To improve process economics of the lignocellulose to ethanol process a reactor system for enzymatic liquefaction and saccharification at high-solids concentrations was developed. The technology is based on free fall mixing employing a horizontally placed drum with a horizontal rotating shaft mounted with paddlers for mixing. Enzymatic liquefaction and saccharification of pretreated wheat straw was tested with up to 40% (w/w) initial DM. In less than 10 h, the structure of the material was changed from intact straw particles (length 1-5 cm) into a paste/ liquid that could be pumped. Tests revealed no significant effect of mixing speed in the range 3.3-11.5 rpm on the glucose conversion after 24 h and ethanol yield after subsequent fermentation for 48 h. Low-power inputs for mixing are therefore possible. Liquefaction and saccharification for 96 h using an enzyme loading of 7 FPU/gÁDM and 40% DM resulted in a glucose concentration of 86 g/kg. Experiments conducted at 2%-40% (w/w) initial DM revealed that cellulose and hemicellulose conversion decreased almost linearly with increasing DM. Performing the experiments as simultaneous saccharification and fermentation also revealed a decrease in ethanol yield at increasing initial DM. Saccharomyces cerevisiae was capable of fermenting hydrolysates up to 40% DM. The highest ethanol concentration, 48 g/kg, was obtained using 35% (w/w) DM. Liquefaction of biomass with this reactor system unlocks the possibility of 10% (w/w) ethanol in the fermentation broth in future lignocellulose to ethanol plants.

Integrated Biomass Utilization System (IBUS) is a new process for converting lignocellulosic waste biomass to bioethanol. Inbicon A/S has developed the IBUS process in a large-scale process development unit. This plant features new continuous and energy-efficient technology developed for pretreatment and liquefaction of lignocellulosic biomass and has now been operated and optimized for four years with promising results. In the IBUS process, biomass is converted using steam and enzymes only. The process is energy efficient due to very high dry matter content in all process steps and by integration with a power plant. Cellulose is converted to bioethanol and lignin to a high-quality solid biofuel which supply the process energy as well as a surplus of heat and power. Hemicellulose is used as feed molasses but in the future it could also be used for additional ethanol production or other valuable products. Feasibility studies of the IBUS process show that the production price for lignocellulosic bioethanol is close to the world market price for fuel ethanol. There is still room for optimization -and lignocellulosic bioethanol is most likely a commercial alternative to fossil transport fuels before 2012.

Skin lesion classification based on in vitroRaman spectroscopy is approached using a non-linear neural network classifier. The classification framework is probabilistic and highly automated. The framework includes feature extraction for Raman spectra and a fully adaptive and robust feed-forward neural network classifier. Moreover, classification rules learned by the neural network may be extracted and evaluated for reproducibility, making it possible to explain the class assignment. The classification performance for the present data set, involving 222 cases and 5 lesion types, was 80.5%±5.3% correct classification of malignant melanoma, which is similar to that of trained dermatologists based on visual inspection. The skin cancer basal cell carcinoma has a classification rate of 95.8%±2.7%, which is excellent. The overall classification rate of skin lesions is 94.8%±3.0%. Spectral regions, which are important for network classification, are demonstrated to reproduce. Small distinctive bands in the spectrum, corresponding to specific lipids and proteins, are shown to hold the discriminating information which the network uses to diagnose skin lesions.

No abstract

In this chapter, we provide an overview of existing algorithms for blind source separation of convolutive audio mixtures. We provide a taxonomy, wherein many of the existing algorithms can be organized, and we present published results from those algorithms that have been applied to real-world audio separation tasks.