Evaluation of Models for Predicting Tensile Strength of 2- By 4-Inch Lumber

Several mathematical models offer potential grading criteria for tensile strength of lumber. Prediction efficiency, falldowns, and A/P ratio are used here to evaluate and compare the models. A common sample, consisting of 287 machine-rated 2 by 4’s. furnished these necessary statistics. The goal of and stress grading method is to maximize prediction efficiency without entailing more than 5 percent falldowns (pieces with actual strength, A, less than predicted strength, P, that is, where A/P is less than unity). Several models employing combinations of strength ratio, modulus of elasticity, specific gravity, or slope of grain yielded prediction efficiencies of 52 to 58 percent without entailing more than 5 percent falldowns. Those prediction efficiencies are up to 10 percent higher than that for the ASTM D 245 bending strength ratio model when the “surface” method is used to estimate bending strength ratio, but only up to 3 percent higher when the “displacement” method is used; however, falldowns exceeded 5 percent by the displacement method. The visual grading model, which is a stepwise version of the ASTM D 245 bending strength ratio displacement model, yielded a prediction efficiency of 45 percent with 3 percent falldowns. The machine grading stepwise model currently in western lumber grading rules yielded a prediction efficiency of 51 percent with 6 percent falldowns, but for one species group, falldowns were as high as 10 percent. Based on the A/P ratios, some predicted strengths were substantially higher than actual strengths even for models without excessive falldowns. The results of this study should be of prime importance to agencies that write grading rules for machine rated lumber. They should also be of interest to researchers in nondestructive testing of wood for strength and to engineers who design with wood. The results concerning visual grading should be of interest to members of ASTM subcommittee D 07.01.

You must be logged in to download any documents. Please login (login accounts are free) or learn how to Become a Member