An automated system for determining soil parameters: Case study

Graz University of Technology, Graz, Austria

^* Corresponding author: islam.marzouk@tugraz.at

Abstract

The success of numerical analysis depends on several factors. One of the keys factors is the accurate determination of constitutive model parameters. Determining these parameters from in-situ tests has several advantages compared to laboratory tests, from lower costs to minimal disturbance of the soil. However, it is not possible to derive soil parameters directly from in-situ tests results, since correlations are required. The literature offers a wide range of correlations which increases the uncertainty during interpretation. The ongoing research project APD (Automated Parameter Determination) investigates the possibilities of automated parameter identification from in-situ tests using a graph-based approach. In the present paper, existing correlations – developed for cone penetration tests - are validated by comparing their output to laboratory results. The Norwegian GeoTest Sites (NGTS) infrastructure project consists of five tests sites in different soils in Norway. The data from the soft clay site located in Onsoy, south-eastern Norway was used in the validation process. A web-based application “Datamap” that has been developed to capture, organize, and classify geotechnical research data has been used to obtain in-situ and laboratory tests data. The further validation of existing correlations to derive accurate constitutive parameters from in-situ tests is part of ongoing research.