Edinburgh Napier University

  This paper presents an experimental investigation in which the torsion testing approach was employed to evaluate the shear strength of Sitka spruce and Norway spruce joists. Rectangular
structural size specimens of lengths ranging from 1.0 m to 3.6 m were tested using a 1 kN-m torsion testing machine. In order to determine the shear strength, specimens were tested until they either fractured, or exhibited pseudo-plastic behaviour under applied torque. The failure modes and the
correlation of shear strength and torsional shear modulus were also studied.
In current testing standards, the measurement of shear strength for wood is mainly based on testing small clear samples and on bending tests of full-size structural lumber. Tests on shear blocks do not account the influences of wood defects and therefore the test procedure underestimates the
heterogeneous nature of wood. In the bending test practice, the interaction of axial bending stresses and loading embedment with shear stress means it is impossible to obtain a state of pure shear.
By overcoming these problems using torsional loading, it was found that the characteristic shear strength of the wood in the tested joists was 166% to 200% higher than the published design values in EN 338 and the mean strength was 8% to 13% higher then the shear block based published values
in the USDA Wood Handbook.
The test joists fractured mostly at the middle with cracks propagating towards either the supports or to the top or bottom surfaces. However, combined tension shear and crushing failure modes were sometimes observed at supports. A correlation (R2 = 0.40) was found between the shear strength
and the shear modulus obtained from torsion tests. However, it appeared from this study that knots do not have substantial influence on the shear strength.