Advances in Biomass Integrated Size Reduction and Separation
A.R. Womac, X.P. Ye, D.G. Hayes, C. Igathinathane, S. Klasek, P. Miu, T. Yang, M.Yu, S. Sokhansanj, S. Narayan
Collaborative Project of: The University of Tennessee, First American Scientific Company, Oak Ridge National Laboratroy, University of British Columbia
Executive Summary
Project objective is to address problems in biomass size reduction (chopping, grinding) and dry
separation of plant components (sort by botanical/ chemical properties). First year priorities were to
identify biomass ultimate failure stress/ energy/ physical properties and to better understand current
equipment to identify grinder and separating actions for in-depth evaluation. Assessment tools were
developed and include rapid imaging for sizing large, non-uniform particles, FT-NIR for rapid
chemical analyses and wet chemical protocols to evaluate targets and separated biomass.
Biomass “models” were selected as follows: corn stover, switchgrass, rice straw, hickory wood, and
bagasse. Performance targets identified for grinding energy from published literature were 40, 20, and
10 kW-h/ton for fine (~5 mm), medium (~10 mm), and coarse (~20 mm) grinds, respectively, for
relatively dry (10% w.b.) fiber-rich biomass. A weakness of published literature was good
documentation of pre- and post- grind particle spectra for comparisons. Most published data dealt with
agronomic crop forage, not an array of biomass properties and conditions. Project original goal was
15% grinding savings, and is now projected to reduce typical grinding cost $3 to $4 per dry ton (about
1/4 of current costs) based current understanding of pre- and post-grinding literature data. Multiple
stage grinding may be most appropriate to maximize efficiency. Target particle size for bio-refinery
was identified as ~6 mm based on input from industry experts. One U.S. Patent (5,677,154,
Production of ethanol from biomass) verified nominal sizes of ~1 to 6 mm. These small sizes will
likely require multiple stage grinding, and one question deals with identifying equipment and
operations for each stage. Most literature report separation effectiveness between grain and
lignocellulosic materials with separation efficiencies 95% or greater. Lower performance targets
between for stalks and foliage material streams are expected because material properties are more
similar than grain versus chaff.
Size reduction technologies were identified for 2nd year instrumented testing as follows: hammer mill,
knife mill, disk mill, and variable-spacing linear knife grid. Rationale for selection included
maximizing more efficient shear failure with knife, shear bar, and pinch points. A Warner-Bratzler
shearing device evaluated different knife bevel angles (30° and 45°) and found 18, 31, and 22 % less
input energy for the 30° bevel angle for corn stover, hickory, and switchgrass, respectively. Direct
measurement of grinder input power is being emphasized to evaluate grinders (not inferred as often
published).
Sieve technologies were prioritized for separation. First, accurate evaluation of particle spectra from
grinding studies is necessary to compare grinder energy results. Second, sieves are commercial-viable
for biomass separation. Terminal velocity was identified as a separation action whereby aerodynamic
and gravitational forces are equilibrated at a given air stream velocity. A 3.5-m tall vertical wind
tunnel was developed for the project. Example terminal velocities of 12-mm long pieces of dry
switchgrass were 5.6 m/s for internode, and 7.6 m/s for node – thereby indicating separation potential
since the terminal velocities varied by more than 10% (published recommendation).
Micrographs of biomass cross sections were made to aid constituent identification (eg, silica). Image
analysis techniques using a flat bed scanner were developed to measure irregular biomass on sieves.
An FT-IR/NIR instrument operated in diffuse reflectance mode was applied to biomass compositional
analysis. Standard wet chemistry techniques were initiated for monosaccharide units of cellulose and
hemicellulose (HPLC), acid-insoluble lignin (oxidation, furnace), acid-soluble lignin (UV 205 nm),
and ash (oxidation, furnace). New areas of improved chemistry include the use of ionic liquids to
solubilize biomass for rapid wet chemistry and supercritical fluid chromatography (SFC) detection
(SFC-UV 190 nm) with no solvent interference, linear calibration, and strong signal-to-noise ratio.
SFC may be a desirable alternative, providing comparable resolution of chromatographic peaks at a
shorter run time with less waste and costs.
Impact of Activities
Individual Stem Properties were measured to better understand basic size reduction processes.
Modified Warner-Bratzler device shown. Results indicate that the higher the biomass strength –
the greater the response sensitivity magnitude and potential energy savings. Moisture and timing
of grinding may be an important consideration to reduce energy for some biomass.
Instrumented Size Reduction of biomass using direct measures of torque and rpm of rotary grinders
and input force-displacement of linear knife grid is underway to provide better, more objective
comparisons of the grinding process using different types of equipment.
Technical Approach & Targets
•Use shear failure for grinding, and particle property differencefor separation
•Separation good for 10% or greater “threshold of difference” in property
•Adjust potential to improve size reduction/ separation performance based on instrumented results
•Target of 40, 20, and 10 kW-h/ton for fine (~5 mm), medium (~10 mm), and coarse (~20 mm) grinds, respectively, for relatively dry (10% w.b.) fiber-rich (tough) biomass
•Target grinding savings from ~$1/ton (stover) to $5/ton (switchgrass) (includes additional expense for providing separation by botanical component)5.717.723.424.932.877.06Terminal Velocity(m/s)InternodeNodeInternodeNodePith(internode)Rind(internode)BiomassPropertyDry SwitchgrassDry Wheat
|
|
