Grass Biomass Energy Field Day Proceedings

September 13, 2012

Vermont Technical College

Randolph Center, VT

Sponsors

Grass Biomass

Energy Field Day

Proceedings

Agenda

9:30 - 10:00 Registration and Networking (Red School House)

10:00 - 10:45 Presentations Red School House Conference Room

10:45 – 11:20 Tour of Frohling pellet boiler/Poster sessions

11:20 – 12:15 Tour of UVM Extension Warm Season Grass Trials

12:15 – 1:15 Lunch (VTC Cafeteria or on your own)

1:15 – 2:30 Grass Densification Demonstrations at the VTC Farm

2:30 – 3:15 Optional follow up and networking opportunities

- RSH tour of Frohling boiler

- Discussion of business opportunities at RSH

- Discussion of VTC Projects and Curricula at RSH

- Revisit grass plots

3:30 Adjourn

Table of Contents

Page

Map of Vermont Tech with directions to Plots and Cafeteria 1

Plot Plan of Warm Season Grass Trial 2

Summary of Field Research Projects on Grass Energy in Vermont 3

Pelletizer Interim Assessment Report from Vermont Tech SOAR Program 11

Special Thanks 22

For More information about the Grass Energy Partnership contact:

Sid Bosworth, (802) 656-0478, [email protected]

Netaka White, (802) 828-1260, [email protected]

Joan Richmond Hall,(802) 728-1717 [email protected]

A special thank you to U.S. Senators Patrick Leahy and Bernie Sanders and the U.S. Department of Energy for their

generous support for ongoing renewable energy research and education.

UVM Extension helps individuals and communities put research-based knowledge to work. University of Vermont Extension, and U.S. Department of Agriculture, cooperating, offers education and employment to everyone without regard to race, color, national origin, gender, religion, age, disability, political beliefs, sexual orientation, and marital or familial status. If you require accommodations to participate in this program, please let Sid Bosworth, (802) 656-0478, know by Sept. 4, 2012 so we may assist you.

mailto:[email protected]



1 of 22

2 of 22

Using Perennial Grasses for Biomass Production in Vermont

A Summary of Research Projects on Grass Energy in Vermont

Sid Bosworth, Extension Agronomist, University of Vermont

In developing the Vermont Grass Energy Partnership, several questions were raised

concerning the potential of using perennial grasses for biomass production:

What grasses and grass cultivars would work best for grass biomass in Vermont?

What are the challenges for establishing these grasses?

What are the production and sustainability concerns for managing these grasses?

What are the harvest considerations for making biomass from grass?

What are the economic considerations for growing, harvesting and processing grasses

for biomass energy?

What is the environmental impact (air quality) on using perennial grasses as a biomass

source for heat energy?

Over the past four years, the University of Vermont and UVM Extension has conducted

field trials to address some of these questions. Starting in 2009, several field trials were

established at various locations in Vermont. The following is a table showing the type of

trial, the locations for these trials and the years conducted.

Table 1 Locations

Title of Each Trial Years South

Burlington Shelburne

Randolph Center

East Randolph

Alburg

Warm Season Grass Specie and Cultivar Evaluation

2009 to

2012 x x x #

Cool Season Grass Specie and Cultivar Evaluation

2009 to

2012 x

Switchgrass Seeding Rate Study

2009 to

2010 x x

Switchgrass Emergence Study

2010 to

2012 x x

Switchgrass Nitrogen Response Study

2010 to

2012 x x

Reed Canarygrass Nitrogen Response Study

2009 to

2010 x x

#failed due to major weed problems

3 of 22

In order to summarize results from these trials as well as observations that we have made

over the years at these sites and other demonstration sites, I will again address each

question one at a time.

What grasses and grass cultivars work best for grass biomass in Vermont?

Just about any type of plant material can be dried, densified and burned for energy.

However, for a particular crop to be sustainable and economical for this purpose, it would

need to be productive and persist over a long period. It would also need to meet minimum

fuel quality characteristics. The focus of this research is on perennial grasses. They can

persist for many years, they generally form a sod that helps to protect the soil from erosion

and runoff, they can be very productive, and can be harvested using standard hay

equipment that is readily available in Vermont.

Up to six warm season grass species were evaluated at one to three locations including

switchgrass, big bluestem, indinagrass, giant miscanthus, cordgrass and coastal

panicgrass. There has also been four cool season grass species at one location including

reed canarygrass, intermediate tall wheatgrass, tall wheatgrass and wild rye. For some of

these grasses, several cultivars were also included in the trials.

Below are some findings from our studies so far:

In the seeding year in 2009 at UVM and Meach Cove Farm, only the switchgrass plots had enough growth at the end of the season to warrant a harvest (Figure 1)

Yield at UVM and Meach Cove was quiet low in the seeding year (about 1 to 1.5 tons per acre), probably not enough to offset the cost to harvest (Figure 1). However, seeding year yields were higher at VTC due to excellent weed control and timely rainfall patterns. Even big bluestem and coastal panicgrass had yields over on ton per acre (Data not shown).

Yields improved each year with the highest in the third and latest year of collecting data (Figure 1). There will be at least one more year to collect yield data.

Of the switchgrass cultivars, ‘Cave-N-Rock’ and ‘Shawnee’ have been the most productive in biomass yield across all locations (Table 2 and Table 3). ‘Kanlow’, a lowland cultivar, has done well at UVM but poorly at the other two sites. ‘Blackwell’ has done well at VTC, its only location. ‘Sunburst’ yields less and consistently shows more leaf diseases. ‘Bowmaster’ was killed out the first winter.

Of the other grass species, ‘Prairieview’ big bluestem has yielded well at the two locations where it is located. ‘Niagara’ big bluestem has had moderate yields. The other grass species have not yielded as well.

Ash content varied by location reflecting both soil fertility levels and water availability (Tables 2 and 3).

Generally, the big bluestem cultivars were consistently lower in ash content than the switchgrass cultivars (Tables 2 and 3).

4 of 22

Figure 1. Biomass yield of switchgrass cultivars in the first three years of the stand. The first year is the seeding year (2009 at UVM and Meach Cove and 2010 at VTC). All harvests were made in October of each year.

5 of 22

Table 2. Average yields and ash content of three year old stands of warm season grasses

plus reed canarygrass harvested in 2011.

Table 3. Average yields and ash content of two year old stands of warm season grasses

in Randolph Center (VTC Farm) harvested in 2011.

6 of 22

Results so far (continued):

Cordgrass also had very low ash content but yields are fairly low (Table 3).

Lodging can be a problem. It seems to be worse with certain species particularly big bluestem and cordgrass (data not shown).

Stand persistence is an issue considering long term sustainability. The most obvious treatment was ‘Bowmaster’ switchgrass with died out the first winter. Another switchgrass cultivar that has been inconsistent across locations is ‘Kanlow’ and the indiangrass cultivar has been inconsistent.

Giant miscanthus, which had to be planted by rhizome cuttings since the seed is sterile, had winter kill of some of the plants especially at VTC and Shelburne. Its yield has varied greatly across locations (Table 2 and 3) and needs several more years of evaluation before being considered as a viable option in Vermont.

One consideration would be to formulate polyculture mixtures of various species. Most compatible would be switchgrass with big bluestem and indiangrass. This would help spread the risk of failure due to some climatic or soil condition The polyculture mix in these trials were a mixture of ‘Cave-N-Rock’ switchgrass and either ‘Prairieview’ or ‘Niagara’ big bluestem. This combination works well with the higher yields from the switchgrass and the lower ash content of the big bluestem (Tables 2 and 3).

Cool season grasses yields have not been as good as some of the warm season grasses and their ash content is fairly high. There is also a lodging problem with many of the wheatgrasses and, other than reed canarygrass, their stands appeared to have declined by 2012 (Table 4).

Table 4. Biomass yield, ash content and lodging of cool season grasses harvested in

2010 and 1022 in South Burlington. Planted on 8/18/2009. Adams sandy loam

soil.

7 of 22

What are the challenges for establishing biomass grasses?

Many of the warm season grass species are very slow to establish. Their seed has innate

dormancy mechanisms which can delay germination by many months and once

germinated the seedlings are relatively slow to grow. Weed control is very critical since

these grass seedlings are not very competitive. We conducted two different studies at two

locations to assess the optimum seeding rate. The following is a summary of the findings

from both the trials and field observations in establishing these grasses.

A seeding rate of 8 to 10 pounds per acre of switchgrass (accounting for both % germination and % dormant seed found on the seed tag) seems adequate to achieve a productive stand (Table 5).

Purchase seed that has a high germination and low dormancy if possible. Seed that is treated with fungicides can also help prevent seedling diseases.

Seed should not be planted until the soil is warm enough to promote quick germination. This is usually within the first two weeks of June in Vermont.

It is best to plant in a field with low weed pressures, particularly summer annual grasses like crabgrass or foxtails.

To manage weeds, a stale seedbed approach can work affectively if weather allows. The field should be prepared one month to six weeks before planting and allowed to have a couple of weed flushes before planting. Emerging weed seedlings can be killed with blind cultivation using a flex tine weeder Set the tines for shallow tillage to avoid bringing up more weed seed form lower soil depths. The flush of weeds can also be killed with a herbicide like glyphosate before the warm season grass is planted. In this situation, the weeds should be allowed to emerge and grow some before spraying in order to get adequate foliage coverage.

For fields with a history of heavy weed pressures, it may be best to rotate the field for a year or more with an annual forage crop such as sorghum-sudangrass that affectively suppresses weed emergence and growth. This can help reduce the weed seed bank before establishing the warm season grass.

Table 5. Biomass, tiller population and tiller weight of four switchgrass cultivars harvested

in autumn of 2010 from plots seeded at four different seeding rates in 2009.

8 of 22

What are the production and sustainability concerns for managing these grasses?

Once established, the major factors that will affect the life of the stand and sustained

productivity include fertility management, weed control, diseases, and cutting

management.

Much research has shown that cool season grasses respond to nitrogen (N) fertilizer when

managed for hay; however, less is known when managed for biomass. Nitrogen

response trials were conducted on established stands of reed canarygrass (‘Palaton’) and

switchgrass (‘Cave-N-Rock’) over a two to three year period to assess biomass

productivity. The following is a summary of the findings for both grass trials and our

observations in maintaining established stands.

Summary of Reed Canarygrass N Response Trials:

In a two year study, reed canarygrass did have a positive yield response to additional N fertilizer with the response being similar in both years. (Table 6)

Linear regression analysis indicates that the optimum N application rate would be about 75 lbs. of N per acre (applied just as the grass is starting to grow in mid spring).

There was visually significantly more lodging in the high nitrogen plots; however, the degree of lodging was mild enough that a normal mower or mower/conditioner should have been able to efficiently handle it if this had been a production field.

Ash content as well as P and K content were lower with increased N fertilization rates which is desirable from a fuel quality perspective.

Removal rates of P were about the same regardless of N fertilization; however, K removal was highest with the 100 lb./acre N rate. At 70 lbs of K removed per year, a producer would most likely need to supplement with either fertilizer or manure to sustain the field over a long period of time.

Summary of Switchgrass N Response Trials

This study was initiated on a four year old stand of ‘Cave-In-Rock’ switchgrass located at the Borderview Farm in Alburgh, VT.

In the first year, there was no difference in yield amongst the N treatments (Table 7). Switchgrass is very efficient at utilizing N from the soil. Studies in New York have found similar results in which switchgrass showed no response to additional N in the first four years of their study. However, by the next year, there was a significant response to N when treatments were applied to the same plots.

There was visually significantly more lodging in the high nitrogen plots in the first year but not to the point of being a mowing problem (Table7).

Ash content was not affected by N treatment (Table7).

Neither phosphorus nor potassium levels were affected by nitrogen fertilizer applications at the rates applied (Table 8). With their low P content, there was relatively little of this mineral removed. However, the average removal of potassium was a little over 40 lbs. of K per acre. So, over a period of a few years, a producer would need to replace this K with either fertilizer or manure.

9 of 22

Table 6. Biomass yields, lodging, ash content and mineral removal of a four year old stand of

‘Palaton’ reed canarygrass harvested August 27, 2010.

Table 7. Biomass yields, lodging, and ash content of a four and five year old stand of ‘Cave-N-

Rock’ switchgrass in Alburg, VT. Treatments were repeated on same plots.

Table 8. Biomass mineral content and removal of a four year old stand of ‘Cave-N-Rock’

switchgrass harvested October 28, 2010 in Alburg, VT

10 of 22

Pelletizer Interim

Assessment

A Report by the

2012 SOAR Interns:

Adam Sprague Ben Cayer

Blaine Conner Cole Predom

James Blakely

June 2012

11 of 22

Pelletizer Interim Assessment 24 July 2012

Summer of Applied Research July 2012 Page 4 of 22

GOALS

Vermont Tech sought to identify the process parameters of

making biomass fuel pellets with a small-scale pellet mill. The

2012 SOAR team developed operational procedures needed to

produce pellets with the Buskirk pellet mill and conducted

pellet production experiments modifying the speed of the

pellet die, feedstock type, and moisture content, mass flow

rate of input material, and additives.

RESEARCH

We began research by conducting an internet search on biomass pellets. We found various sources

suggesting that moisture content and lignin content of the feedstock have the most impact on

biomass pellet quality (Pelheat.com and Biomass Energy Resource Center). Lignin is a naturally

occurring chemical compound that acts as a binding agent when heated in the pelletizing process. It

is general knowledge that attempts to make 100 percent grass pellets were typically unsuccessful

because of low lignin content. The moisture content of the feedstock governs the ability of the pellet

mill to melt the lignin and glassify the exterior of the pellet. If the moisture content is too high, the

pellet mill is unable to evaporate the moisture and pellets are not produced. Too low, the heat

transfer rate from the pellet mill die to the feedstock drops and the feedstock exits the mill as dust

rather than cohesive pellets.

Next we began research on the pellet mill itself. Using U.S. Department of Energy funds obtained

with the help of Vermont Senator Bernie Sanders, Vermont Tech purchased a “turn-key” system

manufactured by Buskirk Engineering of Ossian, IN. This system includes a hammer mill, blower,

surge tank, pellet mill, water pump and cooling conveyor, all mounted on a 5’ x 10’ steel base. We

mounted this self-contained system on a 7’x 16’ trailer. Though the system was advertised as “turn-

key” it was clear that this machine was complex and that a thorough investigation of the operating

parameters was necessary before any laboratory experiments could be conducted.

First, we analyzed each component of the system and became familiar its individual operating

parameters. Documentation included with the system was sparse and required us to look up specific

information from the manufacturer of each specific part or device. Once we understood how each

component worked, we were able to look at the system as a whole and start developing operating

procedures and key parameters.

Figure 1: Ben and Blaine running the Pelletizer. The 55HP Tractor was generously supplied by Roger Howes.

12 of 22



Figure 3: Operation of hammer

mill

COMMERCIAL PELLET MILL TOUR

We were fortunate enough to take a tour of Vermont Wood Pellet Company in North Clarendon,

Vermont on May 23. Although our pellet mill is much smaller than theirs, our Buskirk system

replicated every step and process of their larger mill. Brian Pinsonault the maintenance manager

reiterated the fact that lignin and moisture content of feedstock are major factors in the pellet

making process. He also said that the type and depth of your die can be a factor in production.

Vermont Wood Pellet uses a vertical-ring die where the rollers roll on the inside of the ring. Our

system has a horizontal die in which the rollers rotate on top and force the material down through

the die. The tour gave us

firsthand knowledge about the

process of making a wood

pellet, and we were able to

apply this knowledge directly to

our own system.

SAFETY

As delivered, the pellet mill did

not have all of the safety

equipment in place that we felt

is necessary. The biggest risk we identified with this pellet mill was the PTO (power take off) driven

portion of the machine. We determined that a safety platform would need to be built to keep

operator away from the PTO shaft. The safety platform also includes handrails and is only

accessible from the left side of the trailer. The safety platform

allows the person operating the electrical panel to oversee the entire

machine during production and react quickly to any issues that may

arise. The tractor cannot be operated from the safety platform;

however an emergency cord was run from the fuel shutoff on the

tractor to the platform so if a problem arises the tractor and PTO

can be shut down quickly from the safety platform. Two or three

people can observe the pellet mill during production from the

platform. Figure 1 shows the safety platform as-built drawings.

During production, eye and hearing protection are a must. We used

clear safety glasses to protect our eyes. The hammer mill poses the

greatest risk for eye injuries because when feedstock is fed into the

Figure 1: Top view of Safety Platform with tractor on the right and trailer and pelletizer w/ safety deck on left

13 of 22



mill, the machine tends to kick some particles back out through the hopper. The tractor, pellet mill,

hammer mill and blower are the loudest components of the system and require hearing protection

when they are running. During production runs, we used over-the-ear hearing protection with an

NRR of 29. General-purpose nuisance dust masks are strongly recommended. We used leather

safety gloves to protect our hands when feeding material into the hammer mill as seen in figure 2.

INITIAL COMPONENT TESTING

The primary focus of the first rounds of testing was to familiarize ourselves with the operation of the

“turn-key” pellet mill. Though the manufacturer provided some specifications about the machine

and its capacities, it was necessary to confirm these specifications and determine if correction

factors needed to be included in any of our pre-production calculations.

UNIVERSAL FRICTION CLUTCH

The manufacture did not provide method of separating the

pellet mill from the tractor. For safety purposes we installed a

universal friction clutch (slip clutch) on the pellet mill side of

the PTO. The slip clutch (Figure 4) protects the tractor and the

machine in the event of the mill rollers locking. If the pellet mill

locks, the torque spikes and the tractor slips the friction clutch,

and the pellet mill isolates from the tractor. If the slip clutch

was not in place, the tractor could shear the pellet mill shaft or

break the PTO transmission. We adjusted the clutch springs to allow the clutch to spin during a

lock up situation but hold during normal operation.

HAMMER MILL

The Hammer Mill utilizes a single phase VFD

controlled 10 HP motor that spins a series of beater

arms or knives at 3600 RPM. These beater arms are

encircled by an 18” circular screen with perforations of

a material specific size. Material is inputted through a

Plexiglas hopper that incorporates a plate magnet to

catch any metals that may be in the mix. The material

is hammered with the beater arms against the circular

screen which reduces the input material to the size of

the perforations. The reduced material is then Figure 5: Inside hammer mill

Figure 4: Universal Slip Clutch

14 of 22



Table 1: Mass Flow Rate Results for the Farm Sawdust

transferred from the Hammer Mill to the Surge Tank via a pneumatic air system that draws the

processed material from the bottom of the metal cabinet towards the front of the pelletizer unit.

The air system draws air from both the hopper intake and a 4” drilled plate, and other parts of the

pelletizer unit.

FEED AUGER RATE

The feed auger moves the feedstock from the surge

tank through screw action into the pellet mill. It is

electronically controlled by a variable frequency drive

(VFD) connected to a Human Interface Module

(HIM) located on the main power enclosure. The feed

auger frequency, in hertz, is displayed on the HIM.

The HIM allowed for on-the-fly adjustment of

frequency, which in turn changed the speed of the

auger motor. However, the module does not display

the mass flow rate of the material being moved.

During the initial testing of this component we

realized that we needed to know delivery rate (mass

flow rate) of the auger at different frequencies. It was

important to determine this mass flow rate because

we need to know how many gallons of water per unit

time needed to be added to the feedstock mix

to bring it up to the optimum moisture

content for pellet production. To determine

the mass flow rate of the auger at different

motor speeds we collected several samples at

various feed auger motor speeds and recorded

the amount of time needed to deliver a specific

volume. Figure 4 shows the feed auger and one

gallon bucket used. These measurements

allowed us to calculate a volume per unit time,

or mass flow rate. It should be noted that the

Sample

Moisture

Content

(%) Hertz

Mass Flow

Rate

(lbs/hr)

Farm Sawdust 45 15 97.3





Kiln-Dried

Shavings 9.7 30 85.0

Kiln-Dried

Shavings 9.7 45 130.7

Kiln-Dried

Shavings 9.7 60 151.0

Figure 6: Measuring Flow Rate

Figure 7: Surge mixers in surge tank

15 of 22



Figure 9: Cooling Conveyor

mass flow rate is dependent on the density of the feedstock and its moisture content. These rates

are only applicable for the materials tested and must be determined for any feedstock used to make

pellets.

WATER PUMP FLOW RATE

The second parameter tested was the water pump flow

rate metered by a King 7530 2C-04 inline flow gauge

measures in gallons per hour (GPH). As described on the

manufacture’s website, the meter has an accuracy of ±6%

of full scale flow, or ±.72 GPH. We repeated the method

used to determine the auger mass flow rate and found

that the gauge was inaccurate and required a correction

factor. Table II shows the percent error of the meter at

different gauge readings with an average error of 25%.

COOLING CONVEYOR

The Buskirk pellet mill has a cooling conveyor that is used

to cool down the hot pellets coming out of the die and to

suck out the fines that made it through the pellet mill

(Figure 9). The blower that is hooked up to the conveyor

sucks out the fines to be reused in the surge tank. The fines

become very dry after exiting the die and are blown back

into the surge tank;

this caused some

problems with our

moisture content. We

also noticed that the

pellets were coming

out warm even after

exiting the conveyor.

We decided to slow down

the cooling conveyor speed from 60 Hz to 30 Hz to give the

pellets more time to cool and more time for the fines to be removed.

Gauge

Reading

(gal/hr)

Actual

Output

(gal/hr)

Percent

Error

12 9.5 21%

10 8.8 12%

8 7.5 7%

6 5.1 15%

4 3.1 23%

3 1.9 36%

2 1.2 38%

1 0.5 50%

Figure 8: Pellets entering Cooling Conveyor

Table 2: Water Meter Actual Output

16 of 22



Figure 11: Black Bolt where Thermocouple is Mounted

Figure 10: Thermocouple Lead

DIE THERMOCOUPLE

Buskirk Engineering specified a

normal operating temperature for the

pellet mill die of between 160-200

degrees F. Buskirk did not provide a

method of measuring this

temperature so we installed a

thermocouple to one of the mounting

bolts on the die. This thermocouple is

attached to a multi meter that displays the temperature of the die. We

mounted the multi-meter next to all of our other controls and

included the die temperature in the data collected during a

production run (Figures 10 & 11). We found that a die temperature of

±170°F contributed to a durable pellet.

Moisture Analysis & Feedstock

Before we started to make

pellets we needed to identify a

good test feedstock that would

allow us to determine optimal

moisture content for pellet

production. Our research told us

that lignin holds the pellets

together and is a chemical

compound of wood and found in

plants and algae. We had a few

materials like feedstock from the

farm and commercial bedding

sawdust which were easy to get.

We used an oven and electronic

balance to determine the actual

moisture content, and then we tried to duplicate those results using a microwave. We concluded

that the tests were very similar and that we could use the microwave oven, saving time while

completing our moisture testing.

The first material that we tested was sawdust from the Vermont Tech farm which is used for

bedding. The sawdust appeared very moist in the middle of the pile and drier on the outside so we

Samples before pelletizing Moisture Content

Dry sawdust from the farm 19.47%

Wet sawdust from the farm 46.70%

Pre-hammer mill Control sample 48.10%

Post hammer mill sample 41.70%

Central Supply Commercial bedding

sawdust

9.70%

Round bale of switch grass 13.50%

50/50 mix switch grass and sawdust 25.80%

Table 3: Feedstock Moisture Contents

17 of 22



Figure 12: Inside the Pellet Mill

decided to take samples from each part of the pile for testing. We found that the average dry

sawdust moisture content was about 19.47% and the wet sample had about 46.7% moisture. Both

values seemed relatively high given that we a wanted a moisture content of 10%-15% going into the

die to produce good pellets. We then decided to test moisture content after the sawdust was

processed through the hammer mill and blown into the surge tank. We found the sample to be

about 7% drier after going through the Hammer Mill. This meant that we would have a lower

moisture content going into the die, but that moisture content would still be high enough so that

no water would need to be added.

The next material we tested was commercial bedding sawdust from Central Supply; a much drier

material. We tested the moisture content of the bedding sawdust and found it to be about 9.7%.

Because this material was drier, we needed to add water to bring it up to a moisture content of

about 15% for production.

Our third biomass sample was switchgrass. Switchgrass is native to the United States and has a

lower ash content than other grasses which makes it a promising biomass crop. “Switchgrass, as a

C4 species, has greater photosynthetic, water and N use efficiencies compared to C3 species.” (Jerry

H. Cherney) Switch grass has high productivity, high moisture content and nutrient use efficiency

that adapts to marginal soils. The switchgrass we used came from one of Sid Bosworth’s test fields

in Shelburne Vermont. We used a bale chopper to chop the grass into smaller pieces for the

hammer mill. We found that the moisture content of the switchgrass was about 13.50%.

PELLETIZING TRIALS & RESULTS

In order to learn about this system and see how it

actually worked we ran it several times and focused

more on our safety and getting used to running it than

on pellet production. We couldn’t see the sense in

doing experiments with this system before we could

run it in a somewhat flawless way. In these first trial

runs we had some success, and after running the

machine several times we had a system down, but the

production rate and pellet quality was poor. We

consulted with Buskirk and they suggested that we

adjust the distance between the rollers and the pellet

die. Buskirk said that the rollers in the pellet mill could be

spaced too far away from the die, therefore not making enough contact to push the material

through the die. We tightened the large bolt on top of the roller shaft that you can see in Figure 11.

We immediately saw improvements in our next experiment, when we were able to make some

pellets out of Pine shavings from our local supply store. We noticed that the pellet production rate

18 of 22



was much faster. This small adjustment made the pellet mill work the way that it is supposed to. For

two days we experimented with the commercial feedstock from the supply store and the sawdust

that the college farm uses for bedding. We were able to make consistently good pellets by dialing in

the right amount of water, feeding the stock into the mill at the right rate and monitoring the speed

of the PTO.

Moisture content and relative humidity seem

to be the factors with the most influence on

our pellet production rate. Our data

demonstrates that humidity does not affect

the pellet quality, but they significantly

lower the rate of pellet production. In one of

our experiments, relative humidity was low

and we had great quality pellets out of

feedstock that had a moisture content of

46%. We saw that the pellet mill would go

through a cycle when we were making these

pellets. This was a lag period which would

spin for about 10 seconds without pushing

pellets out, and then send them out for

about 4-5 seconds. We came to the

conclusion that as the feedstock material sat

and the rollers continued to spin, the

feedstock was losing a large amount of the

moisture, until it was at an optimum moisture

content that would allow for pellet production. In fact, the process was evaporating roughly 40% of

the moisture, because the finished pellets from those experiments consisted of 6-7% moisture. The

optimal temperature has been anywhere between 155-175 °F.

Once we achieved some success with wood feedstock, we moved on to switchgrass. We used a bale

chopper to roughly chop up about half of a round bale of switchgrass that was grown on the college

farm last year. We were finally ready to start making pellets with grass. We didn’t want to start with

100% grass because we had read about how challenging grass pellets could be, so we started with a

50/50 (w/w) mixture of wood sawdust from the farm, and switchgrass, Panicum virgatum. We mixed

the two together and ran the mixture through the hammer mill and into the surge tank. We knew

the moisture content of the wood and the switchgrass, but wanted to confirm that addition of water

was not needed. The moisture content of the wood/grass mixture was 25.8%, so we proceeded to

make pellets without additional water. We used old pellets to warm up the die to about 165°F. We

fed the wood/ grass mix in at 40 Hz and immediately pellets started coming out at a very fast rate.

The temperature climbed into the 170-180°F range, and we continued the pellet making process for

Graph 1: Moisture Content of Feedstock and Finished Pellets

0.00

5.00

10.00

15.00

20.00

25.00

30.00

35.00

40.00

45.00

50.00

Moisture Content

MoistureCont.Feedstock (%)

MoistureCont. FinishedPellet (%)

19 of 22



Figure 13: Pellet Durability Tester

about an hour until we ran out of material. This mix of grass and wood made a very durable pellet at

an increased production rate. There was no lag time as there had been with the wood feedstock.

Our next step was to try a mix that was 75% grass / 25% wood. We created the mixture as described

above, with the moisture content of the mixture being 20.5%. We fed the 75 / 25 mixture into the

pellet mill at 40Hz, when the die was about 160°F. Pellets began forming soon after we started the

feed auger. The pellets were not the quality of the ones that we made in the previous experiment,

but they were not the worst pellets we had made either. We saw some signs of insufficient lignin in

this mixture, as some of the pellets began to fall apart as they cooled.

The final experiment of our three-week pelletizing project was to make 100% grass pellets. We knew

that we would have to add a small amount of water because the moisture content of the switchgrass

was 13.8%. We started the process when the die was about 160°F, and did not add water. The initial

result without addition of water was that the pellet mill was pushing out fines, but no pellets. We

started to add water, and nothing seemed to happen. We then added vegetable oil, which acts as

binder. We added small amounts of the oil and began to make pellets. We kept adding about 2

gallons per hour of water, and stopped adding the oil, and the mill continued to make pellets for the

remainder of the batch. We think that the oil helped bind the grass long enough to plug the holes in

the die and slow down the material as it passed through the die.

We did a standard PFI (Pellet Fuel Institute) durability test on all of the pellets we had produced.

All pellets passed the durability test, in which means they are placed in the durability testing

apparatus for 10 min at 50 RPM (See Figure 13). We determined the mass of pellets after removing

fines in a 1/8” sieve before and after

durability testing. If the Pellet

Durability Index number is above

96.5 then the pellets are considered

premium grade, and all of our pellets

were graded as premium.

Graph 2: Heat Content of Pellets

92

93

94

95

96

97

98

99

100

PD

I

Pellet Durability Index

20 of 22



NEXT STEPS

The SOAR research funded by this grant identified the operating parameters of the pelletizing process for this PTO and electricity powered Buskirk mill. Our results are promising, but more work needs to be done to further define the factors required for successful processing and to develop a “recipe” for pelletizing success.

First, we need to determine what blends of wood and grass produce a durable pellet, and what additives, if any, could be introduced to create a stronger pellet. Alongside this, we should develop a reliable method to quickly determine the moisture content of the feedstock. The microwave test worked well but still took about 30 minutes to test the feedstock so we had to wait on production until the test was complete.

Secondly we have only been able to consistently produce pellets at about 60 lbs/hr, though the Buskirk system is said to be capable of producing 800 lbs/hr. A next step would include identifying the characteristics that constrain pellet production rate. Preliminary testing shows a strong correlation between moisture content and production rate.

Lastly we will continue to test these pellets for their heat content with the Parr 1341 Bomb Calorimeter. The standard operating procedures for the bomb calorimeter are attached in the appendix section of this report. Graph 3 shows that our grass pellets have slightly lower heat content than wood ones. These results are approximate, because we were only able to test 1-2 samples of each type of pellet. As we do more tests we will be able to compare different types of feedstock and different pellets.

Taking these steps will allow us to more clearly refine the pellet production process. Wood and grass blends could help create a niche market for these types of pellets. Further research into increasing production rates could make this particular unit feasible as a mobile pellet production facility. If production could be increased with this unit, it could be used similar to the way that some people use mobile sawmills. Someone could have some land that they are logging off, and rather than sell the logs they could have you come in with a pellet mill and means of chipping up the logs and make pellets for them for their home.

Graph 3: Heat content of finished pellets

0

1000

2000

3000

4000

5000

6000

7000

8000

9000

BT

U/l

b

BTU Comparison of Finished Pellets

21 of 22

A special thanks to the many collaborators and advisors in this project:

John Abley, owner, Meach Cove Farm Trust, Shelburne, VT

Chris Davis, manager, Meach Cove Farm Trust, Shelburne, VT

Roger Rainville, Borderview Farm, Alburgh, VT

Sam Lincoln, Lincoln AgriSource LLC, Randolph Center, VT

Sosten Lungu, Vermont Technical College, Randolph Center, VT

Paul Salon, USDA-NRCS Plant Materials Station, Corning, NY

Jerry Cherney, Extension Forage Specialist, Cornell University

22 of 22

Documents

Grass Biomass Energy Field Day Proceedings