Tensile Strength of Redwood Dimension Lumber. II. Prediction of Strength Values

This paper presents results of statistical analyses relating tensile strength of redwood dimension lumber of both structural and nonstructural classification from Clear, Select, and Construction grades in 2 by 6 and 2 by 10 sizes to major defects. Various methods of predicting tensile strength, based either on visual assessment of defects or on mechanical stress rating, or a combination of both were examined. The possible effect of a number of minor variables on tensile strength was also explored. An analysis of defects and lumber characteristics was made to determine their relation to tensile strength parallel to grain. Out of this analysis a model based on visual characteristics was developed which predicted tensile strength with a correlation coefficient of 0.69. The procedure was based on determining the relative area of the specimen cross section which was occupied by the principal defect and its lateral eccentricity with respect to the center line of the piece. The equation encompasses several defects including knots, local cross grain, and birdseye. It was found that relative knot area was nonlinear in its effect on tensile strength, local cross grain acted like a knot scaled to quarter size, and birdseye acted like a knot when it was the only significant defect present. When small clear specimen properties were included in the regression equation, the correlation was improved. The highest correlation (0.76) resulted from the inclusion of dry tensile strength of small clear specimens into the equation. The correlation between tensile strength and strength ratio was 0.62. The use of nondestructive testing to predict working stress values for tension was examined using a static flatwise bending and a vibration stiffness measurement. The two measurements highly correlated to each other (r = 0.988). The correlation between static bending stiffness and tensile strength was 0.49 for all pieces, and varied between grades. The combination of stiffness measurement with visual characteristics improved the estimate. Bending stiffness and strength ratio together gave a correlation coefficient of 0.77, while bending stiffness and the defect parameters of the special model together gave a coefficient of 0.82. The effect of several minor variables was evaluated statistically. Crook, when present, of 0.4 inches or greater in 12 feet of length adversely affected tensile strength. Compression wood when present In up to 50 percent of the cross section did not affect tensile strength. Clustered birdseye had an adverse effect on tensile strength. The presence of checks did not affect tensile strength.

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