SAMPLING SUFFICIENCY FOR MECHANICAL PROPERTIES OF WOOD

Abstract

Based on most recently published studies, there is a large variability in both the mechanical properties of wood and sample sizes selected to evaluate them. This study aims to define sampling sufficiency for some mechanical properties of wood, which were bending strength, bending modulus, compressive strength, compressive modulus, hardness, and shear strength. The mechanical tests were carried out according to an American standard procedure on wood samples cut from clonal Eucalyptus planted in southern Brazil. Sampling sufficiency was determined by an intensive computational method based on resampling of original data using Monte Carlo simulations. The experimental tests data conformed to the normal distribution and most of the obtained sufficient sample sizes determined by Monte Carlo simulation were above those sample sizes used in most already published studies. Furthermore, properties related to wood stiffness presented smaller variabilities than their respective properties associated with wood strength, leading to smaller sample sizes for the former cases.