Ändamålsanpassad TINA-sortering av sågtimmer

Abstract: Swedish sawmills of today suffer from unsatisfying economy with high costs and low benefits. The economical result depends on how well the sawmill can use their raw material. Iggesund Timber AB, a sawmill in the Holmen group, produces a large amount is enduse orientated products for the woodworking industry. Therefore the aim of this student thesis was to investigate the possibility of automatic enduse orientated quality grading of Scots Pine (Pinus sylvestris) saw logs by using a gamma-ray log scanner. Iggesund Timber AB has the possibility to grade logs automatically with TINA, a two way gamma-ray log scanner. Grading and classification of Scots Pine logs in Sweden today are mainly manual but can in some rare cases be automatic. Classification is mostly based on mean quality of the two centreboards and mean value settings for a large group of logs. In aspect of the enduser this classification is not satisfying. The study consists of 178 pine logs in the timber class F290-299 mm which was randomly taken from the timber yard. This timber class is for example used for production of high quality components, mainly for the European market. The pine logs were carefully marked with colours and numbers in one end before sawing. The marking was very important for the re-identification. After the drying all boards were planed. All sorting of the planed boards were made by the personal of Iggesund Timber AB. The length and quality for each component were summed and provide the total value for each log. The value of each log was then used in the statistic analysis (based on multivariate calibration). The analysis resulted in two prediction models for automatic gamma-ray sorting of pine. The models result in two quality indexes that express the expected amount of components for the woodworking industry (mainly furniture production). The new quality indexes are based on two density variables and two external shape variables. The expected precision of the models is that they give about 70 % correct classification when sorting the logs into two quality groups. With developed log scanners and analysis methods it is possible to improve the precision of the sorting. In my opinion, it is obvious that the Swedish sawmill industry could improve their results by using defined methods for automatic timber sorting. But that would also mean a major change in thinking about market and quality, which I think may be the hardest part in this traditional industry.