Hållfasthetssortering av plankor med hjälp av röntgendata
The aim of this thesis is to find a statistical model that sort logs according to a predicted modulus of elasticity on the sawn boards. The input parameters for the model are X-ray variables from an X-ray scanner. For some of the log classes sawn at the sawmill it is desirable to have a narrow distribution of modulus of elasticity on the sawn boards. 250 pine logs from five different diameter classes were numbered and followed through the sawmill process. Log data were collected from X-ray and 3-D scanners in the log intake. After the logs were sawn in to boards, the boards were numbered. After the boards were dried and conditioned the eigenfrequency were tested to decide the modulus of elasticity. Log and board data were processed in a multivariate statistic program, and sorting models were extracted. The models extracted show satisfying r2 and q2 values, and a Pearson chi squared test showed that the models are predicting modulus of elasticity better than random sorting. The models are now being used in the production at Munksund sawmill.