diff options
author | Jaron Kent-Dobias <jaron@kent-dobias.com> | 2020-01-30 13:06:58 -0500 |
---|---|---|
committer | Jaron Kent-Dobias <jaron@kent-dobias.com> | 2020-01-30 13:06:58 -0500 |
commit | ee932a158bafd9ea0f6c1bf1de4a3140377bbf3c (patch) | |
tree | a39385e7ba88d8ea3c5991d8cd5af2593cb56cf6 /poster_aps_mm_2020.tex | |
download | aps_mm_2020-ee932a158bafd9ea0f6c1bf1de4a3140377bbf3c.tar.gz aps_mm_2020-ee932a158bafd9ea0f6c1bf1de4a3140377bbf3c.tar.bz2 aps_mm_2020-ee932a158bafd9ea0f6c1bf1de4a3140377bbf3c.zip |
initial commit, text still from old statphys poster
Diffstat (limited to 'poster_aps_mm_2020.tex')
-rw-r--r-- | poster_aps_mm_2020.tex | 223 |
1 files changed, 223 insertions, 0 deletions
diff --git a/poster_aps_mm_2020.tex b/poster_aps_mm_2020.tex new file mode 100644 index 0000000..7451fff --- /dev/null +++ b/poster_aps_mm_2020.tex @@ -0,0 +1,223 @@ +% +% poster_05.2014.tex - description +% +% Created by on Tue May 13 12:26:55 PDT 2014. +% Copyright (c) 2014 pants productions. All rights reserved. +% +\documentclass[portrait]{a0poster} + +\usepackage[utf8]{inputenc} +\usepackage[T1]{fontenc} +\usepackage[]{amsmath} +\usepackage{amssymb,latexsym,mathtools,multicol,lipsum,wrapfig} +\usepackage[font=normalsize,labelfont=bf]{caption} +\usepackage{tgheros} +\usepackage[helvet]{sfmath} +\usepackage[export]{adjustbox} +\renewcommand*\familydefault{\sfdefault} + +\mathtoolsset{showonlyrefs=true} + + +\setlength\textwidth{194pc} +\begin{document} + +\setlength\columnseprule{2pt} +\setlength\columnsep{5pc} +\renewenvironment{figure} + {\par\medskip\noindent\minipage{\linewidth}} + {\endminipage\par\medskip} + + \renewcommand\section[1]{ + \vspace{3pc} + \noindent\huge\textbf{#1}\large + \vspace{1.5pc} + } + + \newcommand\unit[1]{\hat{\vec{#1}}} + \renewcommand\vec[1]{\boldsymbol{\mathbf{#1}}} + \newcommand\norm[1]{\|#1\|} + \def\rr{\rho} + \newcommand\abs[1]{|#1|} + \def\dd{\mathrm d} + \def\rec{\mathrm{rec}} +\def\CC{{C\kern-.05em\lower-.4ex\hbox{\cpp +\kern-0.05em+}} } +\font\cpp=cmr24 +\def\max{\mathrm{max}} + + \noindent\hspace{177pc}\includegraphics[width=18pc]{CULogo-red120px.eps} + \vspace{-24.5pc}\\ + \Huge \textbf{Elastic properties of hidden order in URu$_{\text2}$Si$_{\text2}$ are reproduced by\\ staggered nematic order} + \bigskip\\ + \huge \textbf{Jaron~Kent-Dobias, Michael Matty \& Brad J Ramshaw} + \vspace{1pc} + + + \begin{multicols}{2} + + \section{Introduction} + \Large + + \begin{wrapfigure}{r}{.25\textwidth} + \centering + %\includegraphics[width=0.25\textwidth]{imgs/crack.jpg} + \caption{Cracking in concrete.} + \end{wrapfigure} + + Understanding material cracking and fracture is necessary for + understanding the aging and failure of those materials in our buildings + and infrastructure. In ordinary brittle materials like glass, stress at + the tip of a crack causes it to quickly and cleanly propagate through the + material. In ductile materials like metals, this stress is reduced by + plastic deformation around the crack tip, forming the crack's + \textbf{process zone}. In quasi- (or disordered) brittle materials like + concrete, this stress is reduced by opening a complicated network of + microcracks in the process zone. This makes the structure of the + quasibrittle process zone and crack propagation + difficult to study by ordinary means. + + + \section{Fuse Networks} + + \begin{wrapfigure}{l}{.25\textwidth} + \centering + %\includegraphics[width=0.12\textwidth]{imgs/square_network.pdf} + %\includegraphics[width=0.12\textwidth]{imgs/square_high_beta.pdf}\\ + %\includegraphics[width=0.12\textwidth]{imgs/voronoi_network.pdf} + %\includegraphics[width=0.12\textwidth]{imgs/voronoi_high_beta.pdf} + \captionof{figure}{Contrasting the square (top) and voronoi (bottom) networks. {\bf Left:} Unbroken fuse networks. {\bf Right:} A + fracture surface in each at low disorder ($\beta=10$).} + \label{nets} + \end{wrapfigure} + + \Large + + We model quasibrittle cracking and fracture using simulations of fuse + networks, which are electrical systems of resistive fuses whose random + current thresholds $t$ are cumulatively distributed by $t^\beta$. $\beta$ + parameterizes the amount of disorder in the system: large + $\beta$ corresponds to vanishing disorder, small $\beta$ to very large + disorder. Fracture is performed adiabatically: the fuse whose ratio of + current to threshold is largest breaks, and the current across the + networks is recomputed. In order to reduce lattice effects, which become large for + small disorder (see Figure \ref{nets}), we use voronoi meshes for our fuse networks. + + + \section{Homogeneous Scaling} + + \begin{wrapfigure}{r}{0.25\textwidth} + \centering + %\includegraphics[width=0.25\textwidth]{imgs/ashivni.png} + \captionof{figure}{The `phase diagram' for fracture in homogeneous + systems.} + \label{ashivni} + \end{wrapfigure} + + \Large + + The problem of fracture in fuse networks was unresolved until recently. + For low disorder, fracture is nucleation-like, similar to that of ordinary + brittle systems. At large disorder, fracture occurs after a very large + amount of uncorrelated damage, and appears percolation-like. + Sethna and Shekhawat developed a theory which unifies these behaviors with + an {\sc rg} crossover at intermediate disorder characterized by mean-field + avalanches. The percolation-like behavior at high disorder was shown to + be unstable under course-graining, and therefore any nonzero $\beta$ + will cause nucleated fracture at a sufficiently large system size (see + Figure \ref{ashivni}). + + In unpublished work, Shekhawat and Sethna found a scaling form for the + distribution of network strengths $\sigma_\max$, the largest current + applied to the network before it has broken. It is given by + \begin{equation} + P(\sigma_\max\mid\beta, L,u)=\sigma_\max^{-\tau_\sigma}\mathcal P(\beta + L^{1/\nu_f},\sigma_\max L^\delta,uL^{-\Delta/\nu_f})\label{form} + \end{equation} + where $\tau_\sigma$, $\nu_f$, $\delta$ and $\Delta$ are universal + exponents, $L$ is the system size, and $u$ is an irrelevant scaling + variable. + + \begin{figure} + \vspace{1pc} + \centering + %\includegraphics[width=0.325\textwidth]{imgs/voronoi_homo_big.pdf} + %\includegraphics[width=0.325\textwidth]{imgs/voronoi_homo_med.pdf} + %\includegraphics[width=0.325\textwidth]{imgs/voronoi_homo_lit.pdf}\\ + \begin{minipage}[c]{0.325\textwidth}\centering$\pmb{\beta=3}$\end{minipage} + \begin{minipage}[c]{0.325\textwidth}\centering$\pmb{\beta=0.5}$\end{minipage} + \begin{minipage}[c]{0.325\textwidth}\centering$\pmb{\beta=0.03}$\end{minipage}\\ + \captionof{figure}{Fractured fuse networks at various $\beta$. Each colored region shows a + contiguous cracked cluster. The black region shows the surface of + the spanning crack.} + \label{homo} + \vspace{1pc} + \end{figure} + + + \section{Scaling in the Process Zone} + + \begin{figure} + \centering + %\includegraphics[width=0.245\textwidth]{imgs/sample_16.pdf} + %\includegraphics[width=0.245\textwidth]{imgs/sample_32.pdf} + %\includegraphics[width=0.245\textwidth]{imgs/sample_64.pdf} + %\includegraphics[width=0.245\textwidth]{imgs/sample_128.pdf}\\ + \begin{minipage}[c]{0.245\textwidth}\centering$\pmb{L=16,\;\beta=1.9}$\end{minipage} + \begin{minipage}[c]{0.245\textwidth}\centering$\pmb{L=32,\;\beta=1.2}$\end{minipage} + \begin{minipage}[c]{0.245\textwidth}\centering$\pmb{L=64,\;\beta=0.78}$\end{minipage} + \begin{minipage}[c]{0.245\textwidth}\centering$\pmb{L=128,\;\beta=0.5}$\end{minipage}\\ + \captionof{figure}{Notched fuse networks at critical stress. The + disorder for each system is tuned so that $\beta L^{1/\nu_f}$ is + constant, and the statistics of each should scale trivially.} + \vspace{1pc} + \label{notches} + \end{figure} + + \Large + We have made progress for developing a scaling theory of damage and stress + in the process zone of quasibrittle cracks. We have essentially taken the + scaling behavior of \eqref{form} as an ansatz for how the corresponding + qualities of a critically semi-cracked network should scale. We have + begun demonstrating the validity of this theory. Figure \ref{notches} + shows critically cracked networks at constant $\beta L^{1/\nu_f}$, an + invariant scaling combination under this theory. As can be seen in Figure + \ref{collapse}, the disorder-averaged stress profiles caused by each + collapses nicely. + + \begin{figure} + \centering + %\includegraphics[width=0.5\textwidth]{imgs/legends.pdf}\\ + \vspace{-.5em} + %\includegraphics[width=0.505\textwidth,valign=t]{imgs/sample_collapse_1.pdf}\hfill + %\includegraphics[width=0.48\textwidth,valign=t]{imgs/sample_collapse_2.pdf}\\ + \vspace{-.5em} + + \captionof{figure}{Disorder-averaged stress $\sigma$ as a function of + distance $x$ from the tip of a critical crack. $\beta L^{1/\nu_f}$ is + constant for each curve. \textbf{Left:} The unmodified stress. + \textbf{Right:} The stress collapsed.} + \label{collapse} + \end{figure} + + \section{Next Steps} + + \begin{wrapfigure}{r}{0.25\textwidth} + \centering + \vspace{-2em} + %\includegraphics[width=.25\textwidth]{imgs/voronoi_heir.pdf} + \captionof{figure}{A hierarchical voronoi lattice.} + \end{wrapfigure} + + \Large + + The voronoi networks we are able to generate allow us great flexibility + for future multiscale computational modelling. Once we have hashed out + our scaling theory more thoroughly, we plan on using it to probe much + larger systems than previously feasible using networks whose fuse density + becomes smaller in regions of less importance. This should allow us to + see cleaner stress and damage scaling and cleanly stitch our discrete system to a + continuum approximation. + + \end{multicols} +\end{document} + |