Osterberg Lecture 2018: David Muir Wood

On May 22nd, 2018, the Civil and Environmental Engineering Department hosted David Muir Wood as part of our annual Osterberg Lecture. David Muir Wood visited Northwestern for an extended period of time, concluding his visit with the seminar titled "Particle-Continuum Duality: Intermediate Scales and Representative Elemental Volumes." Please see the video below for the full lecture. 

Abstract

Soils are particulate materials but it is convenient to describe them using familiar continuum concepts of stress and strain. What is the size of an appropriate representative volumetric element? It must be larger than the median particle size. There are various patterns of particle arrangement within granular assemblies which have dimensions 10d50 − 30d50. We cannot determine stresses within real soils. However, studies of transmission of stress within assemblies of photoelastic discs show chains of heavily loaded particles and adjacent particles carrying no load. The heavily loaded chains form a network with a cell size typically 5d50 − 10d50.
Radiographic studies of shearing of sand show dilation bands of typical width 10d50 − 20d50 forming rather regular networks with cell size typically ~ 30d50. As a shear band starts to develop there will be an activation length over which the mobilised friction falls from a peak towards the critical state – this length is typically around 150d50 − 200d50, which may be long compared with elements of a failure mechanism developed in a small laboratory or centrifuge model test, but insignificant compared with a prototype failure.
If we reckon that a typical sample size should cover 10-20 of our identified cells, then that sample size should be around 300 − 600d50. Hostun sand has particle size around 0.3 mm; the required sample dimension is then 90 – 180 mm, which is of the same order of magnitude as some typical test samples.
A triaxial cell provides a three-dimensional loading environment. Studies of internal strains indicate a complex pattern of mechanisms which may combine to give a sort of axially symmetric response though the mechanisms themselves are in no way axially symmetric. What is the ideal laboratory test from which to collect data with which to challenge our candidate constitutive models.

Biography

Soils are particulate materials but it is convenient to describe them using familiar continuum concepts of stress and strain. What is the size of an appropriate representative volumetric element? It must be larger than the median particle size. There are various patterns of particle arrangement within granular assemblies which have dimensions 10d50 − 30d50. We cannot determine stresses within real soils. However, studies of transmission of stress within assemblies of photoelastic discs show chains of heavily loaded particles and adjacent particles carrying no load. The heavily loaded chains form a network with a cell size typically 5d50 − 10d50.
Radiographic studies of shearing of sand show dilation bands of typical width 10d50 − 20d50 forming rather regular networks with cell size typically ~ 30d50. As a shear band starts to develop there will be an activation length over which the mobilised friction falls from a peak towards the critical state – this length is typically around 150d50 − 200d50, which may be long compared with elements of a failure mechanism developed in a small laboratory or centrifuge model test, but insignificant compared with a prototype failure.
If we reckon that a typical sample size should cover 10-20 of our identified cells, then that sample size should be around 300 − 600d50. Hostun sand has particle size around 0.3 mm; the required sample dimension is then 90 – 180 mm, which is of the same order of magnitude as some typical test samples.
A triaxial cell provides a three-dimensional loading environment. Studies of internal strains indicate a complex pattern of mechanisms which may combine to give a sort of axially symmetric response though the mechanisms themselves are in no way axially symmetric. What is the ideal laboratory test from which to collect data with which to challenge our candidate constitutive models.
David Muir Wood obtained his PhD at Cambridge University in 1974 performing true triaxial tests on kaolin clay under the supervision of Peter Wroth. He held a lectureship at Cambridge until 1987 when he took up the Cormack Chair of Civil Engineering at Glasgow University. Having served as Head of Department of Civil Engineering and Dean of the Faculty of Engineering he took up the Chair of Civil Engineering at Bristol University in 1995 - and again served as Head of Department of Civil Engineering and Dean of the Faculty of Engineering before retiring in 2009 and moving to a half-time Chair of Geotechnical Engineering at the University of Dundee. He is now Emeritus Professor at both Bristol and Dundee. Since retiring he has held an affiliated professorship at Chalmers Technical University, Gothenburg, Sweden and visiting professorships at Politechnic of Milan and universities of Western Australia, Dresden, Innsbruck, Yokohama, and Northwestern University. His research has covered laboratory testing of soils in different apparatus and the development of adequately complex constitutive models. He is presently part of a collaborative project 'Rooting for sustainable performance' with Universities of Dundee, Southampton, Aberdeen studying the interaction of roots and soils in laboratory tests and centrifuge models, and developing models to describe the deformation of soil-root composites. He delivered the 20th Bjerrum Lecture in Oslo in 2005 and the 4th Bishop Lecture in Seoul in 2017. He has published 4 books: 'Soil behaviour and critical state soil mechanics' (1990), 'Geotechnical modelling' (2004), 'Soil mechanics: a one-dimensional introduction' (2009), 'A very short introduction to civil engineering' (2012).