Development of a Flail Harvester for Small Diameter Brushand Coppiced Trees to Produce Energy/Chemical Feedstock

Robert A. McLauchlan, Andrew Conkey and Greg Scherer, Department of Mechanical and Industrial Engineering

Peter Felker, Center for Semi-Arid Forest Resources, Caesar Kleberg Wildlife Research Institute, Texas A&M University-Kingsville, Kingsville, Texas 78363

Stan Brown, Brown-Bear Corporation, Corning, Iowa

Abstract

The design, fabrication and field testing is described for a harvester to produce shredded woody biomass from small diameter natural brush stands and coppiced energy plantations. In the southwestern United States, typical applications include harvesting of mesquite, juniper and sage brush, while in the southeastern United States, typical applications would include precommercial thinning of pine stands too small for pulpwood, and thinning of pioneer hardwood stands. These small diameter stems have not been economically harvested to-date because the brush is too small to be handled by conventional forestry equipment, yet too woody to be harvested by conventional agricultural equipment. A harvester was built on a John Deere 210 kw forage harvester. A flail shredder and auger conveyance system was used to sever the trees, shred them and blow them behind the harvester in a vehicle pulled behind the harvester. The harvester was examined for 3 weeks in the southeastern United States in 2 m tall coppiced sweetgum, sycamore and water oak, in 4 m tall seedling sycamore, in 2.4 m tall pines, in 6 m tall pines and in seedling hardwood site with 4 to 9 m tall sweetgum. When the trees were less than 3 m tall and less than 6 cm in diameter they were severed, captured and blown behind the harvester with little difficulty at 3.6 km/hr (0.9 ha/hr). Trees greater in diameter than 6 cm were not completely severed due to a non-overlapping knife pattern in the cutterhead. Nevertheless 12 cm diameter pines and sweetgum trees were mowed down and partially captured at a rate of 1.6 km/hr (0.4 ha/hr). The operating cost of the harvester was estimated to be about $75/hr.

Economical operation of the harvester will require both a silviculture need for thinning and an energy market for the chips. The current version of the harvester is not suited to typical forestry terrain. It will be necessary to build the harvester on a 4 wheel drive, high clearance frame (such as with skidders) in which all components are located internally. A team of forestry companies, equipment manufacturers, universities and government would be best suited to develop this harvester.

Introduction

Our economic analysis indicated that for biomass farming operations in the southwest, over 50 percent of the final delivered cost of the biomass was attributable to the harvesting and transportation of the biomass (Felker 1984). Similarly we believe the major economic constraint to competitively priced biomass production from rangeland brush will lie in the harvesting and transportation. Our particular interest lies in harvesting of mesquite (Prosopis) since it exists on 30 million ha in the United States (Parker and Martin, 1952) with 22 million ha in Texas. Since the standing biomass in mesquite stands ranging from 6 to 18,000 stems/ha varied from about 3 to 40 dry metric tons/ha (Felker and others, 1988), the total biomass resource for mesquite in Texas would range from about 60 to 800 million tons. If expressed in the U.S. Department of Energy units of quads (1015 BTU), the energy resource from Texas brush would range from 1 to 15 quads (at 18 million BTU/metric ton). Total U.S. consumption of energy is about 80 quads and therefore the woody biomass in Texas could be very significant. In South Texas, dense regrowth containing 5 to 10 dry metric tons/ha can occur in 15 years. This renewable resource should be available indefinitely.

Present mechanized forestry harvesting equipment typically harvests 300 38-cm diameter trees per hour (Cullen and Barr, 1981). However, as the stem diameter class decreases, the harvesting cost per ton increases since it is necessary to harvest more stems per ha. As shown in Fig. 1, Kluender and Plummer (1980) found that when the average stem diameter size class decreased below 15 cm, the harvesting cost per ton rose exponentially reaching a cost of $54/ton for 5.1 cm diameter stems. In contrast, Felker and others (1988) measured mesquite densities as high as 18,000 stems/ha with an average biomass of 0.8 kg/stem. These stems cannot be treated individually and require the development of a swath harvester.

Biomass Harvester Development

Significant work regarding the economical harvesting of rangeland brush for energy and chemical feedstocks has been pursued at Texas Tech and Texas A&M University - Kingsville in the 1980s. A "Texas Tech biomass combine" (Ulich,1983) was reported to harvest 4 tons/hr in a 1.52 m wide swath, at a cost of $12.40 per ton of 12 percent moisture chips, with an energy cost of $0.73/million BTUs. If wood chips could be obtained at the rate of 4 ton/hr, with an operating cost of about $30/hr, this would be a cost of about $7/ton.

The basic principle behind the flail harvester is shown in Figure 2. Harvesting Method for a Horizontal Shaft Cutter.

The trees are bent over at about a 55 degree angle using a push-bar. The tree stem is designed to come into contact with the center of the flail shredder which provides the double action of severing the stem and forcing it upwards between the cutterhead knives and the shroud. As the stem progresses farther backwards, the knives continue to reduce the size of the stem as the clearance between the knives and the shroud becomes less. Ulich patented the geometry of this system in which he used 2.5 cm square bars that were welded beneath the shroud to create stages of diminished clearance and thus reduced stem size classes. After the stems have passed through the 4 stages of clearance they are captured in a trough with a flow through auger. This auger strips the head of chips and is the first stage of the materials handling system. The last bar only has about 6 mm of clearance.

The Texas Tech harvester was mounted on a farm tractor with a standard mechanical transmission. The severing and shredding head was mounted on the front of the tractor with front-end loader arms. The cutterhead consisted of a 64 cm diameter horizontal cutting cylinder with 26 10-cm wide stirrup-type knives having a 360 degree swing circle. The cutting blade speed was 2,600 to 3,000 m/min. A 2 ton cotton module basket capable of dumping the chips onto a truck bed was used. The elevator used to take the chips from the cutting head to the basket, consisted of a chain drive with 25 cm semi-circular paddles. The power for the cutting head and elevator assembly was obtained from an 80 kw tractor through a hydraulic pump mounted on the PTO shaft of the tractor.

In year long field trials at Texas A&M University - Kingsville the flail shredder principle used for simultaneously severing and shredding the brush proved satisfactory. However, other aspects of this harvester were most inadequate.

Texas A&M University - Kingsville Biomass Harvester

We significantly modified the Texas Tech (TTU) biomass harvester design discussed above. The major features of the Texas A&M University - Kingsville (TAMUK) design include the following:

1. Use of a John Deere Model 5280 type silage harvester with (a) hydrostatic drive, (b) more powerful (213 kw) engine, and (c) radiator in the rear. The hydrostatic drive was necessary to avoid repeated clutch failure caused by the constant clutching that was necessary to maintain full engine rpm for the cutting head while keeping the forward movement slow enough to avoid clogging the cutterhead. Greater horsepower was required for the cutterhead and associated chip transfer system. The radiator in the rear was necessary to reduce frequent clogging of the radiator by dust from the cutter-head located in front of the tractor causing the engine to overheat.
2. Use of a straight through design of the materials handling/conveyance system from the back of the cutterhead to the trailer. This was necessary to eliminate the chip clogging problems encountered with the previous TTU design that included two right angle turns involving screw augers and chain driven paddles.
3. A redesign of the cutterhead support system to avoid cracking of the front end loader arms and failure of the front rims and lug bolts caused by the excessive weight of the cutterhead. Since the forage harvester operates "backwards" as compared to traditional tractors, the large axle and tires were available to support the weight of the cutterhead and frame.
4. Use of foam filled and forestry tires with reinforced rims to eliminate (a) flat tire problems in the field and (b) failure of the front rims and lug bolts.

Figure 3. Texas A&M-Kingsville Biomass Harvester during Spring 1994 Field Test shows the Texas A&M - Kingsville biomass (mesquite/rangeland brush) harvester in action during field trials in Spring 1994. It shows the tractor with the cutterhead in operation and moving forward to harvest brush. This is the opposite of conventional tractors. The remainder of this section discusses the steps involved in the development of the biomass harvester design.

In 1989 we purchased a used John Deere 5280 forage harvester and mounted the original 1.5 m Texas Tech cutterhead and associated hydraulics, but without the materials handling system. This system performed quite well in harvesting and chipping mesquite. (It is relevant to point out that mesquite has a specific gravity of 0.7 kg/l and is most similar to hickory in hardness and density).

To test suitabilty for an improved chip transfer system, shredded chips were gathered from the ground for use in a materials handling test bed. We found that the 2 23-cm diameter augers that came with the forage harvester quite adequately transferred the chips to the blower paddle assembly and out the spout. The materials handling system even successfully transferred some stem pieces as long as 40 cm. Thus, it appeared that the transfer of the chips from the cutterhead to the blower paddles would not create a problem.

Therefore, a materials handling system was designed that consisted of (1) a cross auger with main axis parallel to the cutterhead to strip the chips from the auger and bring it to the center of the cutterhead, (2) 2 23- cm flow-through augers to transfer the chips from the cross augurs immediately behind the cutterhead under the tractor to the tractor blower assembly, and (3) the tractor blower assembly rotating at 1,000 rpm from the PTO shaft that blows the chips into a van behind the harvester.

Since we had more horsepower than the 80 kw engine used in Ulich's design, we would be able to drive a wider cutterhead. While Ulich fabricated his own flail cutterhead, we desired to obtain a commercially available flail cutterhead to facilitate rapid commercial adoption of this technology. After a review of commercially available flail cutterheads and discussions with cutterhead manufacturers, we chose to employ the 2.6 m wide Brown Bear Flail cutterhead. A joint development agreement between Brown Bear and Texas A&M Kingsville was signed to develop a new harvester using their cutterhead.

Brown-Bear used this cutterhead in a 200 kw brush clearing machine to shred the chips and leave them on the ground. In the Brown Bear design, the rotation of the cutterhead was such that the top of the cutterhead moved in the same forward direction as the harvester. In the Ulich and TAMUK design the cutterhead rotated in the opposite direction, cutting the stems and bringing the severed stems over the top of the cutterhead between the shroud and the cutting knives. Due to the differing objectives of chipping the severed stems on the ground by the Brown-Bear cutterhead, and of severing the stems and dragging them over the top of the cutterhead in our design, there were fundamental differences in the geometry requirements of the cutterhead knives.

Having established basic requirements for the cutterhead geometry and the materials handling system, we found it necessary to design a frame to house the cutterhead and augers and to design a hydraulic system for the cutterhead and cross augers as well as the flow through augers.

Cutterhead/auger housing frame:

A weight of 2,268 kg at a distance of 1.22 m from the center of the drive tires of the tractor was used to design the frame to support the cutterhead and auger housing. Sway bars were also installed to prevent side to side movement of the head. To enable the cutterhead to be raised and lowered, hydraulic cylinders were mounted at the front of the tractor. These cylinders were powered by tractors hydraulic system and used 2 way check valves to prevent them from creeping up or down. One end of this frame was attached with pins that allowed it to pivot up and down, under the bottom of the tractor near the center.

Hydraulic system:

A closed loop hydraulic system capable of transferring the full 213 kw of the tractor engine to the cutterhead was designed with the assistance of Womack Hydraulics of San Antonio, Texas. The complete hydraulic system used the 76 l/min - 16 MPa system on the tractor as well as 3 hydraulic pumps, an auxiliary heat exchanger and 5 hydraulic motors.

The main hydraulic system that drove the cutterhead was driven with a fixed displacement pump coupled to the 1000 rpm tractor PTO shaft. This system was designed to transfer 360 l/min of 32 MPa fluid to two hydraulic motors on either side of the cutterhead. A high tension cog belt connected the tractor PTO shaft to two smaller pumps connected by the same shaft. One of these pumps was a charge pump used to keep the main pump primed. The other pump provided fluid for the two hydraulic motors for the cross augers and the flow through augers. Variable flow dividers were used to regulate the speed of these augers.

An auxiliary heat exchanger to cool the hydraulic fluid from the cutterhead was mounted on the back of the tractor. The fan for this heat exchanger was powered from the tractors hydraulics.

Preliminary Field Evaluation of Harvester on Pines and Stweetgum in the Southeastern United States

In the spring of 1994, the biomass harvester was transported to Alabama for a demonstration for a Short Rotation Intensive Culture (SRIC) mechanization workshop and then to South Carolina and Georgia for field trials harvesting dense pine stands and sweetgum regrowth sites.

At the Alabama site, the harvester was examined on 1 year old coppice regrowth from several year old sycamore, sweetgum, and water oak stools, and on 3 year old sycamore seedlings. When sycamore, sweetgum or water oak stems were less than 5 cm in diameter and the height less than 2.4 m tall, the harvester passed through the material at about 3.5 km/hr capturing nearly all of the stems and then passing it through the blower. Given the 2.6 m width of the cutterhead, a 3.5 km/hr forward speed corresponds to 0.91 ha/hr.

There was concern that the flail cutterhead did not cut the coppice cleanly enough to allow adequate resprouting. Clearly the severed stems were split several cm below the site of the cut. If shoots will emerge below the location where the stems are split, it should be possible to have a high percentage resprouting by cutting the shoots higher on the stump. It is important to base the decision on whether to pursue development of a harvester for hardwood coppice upon quantitative measurements of percentage stool resprout as a function of knife geometry rather than conjecture.

While the harvester was capable of mowing down the 4.5 m tall 10 to 12 cm diameter sycamore seedlings, the trees were not severed and captured by the shredder. We found that a major contributing factor to the failure to capture the 4.5 m tall sycamore trees was the geometry of the cutterhead. The knives on the cutterhead extended only 5 cm beyond the disk that held them to the head. Since none of the knives overlapped, the maximum depth of cut was only 5 cm. As a result, trees with a greater diameter than 5 cm were merely pushed over and 2 parallel 5 cm deep gouges extending the length of the tree were found on the trees.

When operating in the southeast, it was necessary to operate the engine at 1,700 rpm rather than the desired 2,100 rpm due to excessive pressure on some hydraulic components. This also slowed the blower paddle speed giving the chips less trajectory and provided less horsepower to the cutterhead. This problem is currently being corrected in Texas.

The harvester was examined on 2 pine sites near Greenwood, South Carolina. Both sites were hand planted, but large numbers of additional pine seedlings became established from natural seedlings. At the first site, the originally planted trees were about 13 cm in diameter and 6 m tall. The naturally established pine seedlings were less than 6 m tall but so thick that it was often not possible to discern where the original rows were. At the second site of pine, the trees were typically 1.8 to 2.4 m tall. As discussed later, some portions of this field had pine seedling densities as high as 13 trees per square m (130,000 trees/ha). On this site 1.8 m square plots were clipped to measure the standing biomass.

The goal at both of the South Carolina pine sites was to clear 2.6 m wide paths, leaving a width of 0.5 m between the 2.6 m wide lanes. The trees in the 0.5 m width could then grow faster with reduced competition. Six lanes about 230 m long were made in the 6 m tall trees at 1.6 km/hr, before it was decided these trees were just too large and the terrain too difficult for the harvester.

In contrast to the flat agricultural type field sites in Alabama, the South Carolina sites had old windrows, 30 cm diameter rocks and ditches that required that the cutterhead be maintained about 10 to 20 cm above ground. This in turn, created 10 to 20 cm diameter stumps that precluded pulling an agricultural type trailer behind the harvester.

On two occasions the harvester became stuck despite no indications of moisture from the surface. This was no doubt compounded by the fact that the 2,270 kg head was mounted in front of the front tires. It is clear that future harvesters will have to be mounted on a 4 wheel drive "skidder type" frame to provide greater maneuverability, clearance and traction in moist conditions.

At the second South Carolina site the trees were 1.6 to 2.4 m tall with pine tree densities as great as 130,000/ha. At this site the harvester was capable of mowing and harvesting the trees at 3.6 km/hr thus clearing 0.93 ha/hr. If one included the 0.5 m width in which the trees remained, then the area treated for thinning was about 1.08 ha/hr. The harvesting speed at this site was more dependent on the terrain than the amount of biomass to be processed.

Following testing at the South Carolina pine sites, the harvester was tested on a sweetgum regrowth site near Athens, Georgia. The sweetgum was typically 5 to 10 cm in basal diameter and 6 m tall but there were some 15 to 20 cm basal diameter 9 m tall trees. This site was scheduled for bulldozing, stacking into windrows in preparation for planting back to pines. To measure the biomass at this site, trees were harvested and weighed for regression equation development in three plots of 40 square m.

The following regression equation was developed based on 13 trees ranging in circumference and weight of 1.04 cm and 0.68 kg to 6.00 cm and 15.25 inches and 28.8 kg.

log biomass (kg)=2.145908 log circumference (cm) - 1.90783

The detransformed r square was 0.928 and the root mean square error 2.13 kg.

The biomass estimates were derived from basal diameter measurements on 3 plots of 40.5 m2. By applying the regression equation to the 3 plots we estimated the biomass/ha as follows:

• Plot 1 with 25 trees had a brush biomass of 89.72 kg (22,161 kg/ha) (each of the plots were 40.5 m2).
• Plot 2 with 19 trees had fresh biomass of 117.26 kg (28,963 kg/ha).
• Plot 3 with 23 trees had a fresh biomass of 129.87 (32,078 kg/ha).

Thus, the mean biomass/ha was 27,733 kg/ha with a standard deviation of 5,078 kg/ha. The mean and standard deviation for the number of trees per ha was 5,515 +/- 755 trees/ha.

When harvesting the sweetgum trees at 1.6 km/hr, we harvested about 0.4 ha/hr. Given the fact that the standing biomass of this site was 27,731 kg/ha, the theoretical maximum capture would be 11,092 kg/hr. Thus, there is enough biomass at this site to provide 11 ton/hr of biomass which at $9/green ton would be a $90/hr credit. If 85 percent capture could be achieved, the harvesting could generate $76/hr of revenue.

While we had mechanical problems at this site, related to trees hitting exposed hydraulic lines, we cleared about 1 ha at a harvesting rate of about 1.6 km/hr (0.4 ha/hr). For the first time at this site, the power of the harvester became limited in harvesting isolated 17 cm basal diameter, 9 m tall sweetgum trees which fell in front of the harvester and entered the cutterhead to be chipped. This resulted in slower cutterhead and engine speed which resulted in some large wood pieces (15 cm by 4 cm) becoming lodged in the augur. Nevertheless, with improved cutterhead geometry, and the harvester operating at 2,100 rpm instead of 1,700 rpm (due to excessive hydraulic pressures at heat exchanger), it should be possible to routinely harvest 10 to 12 cm diameter sweetgum trees at 1.6 km/hr.

Conclusion

In summary, the basic concept of the harvester worked well but it will be essential to mount the cutterhead on a skidder type frame if it is to operate in the woods, as opposed to short-rotation coppiced fields on farm land. It appears as if harvesting rates of 0.4 to 1.0 ha/hr are possible. Based on the experience in operating a flail shredder, one of us (SB) estimates the operating cost of this harvester (with operator) to be about $70/hr. Harvesting in light stands of 1.6 to 2.4 m tall trees at 3.5 km/hr (0.9 ha/hr) would cost about $76/ha. If in addition, 4 ton of green biomass at $9/ton could be harvested, the thinning cost would be $40/ha. This compares favorably to costs of $384/ha for bulldozing, stacking and burning in windrows. Furthermore the absence of soil disturbance or windrows leaves the site in a better condition for planting.

It is clear that the successful economic operation of this harvester will require both a silvicultural need for thinning pines or hardwoods, and a market for chips used for energy. There are thousands of ha in the southeastern United States where thinning hardwood and volunteer pines is desirable, but not economically feasible, due to the high cost of manual labor and the unavailability of suitable equipment. By combining a silvicultural need with a market for energy, it should be possible to economically thin these stands, allowing them to reach their full economic potential.

Upon our return to Texas, we are making modifications to allow operation of the harvester at full engine rpm by correcting the hydraulic overload on the heat exchanger. This will allow greater power at the head and increase the discharge speed of the chips from the blower. In the summer of 1994, we will mount a revised cutterhead with overlapping tooth geometry in the harvester.

Neither the government, nor industry, nor the university community has the time or financial resources to solve this problem alone. At this point there exists (1) a prototype machine that serves as an excellent test bed for commercial development, (2) a university with staff and students committed to further development, (3) a commercial forestry equipment manufacturer (Brown Bear) interested in building the machines, (4) government agencies (SERBEP) and Western Area Power Authority providing modest funds, and (5) a need of pulp and paper companies to more economically thin stands.

It would appear prudent to create a working group among private forestry companies, universities, government and equipment manufacturers to provide their own unique capabilities to further the development of such harvesters. It will be necessary to examine the engineering aspects, the ecological aspects of site harvesting, sylvicultural aspects, economics, and the matching of the biomass harvested with end-users. A pragmatic approach would be to make the modifications necessary based on the southeastern USA spring 1994 trip (including better cutterhead design) and then seek support from commercial forestry companies to reexamine the machine in the fall of 1994 in the southeast. Based on those results it would be best if several commercial forestry companies could jointly share the purchase of 1 or 2 harvesters (about $230,000), incorporating the best designs, and built on forestry skidder frames. Following the field testing of the new skidder based machines the working group would re-evaluate costs, engineering aspects, sylvicultural aspects, biomass fuel markets, and site ecological aspects for future development.

Acknowledgement

The biomass harvester work reported in this paper was funded under the Texas/DOE Energy Research in Applications Program, The Texas A&M Engineering Experiment Station, The USDA/CSRS, and the U.S. Department of Energy Southeastern Regional Biomass Energy Program (SERBEP). We are most grateful to Tom Morgan and Jim Dewitt of Scott Paper Company and to Bill Moore, Virgil Wall and John Cheatham of Canal Wood Corporation for providing sites and logistical support in Alabama and South Carolina, and in Georgia, respectively.

Bibliography

1. Cullen, D.E. and Barr, W.J. 1981. Harvesting of Close Spaced Short Rotation Woody Biomass. Report to the US Department of Energy. National Technical Information Services, Springfield, Virginia, 99 pp.
2. Felker, P. 1984. Economic, Environmental and Social Advantages of Intensively Managed Short Rotation Mesquite (Prosopis spp) Biomass Energy Farms. Biomass 5: 65-77.
3. Felker, P, J.M Meyer, and S.J. Gronski. 1988. Application of Self-Thinning in Mesquite to Range Management and Lumber Production." Forest Ecology and Management 31:225-232.
4. Kluender, R.A. and G. Plummer. 1980. Whole Tree Chipping for Fuel: The Range of Diameter Limits. American Pulpwood Association, Report No. 80-A-19. 6 pp.
5. Parker, K.W. and S.G. Martin. 1952. The Mesquite Problem on the Southern Arizona Range, USDA Circular 968, 70 pp.
6. Ulich, W.L. 1983. Development of a Biomass Combine. Texas Tech University. Report No. T-3-103.

File posted on March 5, 1996; Date Modified: February 21, 1999