Lumber Strength Estimation by Multiple Regression

Bending and tensile strength of 2 by 4 lumber was estimated by one parameter and multiple parameter regression analyses. The strength predictors evaluated include static and dynamic modulus of elasticity (MOE), shear modulus, density, velocity, screw withdrawal resistance, acousto-ultrasonic parameters, visual parameters (knot diameter ratios and modified knot diameter ratio). Remarkable improvement was found in strength estimation when the regression model was changed from single parameter regression to two parameter regression. The best bending strength predictor is the MOE followed by the modified knot diameter ratio, while the best tensile strength predictor is the 4-face concentrated knot diameter ratio and followed by the dynamic MOE obtained through longitudinal vibration. The result affirm the importance of simultaneous grading of lumber based on MOE (machine stress rating) and appropriate visual grading of lumber.