Analysis of strength grading of sawn timber based on numerical simulation

The main problem in machine strength grading of wood lies in the grade determining parameters and machine grading parameters, which are all stochastic, and the correlation between them is of critical importance. Therefore it is difficult to reliably determine the grade of a particular piece. In order to understand the effectiveness and accuracy of machine grading a large number of experiments should be performed in which the non-destructive machine grading parameters as well as destructive grade determining parameters should be determined. So far we have had access to only a limited number of such experiments. The basic idea of this paper is to use experimental data only for the determination of statistical parameters - mean vector and variance - covariance matrix - and afterwards use these statistical parameters in numerical simulation of machine grading. This numerical procedure gave some valuable data on the average values and the variability of grading results which now make it easier to decide what sample size should be used in determining machine settings.