Experimental characterization and numerical modeling of compression perpendicular to the grain in wood and cross-laminated timber

Abstract: Compression perpendicular to the grain (CPG) of wood is a typical loading situation in timber structures. It has been an extensively studied research topic for decades, due to the highly ductile behavior of wood under such loading, the large variations in mechanical properties, and the relevance of these properties in structural design. Among others, the main influencing factors for CPG properties are stressed volume, load and support configurations, and annual ring orientations to the loading direction. After the innovation of the massive, engineered wood based product, cross-laminated timber (CLT) and its application in high rise buildings, CPG of wood has gained further importance. The development of a non-homogeneous, undesired and combined stress state under CPG in solid wood, due to the material anisotropy in the radial-tangential plane, can build up a complex multi-axial stress state in CLT. As a comparatively new product, the study of the influencing factors for CPG properties of CLT, and an understanding of the local material behavior under such loading, is essential for product characterization and for the development of design guidelines to ensure safe and efficient design.The main aim of the doctoral thesis is to establish a relationship betweenthe anisotropic behavior of clear wood in the transverse plane and the structural response of CLT under CPG loading. Both experimental and numerical studies were adopted herein, to enhance the understanding of the basic material behavior and the product and structural behavior. On the clear wood scale, the focus was on developing a test setup for uniaxial and biaxial loading in the radial-tangential(RT) plane. The potential of the developed test setup for the biaxial testing in the transverse plane was exploited for the investigation of the moisture and time dependent behavior of clear wood under radial compression and rolling shear loading. For data acquisition, in addition to the force and displacement data measured by the internal actuators of the testing machine and an external load cell,a contact-free digital image correlation (DIC) system was used in the experimental investigations. A numerical model was developed, which can describe the elasto-plastic behavior of wood under compression in the transverse plane and predict the structural behavior of solid wood and CLT. For that purpose, a novel Quadratic multi-surface (QMS) failure criterion and a simplified Hoffman failure criterion were implemented in a user-subroutine in the finite element software Abaqus®, and their suitability was compared with the Abaqus implemented Hill’s criterion.The validation of the material models was based on the experimental investigations of failure behavior of clear wood under stress perpendicular to the grain with rolling shear interaction. The material models were further utilized to predict the structural response of solid wood and CLT wall-to-floor connections under CPG loading. The predicted response of CLT connections under CPG by using the above-mentioned material models was compared with experiments, which investigated the influences of different connection types, wall and floor thicknesses, positions of walls, and outer deck layer orientations. The models were then applied to investigate the influence of the pith location in the boards, the number of layers and the thickness of walls and the floor on the stiffness and strength of CLT connections. Moreover, the CLT connection’s rotational rigidity as a consequence of compressive force from the upper floor in a multi-story building was studied by means of finite element calculations.The DIC measured strain fields from the experiments on clear wood confirmed the dependence of strain field on the curvature of the annual rings. As regards the material models, Hill’s model resulted in significantly higher force carrying capacity than experiments on clear wood, whereas Hoffman’s and QMS models predicted reasonably well the force-displacement relationships as found in experiments. The Hoffman’s and QMS models predicted stiffness was about 5–10% higher than corresponding experimental results on clear wood, and about 25% higher for CLT connections. The higher difference in the latter case is due to the difference in material properties of clear wood and structural timber, and the contact behavior between the structural members. The results from CLT wall-to-floor connections revealed a strong influence of loading and supporting configurations, wall thickness and pith locations on their stiffness and strength. A compressive loading on the CLT wall showed a positive effect on the rotational stiffness of CLT wall-to-floor connections, which considerably reduces the CLT floor mid-span deflection in comparison to a simply-supported floor.The thesis work contributes to an enhanced understanding of the anisotropic material behavior of wood in the RT-plane and of its effects on structural timber and CLT under CPG loading. The outcomes of the thesis are beneficial to the product design and standardization of CLT and can be applied in further product development and in optimized structural design.