June 16-21, 2019
The design of a fast pyrolysis reactor to convert biomass has a decisive influence on quality and yield of fast pyrolysis bio-oil (FPBO). Quality requirements are comparably low for the application of FPBO as gasifier fuel for subsequent conversion to synthesis gas, e.g. in the case of the bioliq® concept to convert (ash-rich) agricultural residues to drop-in, 2nd generation biofuels. Within this concept, one optimization parameter of fast pyrolysis is to maximize carbon yield in the liquid product while keeping product requirements that allow feeding into a high-pressurized entrained flow gasifier. This optimization space allows for a more flexible choice of reactor design. The aim of this study is to investigate the influence of reactor type on production cost of FPBO within above outlined framework, i.e. as feedstock for a downstream gasifier.
The investigation will be based on two different type of reactors. First, a twin-screw mixing reactor (TSMR) is being considered, which resembles the actual realization of the 500 kg h-1 fast pyrolysis pilot unit that is being operated as part of the bioliq® project. Second, a fluidized bed reactor (FBR) will be compared to that, which represents state of the art technology of industrial fast pyrolysis units. One important difference between the two reactors is the necessity of a fluidizing agent in case of the FBR, which in turn influences process design and equipment size, specifically in the product recovery section. This additional (inert gas) volume flow is not required in the case of a mechanical mixing, as is the case in the chosen TSMR. At the same time it is obvious that there will be a significant difference in mixing conditions of biomass and heat carrier particles in the two types of reactors, which will translate to a difference in heating rate of the biomass particles. This in turn might affect FPBO quality and yield.
Experiments have been conducted to compare FPBO yields from process development units that feature a TSMR and an FBR, respectively. No significant differences in FPBO yield have been observed. On the one hand this leads to the conclusion that the high heat transfer required to achieve one of the fast pyrolysis conditions (i.e. high temperature of primary pyrolysis inside the biomass particle) is comparable in both types of reactors. This could be explained by the high bulk density achieved during mechanical agitation as compared to that of a fluidized bed, which is capable of making up the lower mixing intensity if a proper ratio of biomass to heat carrier particles is kept. On the other hand, wheat straw (which is the ‘model’ feedstock for the bioliq® project) was used as feedstock in these experiments This choice might also lead to not observing differences between the two reactor types. Wheat straw is characterized by high ash content (around 8%) which increases the significance of secondary cracking reactions and thus lowers any effects of reactor type. Wheat straw also exhibits high heterogeneity which translates to increased standard deviation of the results (confirmed by multiple test runs) and an increased difficulty to detect differences in FPBO yield.
Based on the experimental results, the effect of process design on FPBO production cost are reduced to investment and operation cost. Existing production cost calculations for the bioliq® concept have been reviewed and updated due to the currentness of the underlying data. Additionally, relevant process design changes and equipment cost will be implemented for consideration of an FBR instead of the TSMR. Finally, a sensitivity analysis is conducted to reflect changes in product yield based on available literature data for fast pyrolysis of wheat straw in order to account for the previously discussed uncertainty of the obtained experimental results.
Nicolaus Dahmen, Axel Funke, and Jörg Sauer, "Influence of reactor type on production cost of fast pyrolysis bio-oil" in "Pyroliq 2019: Pyrolysis and Liquefaction of Biomass and Wastes", Franco Berruti, ICFAR, Western University, Canada Anthony Dufour, CNRS Nancy, France Wolter Prins, University of Ghent, Belgium Manuel Garcia-Pérez, Washington State University, USA Eds, ECI Symposium Series, (2019). https://dc.engconfintl.org/pyroliq_2019/17