(714d) Feedstock Preprocessing, Fractionation, and Blending to Improve Overall Cost, Supply, and Quality Considerations for Catalyzed and Uncatalyzed Fast Pyrolysis
The presented work examines the interdependencies of these variables and how they affect the biomass blends required by biomass depots and/or biorefineries to achieve the lowest cost feedstock with sufficient quality at the quantities needed for biorefinery operation. Four biomass depots were proposed in South Carolina to each produce 200,000 tons of feedstock per year. These depots supply a centrally located 800,000 ton biorefinery that converts the feedstocks to bio-oil using either catalyzed or uncatalyzed fast pyrolysis. The four depots utilize biomass based upon availability, but the feedstock or feedstock blend still met the minimum quality requirements for the biorefinery. Costs were minimized by using waste biomass resources such as construction and demolition waste, logging residues, and forest residuals. Preprocessing methods such as air classification and water washing or acid leaching were used to upgrade biomass quality. Air classification was found to be an effective method for separating low- and high-ash fractions of woody residues. For example, a high-ash fraction of logging residues was generated that contained approximately 50 wt% of the total original ash content, but less than 8 wt% of the original organic content. Dilute-acid leaching was then effectively used to remove between 85% (25°C, 0.5 wt% acid) and 98% (90°C, 1.0 wt% acid) of the alkali and alkaline earth metals from the high-ash fraction, making it suitable for recombination with the low-ash fraction. Various combinations of such preprocessing methods were tested to produce an array of blends that were available for recombination to produce least-cost blendstocks that met various conversion specifications. For both uncatalyzed and catalyzed fast pyrolysis, all four depots produced blends that met quality and quantity specifications at a cost lower than using a single feedstock.