000016043 001__ 16043
000016043 005__ 20161115135336.0
000016043 04107 $$aeng
000016043 046__ $$k2013-06-12
000016043 100__ $$aAsteris, P.
000016043 24500 $$aNeural Network Approximation of the Masonry Failure under Biaxial Compressive Stress

000016043 24630 $$n34.$$pComputational Methods in Structural Dynamics and Earhquake Engineering
000016043 260__ $$bNational Technical University of Athens, 2013
000016043 506__ $$arestricted
000016043 520__ $$2eng$$aMasonry is a material that exhibits distinct directional properties because the mortar joints act as planes of weakness. To define failure under biaxial stress, a threedimensional surface in terms of the two normal stresses and shear stress, or the two principal stresses and their orientation to the bed joints, is required. Researchers have long been aware of the significance of the bed joint angle to the applied load and many experimental tests have been carried out on brick masonry discs to produce indirect tensile stresses on joints inclined at various angles to the vertical compressive load. The highest strength of the masonry has been observed for the cases when the compressive load was perpendicular to the bed joints or when the principal tensile stress at the center of the disc was parallel to the bed joints. In this case failure occurred through bricks and perpendicular joints. The lowest strength has been observed when the compressive load was parallel to the bed joints or when the principal tensile stress at the center of the disc was perpendicular to the bed joints. In this case failure occurred along the interface of brick and mortar joint. In the present study, the preliminary results of an ongoing research on the failure of brittle anisotropic materials are presented. In particular, Neural Networks (NNs) are used in order to approximate the experimental failure curves of a brittle anisotropic material such as masonry, that has been investigated in depth by Page [1]. For each angle θ (0º, 22.5º, 45º), a Neural Network is trained with the experimental data of Page as inputs. Then the NN is asked to produce the whole failure curve for each angle as its output, filling also the gaps between the experimental points with appropriate approximations. The results show the great potential of using NN for the approximation of the masonry failure under biaxial compressive stress.

000016043 540__ $$aText je chráněný podle autorského zákona č. 121/2000 Sb.
000016043 653__ $$aMasonry, anisotropy, failure criterion, failure surface, biaxial stress, Neural Network, NN, approximation.

000016043 7112_ $$aCOMPDYN 2013 - 4th International Thematic Conference$$cIsland of Kos (GR)$$d2013-06-12 / 2013-06-14$$gCOMPDYN2013
000016043 720__ $$aAsteris, P.$$iPlevris, V.
000016043 8560_ $$ffischerc@itam.cas.cz
000016043 8564_ $$s858578$$uhttps://invenio.itam.cas.cz/record/16043/files/2158.pdf$$yOriginal version of the author's contribution as presented on CD, section: CD-RS 27 SOFT COMPUTING APPLICATIONS
.
000016043 962__ $$r15525
000016043 980__ $$aPAPER