 
          1077
        
        
          One-dimensional cracking model in clayey soils
        
        
          Modélisation unidimensionnel de la fissuration des sols argileux
        
        
          Ávila G.
        
        
          
            Universidad Nacional de Colombia, Bogotá, Colombia
          
        
        
          Ledesma A., Lloret A.
        
        
          
            Universitat Politècnica de Catalunya (UPC) – BarcelonaTech, Barcelona, Spain
          
        
        
          ABSTRACT: It is difficult to formulate a general model to capture cracking initiation and evolution, because there are multiple
        
        
          factors that influence these processes. In order to simplify the evaluation and get more insight on the phenomenon, a series of
        
        
          shrinkage tests were performed inducing one-dimensional primary cracking on remolded clay from Bogotá. The results allowed the
        
        
          proposal of a conceptual model that, based on mould shape factor, initial moisture content and evaporation rate, may predict time for
        
        
          crack initiation and identify the location and direction of primary and secondary cracks. Results indicate that boundary conditions play
        
        
          a significant role in cracking evolution, due to the restrictions imposed to the free shrinkage and to the homogeneous evaporation rate.
        
        
          The moulds used for soil desiccation, allowed the induction of predefined cracks when their shape factor values were greater than 1.5.
        
        
          In these cases the model is simple and may be considered as a model of cracking with one degree of freedom. When moulds have a
        
        
          shape factor between 1 and 1.5, the degrees of freedom increase dramatically leading to much more complex crack patterns.
        
        
          RÉSUMÉ : La fissuration des sols est un phénomène complexe contrôlé par de multiples facteurs qui rendent difficile la formulation
        
        
          d'un modèle qui permette de capturer l'initiation et l'évolution des fissures. Afin de simplifier l'évaluation des facteurs et d'améliorer la
        
        
          compréhension du phénomène, une série d'essais de retrait, induisant une fissuration primaire, ont été effectuées sur de l'argile
        
        
          remoulé de Bogotá. Les résultats ont permis de proposer un modèle simple qui, en fonction du coefficient de forme du moule, la
        
        
          teneur en eau initiale et le taux d'évaporation, peut prédire le moment d'initiation des fissures et identifier l'emplacement et
        
        
          l'orientation des fissures primaires et secondaires. Les résultats indiquent que les restrictions portant sur le libre retrait du sol et
        
        
          l'homogénéité de la vitesse d'évaporation jouent un rôle important dans l’évolution des fissures. Les moules qui ont permis l'induction
        
        
          de fissures prédéfinies étaient caractérisés par un facteur de forme supérieur à 1.5. Dans ces cas-là, le modèle est simple et peut être
        
        
          considéré comme un modèle de fissuration avec un degré de liberté. Lorsque les moules ont un facteur de forme compris entre 1 et
        
        
          1.5, le degré de liberté s'accroit considérablement, conduisant à des schémas de fissuration beaucoup plus complexes.
        
        
          KEYWORDS: Cracking, unsaturated soil, shrinkage test, Bogotá clay.
        
        
          1 INTRODUCTION
        
        
          Cracking has an important impact on soil behaviour because it
        
        
          affects aspects such as drainage, compressibility and strength.
        
        
          Desiccation processes, conditions of crack initiation, crack
        
        
          evolution and cracking patterns have been studied on different
        
        
          soils under distinct boundary conditions (Marinho, 1994, Miller
        
        
          et al. 1988, Lloret et al. 1998, Kodikara et al. 2000, Yessiller et
        
        
          al. 2000, Vogel et al. 2005, Ávila et al. 2005, Lakshmikantha
        
        
          2009, Peron et al. 2009, Tang et al. 2010, Lakshmikantha et al.
        
        
          2012). When drying test on soils are conducted in square or
        
        
          circular moulds, the cracking patterns are in general complex
        
        
          and very difficult to predict. Nevertheless, when the shape of
        
        
          the moulds force a predominantly one-dimensional contraction,
        
        
          cracks tend to appear in a systematic pattern and it is possible to
        
        
          make predictions about the place and direction of the primary,
        
        
          secondary and in some cases, tertiary cracks. One-dimensional
        
        
          conditions lead to a very favourable situation to get some
        
        
          insight in the apparently erratic evolution of the cracking
        
        
          patterns.
        
        
          This article shows the experimental results obtained in
        
        
          desiccation tests performed on clay samples from Bogotá city
        
        
          where cracking sequences were observed. From these results a
        
        
          simple one-dimensional cracking conceptual model is proposed
        
        
          based on contraction restrictions and soil tensile strength.
        
        
          This work is a part of a comprehensive research program to
        
        
          study the problems of soil shrinkage and cracking that affect
        
        
          Bogotá city.
        
        
          2 CRACKING SEQUENCE DURING DESSICATION
        
        
          TESTS ON BOGOTÁ CLAY
        
        
          Geological and geotechnical characteristics of the subsoil of
        
        
          Bogotá and the effects of surface deformation and cracks on the
        
        
          infrastructures have been discussed by some authors (Lobo-
        
        
          Guerrero et al. 1992, Ingeominas 1996, Ingeominas and Los
        
        
          Andes University 1997, Ávila 1998, Ávila 2003, Ávila et al.
        
        
          2005). The test described in this paper have been done on
        
        
          samples of   Bogotá clay taken between 2 and 4 m depth (w
        
        
          L
        
        
          =
        
        
          62-65%, PI = 30-35% and Activity = 0.52-0.57). These samples
        
        
          are representative of the layers subjected to desiccation and
        
        
          cracking in many sectors of the urban area.
        
        
          Although different type of desiccation tests were conducted
        
        
          for the comprehensive evaluation of shrinkage and cracking of
        
        
          this clay, the discussion presented in this paper is basically
        
        
          focused on the cracking sequence observed in the experiments
        
        
          made in the double T shaped moulds shown in Figure 1. These
        
        
          moulds impose restriction to shrinkage and induce systematic
        
        
          cracking patterns. Those patterns are particularly interesting for
        
        
          the analysis of the process of formation and propagation of
        
        
          initial cracks and the subsequent process of cracking.
        
        
          Reconstituted clay samples were prepared at different initial
        
        
          water content and they were left to dry to an open atmosphere to
        
        
          observe the characteristics of the cracking process evolution.
        
        
          Figure 2 shows three stages of the desiccation process that
        
        
          were observed on three different tests (labeled 1 to 3 in Figure
        
        
          2) conducted at the same time and under equal atmospheric
        
        
          conditions. First sequence corresponds to a picture taken 4
        
        
          hours after the initiation of the desiccation process for which
        
        
          moisture contents of the samples ranged between 40% and