Bioenergetische karakterisering van tropische houtsoorten

The overall aim of this dissertation was to examine the potential energetical valorization of sawmill residues from tropical areas, specifically Suriname.
The objectives were threefold: to provide fundamental knowledge on the studied timber species, to investigate whether state-of-the-art proxies like X-ray computed tomography (X-CT) and Near Infrared Hyperspectral Imaging (NIRhi) are good proxies to obtain this knowledge in a fast and non-destructive manner, and finally to use this information to make statements about the variability in traits between and inside species and its implications to a potential valorization. Variables like wood density, Higher Heating Value (HHV), ash content and extractive content were determined by classical methods and assessed. In order to link structural traits with some ecological aspects of the tree itself, radial trends in wood density were coupled to life-history strategy and growth of the tree. However, these classical methods often involve tedious procedures.
A significant relation was found between extractive content and HHV (R² = 0.53). No relations were found between wood density, ash content and HHV (R² respectively 0.00 and 0.39). Further investigation with a bigger dataset and more species could give more statements about these relations. Knowledge of the lignin and carbohydrate content (which was not investigated here) could influence these relations. X-CT was used as a proxy for wood density measurements. Statistical tests showed no significant differences between both methods, so X-CT is an appropriate method for densitometric measurements. Moreover, this technique is able to generate data at a high spatial resolution.
A proof of concept study was made on the use of NIRhi for the prediction of wood density and HHV. This system was built and optimized to obtain good quality hyperspectral data which were used to construct models based on Projection to Latent Structures (PLS). On the one hand it was shown that constructing models based on different species did not give adequate model quality. On the other hand, if a model was constructed based on only one species, its radial variation (in wood density and HHV) was well predicted. This means that the diversity of different components (extractives) has a great influence on the model. By using only species, this factor is held constant. This study illustrates that NIRhi is a suitable technique for predicting HHV when models are constructed for each species apart. Further investigation with lignin and cellulose concentrations could further explore the potential of NIRhi.
The results showed a significant variability in HHV which could, next to moisture content, have its practical implications for valorization. Wood density is also a very important parameter for the wood industry. When energy was expressed in volumetric terms, it indicated that the high variability in wood density of the investigated timber species, was more important than the subtle differences in HHV (which is expressed in gravimetrical terms).