1637
journal of materials research and technology 2021;14:1630 e 1643
of the experimental value. The sigma8 and sigma4 models were inadequate for the final loading steps and in the range of 60 e 90%of P max , respectively. The multi-linear models showed the possibility of almost exact adjustments, with the greatest difference being registered in Mult_sigma7.6, for a load range close to 75% of P max . Non-linearity is essential in the model for the behavior of wood-based products to bending and its effects are strongly amplified for load levels closer to the maximum. The experimental results of this study corroborate the adoption of models in which the non-linearity is represented by straight lines, in the so-called multi-linear model. For the first stage, which consists of defining the straight-line segment coinciding with the elastic-linear phase, based on the tests, we adopted the validity for 60% of the maximum bending load. However, other experimental researches show that this point can be in a rangeof 40 e 60% of the maximum load [10,45]. This difference can be explained by the nature of this composite, whose me- chanical and physical behavior depends on several factors, such as the type and content of adhesive wood species, and panel densification.
Fig. 10 e Step of maximum load versus deflection with non- linearity applied to the layered model (a) Longitudinal (b) Transverse.
3.1.
Modeling with non-linear behavior
The flexural rigidity analysis was based on the evaluation of how representative each model implemented for the panel behavior was, to the experimental results of load versus deflection. For the longitudinal direction (Fig. 9(a)), the purely elastic-linear model (Elastic-linear) resulted in behavior that can be accepted up to a certain level of the maximum load (approximately 80% of P max ), followed by the results of a very rigid model with maximum displacement 79.8% lower than the experimental one. The results obtained with the sigma20 model were very similar to those obtained with the Elastic- linear model, not representing the behavior for more advanced step loads. The sigma10 model was adjusted to represent the maximum experimental displacement value. On the other hand, there was a range in which the displace- ments were greater than the experimental ones, this was observed for loading from approximately 60% of P max . Both the Mult_sigma15 and Mult_sigma19.7 models were very representative of the material ' s behavior, being virtually equal for all step loads. Of these models, it must be considered that Mult_sigma19.7 benefits from simpler implementation procedures to the stress e strain relationship. For the transverse direction, the behavior of each model (Fig. 9(b)) was remarkably similar to those when applied in the longitudinal direction, in which the maximum vertical displacement obtained with the Elastic-linear model was 81.3%
Fig. 11 e Maximum bending stress for SL specimen - case with a single-layer [MPa]: (a) tension in direction X (b) compression in direction X (c) tension in direction Z (d) compression in direction Z.
Made with FlippingBook - Online magazine maker