Documents

A Contribution to the Theoretical Prediction of Life-time in Glass Structures

Categories
Published
of 20
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Description
Theoretical Prediction of Life-time in Glass Structures
Transcript
  A Contribution to the Theoretical Prediction of Life-time in Glass Structures Manuel Santarsiero Civil Engineer University of Pisa, Pisa, Italy Manuel Santarsiero, born 1985, received his civil engineering  bachelor degree with honors from the Univ. of Pisa in July 2009, with a thesis on “  Proposal for a  New Failure Prediction Method  for Glass Structures  ” .  manuel.santarsiero@gmail.com Maurizio Froli Professor Dep. of Civil Engineering, Structural Division University of Pisa, Pisa, Italy Maurizio Froli, born 1954, received his civil engineering degree from the Univ. of Pisa. He is presently Scientific Responsible of the Laboratory for Testing Materials and Structures of the University of Pisa m.froli@ing.unipi.it Summary In order to assess safety levels in glass structures a scattered and inhomogeneous variety of mostly complicated resistance criteria is presently available, very often requiring specially developed softwares. For this reason engineers who wants to assess with reliability the actual safety level of glass structures of relevant economical importance are still obliged to undertake expensive experimental tests. In the attempt to overcome this problem, it was formulated a new semi-probabilistic failure  prediction method called  Design Crack Method  ” (DCM), which is a compromise between the necessity to accurately model the complex mechanical behaviour of glass at breakage and the need to reduce the analytic complexity of the calculations. On the basis of Linear Elastic Fracture Mechanics, such aim has been reached in the present work by defining a new quantity called  Design Crack  , characterized by a mathematical expression that depends only on the probability of failure and on the surface damaging level. The proposed method, which is in accordance with the basic principles of the Structural Eurocodes, allows to predict glass lifetime taking into due account the influence of parameters like the surface extension and the loading time-history of the structural element. The results obtained in some applications with the D.C.M. method have been numerically compared in this paper with those of the existing most frequently used theoretical methods. Keywords: glass strength, design crack, fracture mechanics, static fatigue, surface flaws, life time,  probabilistic model, analytical method.    1 Introduction Glass is increasingly used in modern Architecture to resolve non secondary structural tasks. As a result there has risen an urgent demand for a reliable, theoretical method to perform engineering assessments of the remaining lifetime in glass structure. The availability of such a method would permit a safe design of glass structures and the prediction of their mechanical behaviour even if, like presently, the data base of experimental results obtained on full scale structures is not yet sufficiently extensive. The peculiar sensitivity of glass to surface imperfections, the absence of plasticity and phenomena like the so called “static fatigue” for example and the diminished resistance of the material to long term loads prevents the use of the traditional calculation methods that apply for other building materials. Also standard safety assessment procedures specifically developed for windows cannot be employed for glass beams or glass floors because window glasses must prevalently face short time loadings such as wind gusts. The first scientifically consistent, but engineering oriented, theoretical model of glass strength was formulated in 1972 by Brown [1] who called it “Load Duration Theory” (LDT). Brown combined the static fatigue theory of Charles & Hillings [2] with the concept of failure probability expressed  by Weibull [3]. In 1974 Evans [4] developed the “Crack Growth Model” (CGM) on the basis of the principles of Linear Fracture Mechanics. This method makes use of the empirical description of the sub-critical  propagation of cracks (deduced from the experimental relationship between crack growth speed and stress intensity factor K  I ) together with the Weibull failure probability concept under the hypotheses that in all surface micro cracks a sub-critical crack growth takes place no matter how little their stress intensity factor K  I  is. Between 1980 and 1984 Beason and Morgan formulated the “Glass Failure Prediction Model” (GFPM) [5] devoted to the safety assessment of rectangular glass plates simply supported along their borders. Here also the static fatigue theory of Charles & Hillings is used together with the Weibull concept of failure probability. Calculated stresses include geometrical non linearities caused in thin plates  by large deflections. The GFPM constitutes the theoretical fundament of Canadian Standard CAN/CGSB 12.20-M89 and of ASTM 1300-04.  Nevertheless, Fisher-Cripps & Collins [6] demonstrated in 1994 that the GFPM was not able to  predict glass cracks under short time not under long time loads. However, the CGM also demonstrated a lack of capability of predicting fractures under long time loads. In order to obviate this problem, the same Authors introduced in CGM an additional, experimentally based condition which states that any crack growth occurs if the K  I  factor is less than a limit value called Static Fatigue Limit. The new approach was called Modified Crack Growth Model (MCGM) [6]. Between 1995 and 1999 Sedlacek and Others [7] developed an engineering method to check structural safety in glass structures based on Weibull failure probability and on Linear Fracture   Mechanics. The final verification expression obtained by Sedlacek is formally similar to the Miner rule of progressive damaging, commonly used by designers to assess the remaining fatigue life of steel structures. In 2001 Shen [8] and Siebert [8] proposed their own versions of Sedlacek method, both not very different from the srcinal formulation and giving comparable results. The Sedlacek approach is presently at the basis of Eurocode prEN 13474-3. In the same year Porter [10] elaborates the Crack Size Design (CSD) where he defines a “design crack”, i.e. a maximum design depth that surface cracks, supposed uniformly distributed on a glass  pane, may reach before failure. In 2006 Haldimann gave an important contribution to the solution of this problem with his Lifetime Prediction Model (LPM) where he avoids the introduction of equivalent quantities but calculates directly the failure probability of a glass element starting from the probability distribution of its defects and from the deterministic knowledge of loading time-history [11] . Devigili in the same year [12] removed also the conceptual limitation, in Haldimann’s theory, of the deterministic definition of the time history by introducing the hypothesis that the random properties of the surface micro-defects and of the loading time histories can be described by Markov’s  probability distributions. The results appear very rigorous but the associated calculation difficulties prevent the application of Devigili’s method for current engineering purposes. In this paper we tried to formulate a method for extending LPM but maintaining at the same time a moderate level of analytical difficulty in order to let its application remain possible also for normal design activities. 2 Basic Concepts about the Mechanical Behaviour of Glass Every glass surface, although apparently intact, is inevitably affected by microscopic randomly distributed cracks. When the glass element is subjected to mechanical stresses, high stress concentrations occur at the tip of the micro cracks which can not be plastically redistributed because of the amorphous crystalline structure of the material, lacking in preferential plastic-flow plans. This peculiar feature causes the typical brittle fractures that characterize this material. The fracture resistance of damaged elements can be analytic described by the principles of  Linear  Elastic Fracture Mechanics . For this reason Irwin [13] introduced the Stress Intensity Factor (K) , in order to describe the behaviour of brittle materials damaged by a single flaw placed perpendicularly to the stress direction (opening mode I): (2-1) Where : ã   Y - Shape factor, that depends on flaw’s geometry and dimension;    ã   ! ( t) - Time-history of the tensile stress near to the crack edge; ã   a(t) - Time-history of crack depth. Failure occurs when the propagation of the crack becomes unstable; that happens when: (2-2) K  IC  represents the Critical Stress Intensity Factor   which depends only on the kind of material and can be usually considered technically constant because of its low statistical spread. Substituting (2-2) in (2-1) easily allows to obtain the a cr    e cr analytic expressions, respectively representing the crack depth and the stress intensity able to induce unstable crack propagation. This pair of values identifies the so-called “ inert strength” . The graph of Fig. 2-1 shows, according to the K-factor, the flaw propagation velocity of a glass element subjected to constant stress during time and immersed in a humid environment. Although in section I the K value is much lower than K  IC , a slow sub-critical growth of flaws depth occurs on the glass surface which gradually reduces the inert tensile glass strength over time. This phenomenon is known as s tatic fatigue  and plays one of the main roles in theoretically determining the ultimate strength of glass structures. Moreover, the only part of Fig. 2-1 that  provides a significant contribution to the design life of a crack, which is submitted to stress intensification during time, is the section I because the failure of the glass plate occurs almost instantly when section III is reached. As shown in Fig.2-1, the v -K relation is represented by a constant slope curve on a bi-logarithmic plot and it can be analytically described  by the following differential equation: (2-3) n  being the curve’s slope in section I and v 0  the propagation velocity when K = K  IC . Both n and v 0  are generally dependent on relative humidity rate, temperature, pH and stress intensity. Fig. 2-1 –   versus K  I

PEST Analysis

Jul 23, 2017
Search
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks