Basedon a preliminay testing, an optimized procedure for in-panel variability evaluation has been developed. A database was created consisting in fifty-five 0-2 OSB panels from nine member mills of the SBA conforming to the Canadian standards. The panels have been evaluated at the Eastern Laboratory of Forintek Canada Corp. and at the Alberta Research Council. Each panel was spit into 18 squares containing specimens for conducting tests on tensible strength perpendicular to surface (IB), thickness swelling (TS), water aborption (WA), moisture content (MC), density, thickness and static bending (MOR/MOE). Variance was used as the criteria for in -panel variability.
A survey with experts from the industry and research centres estimated that the variability on thickness, MOR/MOE parallel to forming, TS and IB is the most critical for the applications of OSB. According to the tests results, the most variable properties were MOE (variance from 405 to 676 GPa), MOR (variances from 28 to 38 MPa) parallel to forming and IB (variances from 5 to 11 kPa). The in-panel variability was on average ten times higher compared to among-panel variability, and the in product vatiability was three times as high as the among product variability. Density variation significantly affected only the variations in MC, WA, MOR and MOE. Density at break point and panel core density did not show better correlation correspondingly with MOR and MOE compared to average density. Further research is suggested to grade the factors of variability in OSB panels by their importance, to determine the criteria for acceptable limits of property variations and to define methods for on-line monitoring of horizontal density distribution.