Numerical Analysis of Compression Perpendicular to the Grain in Glulam Beams With and Without Reinforcement

Abstract: In Sweden there has recently been a change in the code from the former BFS2010:2 BKR 13 to Eurocode 5. This shift has led to changes when calculating the design capacity for all materials including wood. One of the parameters that were changed was the strength of wood perpendicular to the grain. The characteristic value required was decreased from 8 MPa to 2.7 MPa which have a relativity large impact on the size of the contact area, for instance between a beam and a support. A consequence of this is that the support has to be larger and in many situations more expensive. Discussions with SWECO Structures, the Division of Structural Engineering and the Division of Structural Mechanics led to to believe that the decrease in characteristic capacity is not fully motivated for all loading situations. Besides the evaluation of the decrease, an investigation of how reinforcement in form of wooden dowels and threaded steel screws affect the compression capacity perpendicular to the grain. The analysis of the compression capacity perpendicular to the grain, with and without reinforcement, is based on analytical models and by means of numerical calculations by a Finite element program called ABAQUS. To capture the plastic non linear behavior perpendicular to the grain two material models were combined, one which handles the stress in longitudinal direction and one handling the stresses in radial and tangential directions. Parallel to this thesis another Master's thesis is produced with the same objective but by a different method, laboratory testings. This is advantageous since it makes it possible to judge the creditability of the FE-models. A total number of 16 FE-models were produced with different loading situations and setups to investigate how the support length and reinforcement effect the stress capacity perpendicular to the grain. When it was evident that the unreinforced models captured the real behavior 4 new models were produced with support length 100-400 mm to see how the stiffness, maximal capacity and deformations depends on the support length. The results of the unreinforced FE-models show that the decrease from 8 MPa to 2.7 MPa is motivated in some cases where small deformations only are acceptable and with a support length > 400 mm. However, if a deformation of 20 mm is acceptable and the support length is 100 mm, the compression capacity is above 8 MPa according to the FE-results. The FE-models handling wooden dowels captured the behavior very well in the elastic area when comparing the models with laboratory tests. When the specimen began to crack the dowels buckled. This phenomenon has not been taken into account in this thesis since it is complicated and time consuming modeling crack propagation. The models with threaded steel rods were dicult to model even in the elastic area due to the fact that the reinforcing rods were irregularly pushed in from the bottom side of the beam. The result shows that there is an advantage in stiffness if the rods are in fush with the bottom side of the beam because they could take load simultaneously. The mean capacity for the supports with wooden dowels was calculated to 11 MPa and for the threaded steel rods 15 MPa compared to 5 MPa unreinforced.