Modeling Soils' Failure Envelope with Virtual Displacements (LSVD)

TL;DR

This paper introduces the LSVD method for accurately modeling soil failure envelopes by finding the best common tangent line among multiple samples, even with measurement errors, and compares it with existing methods.

Contribution

The LSVD method provides an iterative approach to determine the shear strength envelope for various soil conditions, adaptable to different models and robust against measurement noise.

Findings

LSVD effectively finds the common tangent line among soil samples.

Compared with p-q and CTPAC methods, LSVD shows improved robustness.

The method can be adapted to different failure envelope models.

Abstract

This work proposes the Least Squares with Virtual Displacements (LSVD) method to obtain the shear strength of n soil samples for the cohesionless, frictionless and mixed resistance conditions. Finding a common tangent line of more than two samples may not have an exact solution, especially if there are measurement errors. LSVD is an iterative method to find the best common tangent line or function to the n soil samples. There are soils that can be modeled with a linear failure envelope but the LSVD can be formulated in different forms, adapting itself to other criterion too, like the logarithmic failure envelope. The tested data is analyzed with LSVD and compared between other commonly used methods, like p-q method and CTPAC. The soils' samples were subjected to random noise to simulate errors from measurements to study the method's results. Also with that analysis a common reference point can be found to compare the methods given that all of them take different points of the Mohr Circles.