Computer simulation in the forestry-wood chain

Sammanfattning: The forestry-wood chain today involves many actors, and the decisions taken in the process of making trees into finished products are so many that the effect of each individual decision is difficult to assess, especially if the natural variation of the input material is considered. This means that a simulation approach to the forestry-wood chain is suitable, since it makes it possible to evaluate the effects of decisions in a short timeframe, while the input material can be kept constant for different production setups. The long term aim is to connect tree and log properties to the quality of a final product through simulation, depending on the various operations involved in making the product. A part of this is realized through this thesis. It is shown that a simulation model of a cross-cutting and finger-jointing process is representative of a real process. The model is used to evaluate an adaptive strategy for setting the safety zone size between finger-joints and sound knots, a strategy which improves recovery in a finger-jointing operation by 3 %. Another issue addressed in this thesis is that of simulation input data. The sawing simulation tool used to a large extent in research today, Saw2003, uses the Swedish Pine Stem Bank as input data. This is a very well-documented data source, and computed tomography (CT) scanning of the stem bank logs allows wood features such as knots to be represented in a realistic way in Saw2003. There are limitations to the stem bank, however, mostly due to the fact that CT scanning is a time consuming process, which means that the amount of scanned logs is relatively small. In this thesis, it is shown that it is possible to use a small amount of log features for determining the knot structure in a log, which opens up possibilities for using industrial data from two-directional X-ray scanners. This would increase the amount of logs to be used as simulation input data. A second set of data used for simulation was also collected in a study of a production process, where thewood raw material was followed from the log yard through all production operations of making finger-jointed furniture components. Each individual piece of wood was traced through the operations, thus ensuring a link between the properties of logs and those of the finished product. This second data set was collected by grey-scale camera scanning of boards prior to cross-cutting and finger-jointing, and was used in the development of cross-cutting and finger-jointing simulation. It contains information on non-clearwood features of the board surfaces such as knots, cracks, and pitch pockets. It can be concluded from this thesis that it is possible to increase both functionality and the amount of input data in the simulation of the forestry-wood chain, and by doing so production strategies and decisions can be evaluated. Wood quality discussions may be simplified by being able to assess the effects on the production process of decisions being made. Future work involves adding more functionality to the simulation environment as well as evaluating the methods proposed in this thesis industrially. The long term vision is to be able to integrate the forestry-wood chain from log to finished product in one simulation model.