsurface data to produce a quantifiable measure of exploration sufficiency, which is directly related to the probability of project success. RBE employs a first-order Taylor series expansion to combine the performance model sensitivity and input data uncertainty to calculate the variance in project performance in real time, which, given enough time and computer resources, could be calculated by conventional Monte Carlo (MC) simulations. The study shows that the Taylor series together with direct derivative coding (DDC), can generate the performance variance much faster than MC simulations. For the example presented, the Taylor series with DDC calculates results 160 times as fast as MC. The RBE method begins with a previous model of the subsurface that can include engineering judgment. As new information is collected at a site, the prior subsurface model is updated with new site data. The robustness and efficiency of RBE are illustrated with a practical example of exploration associated with two designs of a highway interchange. The case study illustrates the RBE approach by identifying the next location for exploration and evaluating characterization sufficiency. In addition, the effects of different designs on exploration are described. The example shows that design affects surface settlement, location of the next boring, and termination of exploration (number of required borings).

Teknik från Koha