CONTINUOUS FLOW BIODIESEL PRODUCTION

C. L. Peterson, J. L. Cook, J. C. Thompson, J. S. Taberski

ABSTRACT. Biodiesel, which consists of the fatty acid esters of simple alcohols, is a potential replacement for a portion of the
diesel fuel used in transportation. It is produced from both used (oil that has been utilized for frying and discarded) and new
vegetable and animal fats and oils. It has several advantages. Among these advantages are its classification as a renewable
resource, its ability to reduce HC, CO, and CO2 exhaust emissions, its non–toxic character, and its biodegradability. One of
the keys to making biodiesel a viable and profitable energy source is the use of a continuous flow transesterification process
to reduce time and cost, thereby increasing production and profit.
A continuous flow esterification process for producing biodiesel from rapeseed oil and ethanol was investigated. The
equipment consisted of an oil metering pump, centrifugal mixing pump, ethanol and catalyst metering pump, static mixers,

ladder" type retention reactor, water injection system, and continuous flow centrifugal separating system. The oil feed rate
was 0.38 L/min that yields about three times the weekly production of the existing batch type transesterification system that
produces 945 L (250 gal) per batch. It was anticipated that if methanol were used instead of ethanol that the flow rate could
be considerably increased.
The system in its present configuration has met ASTM PS121–99 standard for free and total glycerol. The centrifugal
separation resulted in release of excessive alcohol vapors. A vapor recovery and condensing system should be installed. This
latter system could help further reduce costs of the biodiesel produced by recycling some of the alcohol used.
Keywords. Biodiesel, Transesterification, Continuous flow transesterification.

U
se of biodiesel, which consists of the fatty acid 1998). Carbon dioxide contributions to greenhouse gas
esters of simple alcohols, derived from vegetable emissions are reduced with biodiesel because the CO2
oil and animal fatas as a renewable fuel source resulting from combustion would be consumed by the growth
may provide benefits to the environment. Among of plants used in to produce the oil (Peterson and
its advantages are reduced HC and CO exhaust emissions, Hustrulid, 1998).
non–toxicity, and biodegradability. Carbon dioxide The most widely used method for biodiesel production
production in biodiesel combustion is offset by the carbon utilizes a process called transesterification that uses an
dioxide use of the plant in photosynthesis. Nitrous oxides and alcohol (methanol or ethanol) in the presence of an alkaline
particulate matter changes depend on the particular test catalyst, such as sodium hydroxide or potassium hydroxide,
conditions. However, exhaust particulate matter toxicity and to chemically cleave the raw vegetable oil molecule into
mutagenicity have been shown to be reduced. Biodiesel is methyl or ethyl esters of the renewable oils with glycerol as
biodegradable with potential less harm to the environment in a by–product (fig. 1).
case of a spill. Additionally, vegetable oils contain negligible University of Idaho personnel are producing
levels of sulfur and result in reduced sulfur dioxide biodiesel–using ethanol and KOH in the transesterification
emissions, a key component of acid rain (Peterson and Reece, reaction instead of methanol as is used in most biodiesel
1996; 1996b). production plants. Ethanol is the preferred alcohol by
Petroleum fuels naturally degrade at a much slower rate researchers at the University of Idaho because it is derived
than does biodiesel in the environment and therefore poses a from agricultural products, is renewable, and is biologically
soil and groundwater hazard in cases of spillage and leakage.
Biodiesel fuel is biodegradable and will break down rapidly,
preventing long–term damage to soil or water (Zhang et al.,

Article was submitted for review in May 2000; approved for

publication by the Power & Machinery Division of ASAE in July 2001.
Approved as paper no. 00304 of the Idaho Agricultural Experiment
Station.
The authors are Charles L. Peterson, ASAE Fellow Engineer,
Professor and Interim Head, Jarod L. Cook, Former Graduate Assistant,
Joseph C. Thompson, Engineering Specialist, and Jeffrey S. Taberski,
Former Engineering Specialist, Department of Biological and Agricultural Figure 1. The esterification chemistry involved in production of fatty acid
Engineering, University of Idaho, Moscow, Idaho. Corresponding esters from a triglyceride. R1, R2, and R3 are hydrocarbon chains of fatty
author: Charles L. Peterson, E/P 423 Dept. of Biological and Agricultural acid radicals specific to the particular vegetable oil involved in the
Engineering, University of Idaho, Moscow, ID 83844–0904; phone: reaction. R4 is the radical identifying the particular alcohol being used in
208–885–7906; fax: 208–885–7908; e–mail: cpeterson@uidaho.edu. the reaction, R4 is CH3 for methanol and C2H5 for ethanol.

Applied Engineering in Agriculture

Vol. 18(1): 5–11 E 2002 American Society of Agricultural Engineers ISSN 0883–8542 5
less objectionable in the environment (Peterson et al., 1996). intensity, stoichiometry, and catalyst concentration on the
Compared to methanol, ethanol is safer to handle because overall conversion of vegetable oils to esters.
toxic effects to personnel from exposure to the fumes are re- The reaction materials used by Noureddini et al. (1996) to
duced. Ethanol is more expensive than methanol and cannot produce biodiesel were refined and bleached grade soybean
be derived as readily as methanol. For those reasons metha- oil, anhydrous methyl alcohol, and sodium hydroxide. A
nol is more widely used for biodiesel production than is etha- solution of 25 g of sodium hydroxide per 1 L of methanol was
nol. Benefits of using a continuous flow system include: prepared for mixing with the soybean oil in the continuous
S Production cost and time for producing a small amount of flow system. The article describes the setup of the pilot plant
fuel is high using current batch reactor methods. in detail. The major operations of the system included a pump
S A continuous flow process would allow more fuel to be station, preheating/mixing, mixing, a residence tube,
produced per unit of labor and, on a larger scale, reduce cooling, and settling.
cost. Peterson et al. (1991) discussed the characteristics of the
S With design improvements to optimize equipment, batch type transesterification process for winter rapeseed oil.
continuous flow has the potential for consistently The main variables in the process were: catalyst type, catalyst
producing a higher quality fuel. concentration, oil/alcohol ratio and stirring rate. They found
that 1% NaOH or KOH effectively enhanced the reaction rate
OBJECTIVES at room temperature (a 60–min reaction time was allowed).
Current batch methods of producing biodiesel are slow, A 6:1 molar ratio of methanol to oil gave the best conversions
tedious, labor intensive, and not well adapted to automation. in a batch system. In addition, the rate of reaction was
The objective of this project was to plan, construct, and test satisfactory if the stirring action was vigorous with some
a continuous flow transesterification apparatus for producing splashing.
biodiesel. The literature was reviewed to determine probable Biodiesel composed of hydrogenated soy ethyl esters has
flow–rates, probable raw product quantities, and previous been termed HySEE (Lowe et al., 1998). The
experiments to determine a process that would have a good transesterification process consists of four steps: 1) raw
chance of operational success. material pre–treatment; 2) ethyl ester transesterification;
Other objectives of the project were to: 3) separation of the ester and glycerol phases; and
S Develop design specifications and build a prototype 4) purification of the ethyl esters. The first two steps of this
system. procedure require the addition of heat to the oil and its
S Modify the prototype based on initial tests for better reacting apparatus, because the waste oil is solid at room
performance. temperature. A similar procedure can be used in the
S Analyze biodiesel produced for comparison with batch continuous transesterification process for the conversion of
reactor and ASTM standards. used potato processing oil into HySEE biodiesel fuel in the
future.
Two patents were found relating to biodiesel production.
Bradin (1996) described a method for producing a unique
LITERATURE REVIEW biodiesel fuel that is outside the normal definition of
Typical fuel properties of both raw rapeseed oil and biodiesel being a fatty acid ester of simple alchohols. In his
modified esters derived from rapeseed oil are included in patent he discloses a fuel additive composition that includes
table 1 for comparison with diesel No. 2. fatty acid alkyl esters and glyceryl ethers, as well as a method
Noureddini et al. (1996) detailed an approach for a for preparing the composition. The fuel additive composition
continuous process for the transesterification of triglycerides can be added to diesel fuel to provide an alternative diesel
to methyl esters. This process is based on a combined high fuel composition. Continuous flow production was not a
shear and motionless reactor system. Their experimental claim of the patent. Bam et al. (1995) discussed a method for
studies explored the available variations in the mixing purifying alcohol esters in a continuous batch process. No
patents were discovered that covered a continuous flow
Table 1. Typical fuel properties for raw rapeseed oil,
esters of rapeseed oil, and diesel no. 2.[a]
biodiesel production process as discussed in this article.
Raw Methyl Ethyl
Properties Rapeseed Ester Ester Diesel
Heat of Combustion, 40.40 40.72 40.70 45.71 MATERIALS AND METHODS
MJ/kg (Btu/lb) (17370) (17506) (17500) (19652) Transesterification requires the basic materials including
273.9 170.0 185.0 80.0 vegetable oil [in this case canola, used hydrogenated soy oil
Flash Point, _C (_F) (525) (338) (365) (176)
(used cooking fat), or rapeseed oil], ethanol, potassium
–11.1 –3.3 –1.1 –13.9
hydroxide. Tap water was employed to facilitate product
Cloud Point, _C (_F) (12) (26) (30) (7)
separation from other reaction components. The ethanol was
–15.0 –14.4 –20.6 –27.8
Pour Point, _C (_F) (5) (6) (–5) (–18) 200 proof industrial grade ethanol supplied by the J. R.
Viscosity, CSt 46.7 5.7 6.1 3.51 Simplot Company (Caldwell, Idaho). The J. R. Simplot
Sulfur, % wt 0.022 0.009 0.008 0.36 Company produces ethanol from waste potato products and
0.907 0.880 0.876 0.849 generally supplies the ethanol into the motor fuel market for
Density, kg/L (lb/gal) (7.57) (7.34) (7.31) (7.07) use in gasohol.
[a] University of Idaho, Biological and Agricultural Engineering, Equipment used in the project included two different Alfa
Analytical Laboratory Data. Laval continuous flow centrifuges (Alfa Laval Model WSB
103B–74–60 and Alfa Laval Model Gyro–Tester); oil, ester,

6 APPLIED ENGINEERING IN AGRICULTURE

and glycerol containers; a Fluid Metering, Inc. (FMI)
metering pump (A. O. Smith , Model QDX1, 0 to
550 mL/min); a hydraulic motor–driven centrifugal mixing
pump (Delaval Mfg.); an oil metering pump (Teel, model
1P771 gear pump with 1/3–hp Pacific Scientific DC
permanent magnet motor and a Dart Controls, Inc. model
253G–200E speed controller); and oil flow meter (Omega
model FL–2100). In–line static mixers were obtained from
Cole–Parmer (stock no. H–046667–08, unmounted OD
.5 in., 12 elements per 6.25 in. length). Water flow was
controlled with a Parker no. 600S needle valve monitored
with an M and M, model 406 PX gasoline flowmeter.
Two different continuous flow set–ups were examined in
this project. These will be referred to as CF–1 and CF–2.
Figure 2 is a schematic of the components and flow path of
Figure 3. Continuous flow system (CF–1). The first attempt at a reactor
the final or CF–2 continuous flow system. for the continuous flow system. It consisted of about 30.5 m (100 ft) of
25.4–mm (1–in) clear polyvinyl chloride tubing in a helical pattern. Also
CONTINUOUS FLOW SYSTEM 1 (CF–1) shown in this picture are the centrifugal mixing pump to the left of the
CF–1 consisted of the centrifugal mixing pump, FMI reactor and the FMI metering pump on the shelf to the left and rear of the
reactor.
alcohol/catalyst metering pump, static mixers placed in clear
polyvinyl chloride tubing followed by 100 ft of 1–in. I.D.
clear polyvinyl chloride tubing arranged in a circular pattern separation rate of the glycerol from the ester. The mixture
around a plastic barrel (fig. 3) and either the gyro–tester was then fed into the midpoint of a settling tank. When the
centrifuge or a settling tank used to remove the glycerol. The settling tank was full and had reached equilibrium, ester was
taken off the top and glycerol was drained off the bottom in
settling tank was used on test runs before the centrifuge
became available. proportion equal to the inflow rate.
The first test made with CF–1 was with used frying oil. Subsequent testing with CF–1 involved the use of an
Alpha Laval gyro–tester continuous centrifuge (fig. 4) to
Since this oil is solid at room temperature the reservoir was
heated with a 120–V barrel heater. It was placed on a spring separate the ester from the glycerol phase. Using this system,
scale in order to monitor flow rate by weight loss. The mass a second test was conducted with used frying oil and a third
with canola oil. The canola oil, because of its lower pour
flow rates of oil and alcohol/catalyst were based on
calculations made using the batch process recipe. After being point, eliminated the requirement for heating. Oil flow rate
mixed with alcohol/catalyst and fed into the retention was difficult to monitor and control and the retention time of
the system was only 30 min resulting in less than desirable
network, flow was not a problem. Water was added to the
reactants at the exit end of the retention tubing to increase the conversion efficiency.

Water

Pressure
Regulator

Needle Valve

FMI EtOH and

Pump KOH
Flow Meter

Oil
Retention
Centrifugal Nozzle
Mixing
Network Continuous
Flow
Pump Centrifuge
Flow
Regulating
Pump

Static Mixers

Figure 2. Flow chart showing the components and flow path for the improved (CF–2) continuous flow esterification system.

Vol. 18(1): 5–11 7

 Figure 5.
Ladder type" retention reactor used in the final design for the
Figure 4. Alfa Laval model WSB 103B–74–60 continuous flow centrifuge
continuous flow biodiesel production system. The reactor consists of
used for ester and glycerol separation.
33.5 m (110 ft) of 31.8 mm (1 1/4–in) PVC arranged in 11 steps with 1.8 m
(6–ft). of 22–mm (.5–in) PVC vertical exit section. Nine feet of clear PVC
lines with static mixers used in the feed and exit lines are not visible.
CONTINUOUS FLOW SYSTEM 2 (CF–2)
CF–2 was a second–generation continuous flow system (0.10 gal/min). At the same time, the FMI metering pump
that was constructed using ideas learned during the regulated the flow of alcohol/catalyst at a rate of 111 mL/
development of CF–1. Two major changes distinguish CF–2 min (0.029 gal/min). The catalyst and oil were mixed with
from CF–1. First, a second metering pump was placed ahead high shearing action in the hydraulically driven centrifugal
of the centrifugal mixing pump to more accurately meter the mixing pump. The mixture exited the mixing pump, passed
flow of vegetable oil into the mixer. Second, a redesigned through a series of static mixers, and into the retention reac-
retention reactor was constructed as shown in figure 5. For tor. As the biodiesel exited the retention network, it passed
this reactor eleven, 2.9–m (9.6–ft) sections of 31.8–mm through a series of static mixers. A spray valve then added
(1.25–in.) schedule 40 PVC pipe were assembled into a water to the mixture at a rate of 57 mL/min. The mixture then
vertical
ladder" shaped reactor. Each pipe had a slope of passed through another series of static mixers and was sepa-
25 mm (1 in.) over its 3–m (10–ft) length. Sections were rated in the centrifuge into ester and glycerol phases.
connected with two 90–degree elbows and a short section of A Teel positive displacement pump with a Dart Controls
pipe causing the individual lengths to be separated by Inc. controller was used to meter oil from the supply to the
120.6 mm (4.75 in). At the bottom rung of the network a centrifugal mixing pump. In addition to moving the oil in a
vertical length of 12.6–mm (.5–in.) PVC was used to make positive feed into the mixing pump, the metering pump acted
the outlet of the ladder at the same elevation as the inlet. This as a check to prevent any of the catalyst from flowing back
vertical length of PVC not only helped assure that the into the reservoir of oil. During calibration, the metering
glycerol would not accumulate at the bottom but also ensured pump and controller would not sustain a consistent pumping
that the reactor was filled with reactant. Quick–release rate when the motor driving the pump was operated at speeds
hydraulic couplers were used at the inlet and outlet for less than 20 rpm. Therefore, the motor was geared down
connection into the system. Nine feet of static mixers in using a chain drive to allow the motor to operate at higher
12.6–mm (.5–in.) clear polyvinyl chloride tubing were RPMs while allowing the pump to operate at the desired rate.
integrated into the new system. Separation of the ester from the waste glycerol was
A platform was constructed for the pumping and metering performed by either settling or a centrifuge. The Alfa Laval
portions of the network. The centrifuge was mounted on a centrifuge was incorporated into the purification process. A
movable frame and an explosion proof controller was gravity disk size of 51 mm was used within the centrifuge for
attached. optimum separation. The centrifuge was started 5 min prior
A separate movable self–supporting wall held the to product separation to allow sufficient time for the
retention reactor network. Figures 4 and 5 are photographs of centrifuge to reach full speed and then glycerol was added for
the various components. priming.

CF2 Operation Explosion Proof Considerations

For convenience in these tests canola oil [density = It is necessary to use explosion–proof switches and service
0.915 g/mL (7.61 lb/gal)] was used to produce biodiesel. entrance connections since ethanol and methanol are highly
Canola oil is liquid at room temperature and is pure enough flammable. The centrifugal mixing pump was powered with
to prevent settling of foreign matter within the retention a hydraulic motor, which is a convenient method for
network and connections. Just prior to beginning a run, the providing both speed control and prevention of explosion
catalyst was mixed using a ratio of 11.6 L (3.07 gal) ethanol hazards.
[density = 0.789 g/mL (6.565 lb/gal)] to 2.2 kg (1 lb) KOH.
The process started with oil metered into the inlet of the Calibration Data
centrifugal mixing pump with the positive displacement Calibration was performed on the flow–regulating pump
oil–metering pump at a flow rate of 380 mL/min and a straight line fit to the data provided a model for the

8 APPLIED ENGINEERING IN AGRICULTURE

 Flow Flowrate
Oil Pump Calibration Curve
Gage Reading (mL/min) 1000
18 140 900

Flow–rate
21 300 800

(mL/min)
700
21 300 600
21 330 500
22 350 400
22.5 355 300
200
23 380 100
23 410 0
23 500 15 20 25 30 35
23 500 Dial Position
25 600
Figure 6. Calibration data for the Teel positive displacement oil flow regulating pump.

pump flow adjustments necessary during biodiesel produc- 3. A valve at the exit to the retention reactor network was
tion (fig. 6). Calibration was also performed on the FMI pump turned off and the exit tubing was disconnected and
for modeling the flow of the ethanol/catalyst mixture result- reconnected at the lowest point of the network.
ing in 111 mL/min at the setting used in these tests. 4. A valve at the bottom outlet position of the retention
The final calibration was performed as the biodiesel was reactor network was turned off and the air bleed valve was
produced during the trial run. The water insertion valve was opened allowing the remaining product to drain from the
initially set at 22.7 L/h (6 gal/h), but this flow rate resulted network and through the water injection and centrifuge.
in a foamy product in the exit tubing. Flow was eventually 5. After product stopped flowing from the retention network,
reduced to 9.1 L/h (2.4 gal/h), resulting in a product that was the water and centrifuge were turned off.
similar in appearance to that produced with the batch process. Glycerol will settle from the mixture if left in the system.
An alternative to draining is to fill the system with ester using
Start Up and Shutdown Procedure the oil–metering pump. Filling with ester would take about
One of the problems with a continuous flow system, like an hour and 22.7 L (6 gal) of ester. When the system is
this, is the amount of material required to fill the system. restarted the ester would pass through the centrifuge and be
Approximately 22.7 L (6 gal) of mixture is required to fill the reclaimed.
system. Once the mixture is in the system it is difficult to
empty. The following procedure was used for starting and ANALYTICAL MEASUREMENTS
stopping the production of biodiesel with the continuous flow Free and total glycerol were determined using the
system. Iodometric–Periodic Acid method described in AOCS Ca
Start Up 14–56, Total, Free and combined Glycerol (AOCS, 1989).
1. A brass air bleed valve at the top, entrance end of the Potassium was measured by Atomic Absorption
retention network was opened to allow air to escape while Spectrophotometer by the Analytical Sciences Laboratory,
the system was filling. University of Idaho. Other fuel parameters were measured
2. The flow regulator oil pump and FMI pump were primed according to methods recommended in ASAE Standard
with vegetable oil and ethanol/catalyst, respectively. EP552 (ASAE, 2000) by the analytical laboratory in the
3. The hydraulic driven centrifugal mixing pump was Department of Biological and Agricultural Engineering at
started. the University of Idaho. Only process sensitive parameters
4. The flow regulator oil pump and are reported here. Complete fuel properties have been
ethanol/catalyst–metering pump were turned on at reported in several earlier articles, see for example Peterson
appropriate flow rates. et al. (1997).
5. Filling of the retention network took approximately 1 h.
Once product began to exit the bleed valve, it was closed
and products started to flow out the exit of the reactor RESULTS AND DISCUSSION
network toward the water spray nozzle. The first trials were conducted with the CF–1
6. The water injection and Alfa Laval centrifuge were configuration as previously described. The retention reactor
primed and turned on. was insufficient in size for complete reaction, thus the CF–2
7. Collection of separated glycerol and ester began. configuration was designed, built, and tested. Data and
8. The system flow rates required monitoring to assure a discussion will be provided for both systems.
quality product.
TESTS WITH THE CF–1 CONFIGURATION
Shutdown Tests with the CF–1 configuration produced some
1. The flow regulator oil pump, centrifugal mixing pump, biodiesel; however, free and total glycerol were excessively
and ethanol/catalyst metering pumps were turned off. high. It was found that the centrifuge size limited flow rates
2. A valve at the entrance to the retention reactor network and that the retention reactor was too short. Using a
was turned off. glycerol–settling tank to recover the ester worked better in
this trial. However it required considerably more feedstock

Vol. 18(1): 5–11 9

to begin the process and required more labor for clean up oil source was removed. The tube at the top exit of the net-
when the system was shut down for prolonged periods. work was removed and attached at the lower valve. Once the
lower valve was opened, gravity forced the remaining reac-
TESTS WITH THE CF–2 CONFIGUARATION tion mixture out the tube and on to the rest of the process.
Three tests were performed (called run 1, run 2, and Overall, cleanup took about 1.5 hours.
run 3), each approximately 4 hours long. Run 1 was with a Following each use, the retention network should be
retention reactor with a vertical configuration, Runs 2 and 3 flushed first with water then with ethanol to provide a pure
were with the longer
ladder type" retention reactor as shown test each time. In addition, the centrifuge should be cleaned
in figure 5. Run 3 was essentially a continuation of Run 2 on each time it is used. An emulsive material built up on the
successive days. Run 2 lasted approximately 3 hours on an inner sections, potentially reducing long–term separation
afternoon. To begin Run 2 the system was shut down and effectiveness.
reinitiated the next morning without flushing the system. It
was determined that biodiesel production and quality were Problems
equal to runs performed after total cleaning and flushing. Some problems were noticed during the first run. The
Only slight adjustments were made to water injection chief problem was associated with pressure build up inside
flow–rates due to uneven glycerol settling. Once the the retention network as it was filling with reactants. A brass
ester/glycerol from the previous day’s production had exited, air bleed valve was inserted at the top end of the network. A
all settings remained constant with only slight monitoring. small tube was inserted onto the valve, leading into a beaker.
The network was full when product came out of the tube into
BIODIESEL QUALITY the beaker. The valve was then closed.
Free and total glycerol tests [total, free, and combined A second problem was ethanol vapor. As the centrifuge
glycerol (Iodometric–Periodic Acid Method, AOCS Ca separated the ester from the glycerol, the residual alcohol
14–56)] were performed on the resulting biodiesel produced. vaporized into the laboratory. The alcohol fumes were very
The results for the three test runs are shown in table 2. irritating to those working in the area and potentially
Analysis of the biodiesel for the best run showed it contained hazardous. To eliminate the fumes, the experiment was
0.016% free glycerol and 0.236 total glycerol, both of which moved to a well–ventilated area. Future testing with the
are within ASTM PS121–99 specification for biodiesel. system will require an alcohol vapor capture system.
Potassium analysis provided a means to determine the At a flow rate of 0.38 L/min or 22.8 L/h, the continuous
amount of catalyst remaining in the ester phase. Current flow system would produce as much biodiesel in 41.5 hours
batch methods used in our department, using ethanol, have as one batch from an existing 945 L per batch system.
not been found to always meet ASTM quality standards for Ordinarily, the batch system produces one batch per week
free and total glycerol without additional processing (Lowe (operation is during working hours, but non–working time is
et al., 1998). required for settling) so the small, pilot scale, continuous
Even though the pumps can be set to specific flow rates, flow system would increase present production capacity by
it would improve the operation of the system if flow meters three fold if operated 24 h/day.
and computer controls were added. These improvements
would allow the operator to better monitor the process and to CONTINUOUS FLOW ESTER PLANT CAPITAL COSTS
have the capability of making informed adjustments to The total cost of the continuous flow ester plant as
feedstock flow rates during operation. configured for these tests was determined to be $17,187. As
A second consideration is the amount of ester remaining shown in table 3, the major costs were for the centrifuge
in the glycerol phase. The ester in the glycerol is easily ($14,790), the ethanol and catalyst FMI metering pump
recovered because the ester will be the top layer and can be ($915), the oil flow regulating pump and speed control
recycled through the centrifuge. It was observed that the ester ($604), the centrifugal mixing pump ($300), and the static
in the glycerol phase was very small but no specific mixers ($78). The remainder of the components were
measurements were made. relatively inexpensive and were available locally. No
estimate of the cost of construction and operation of a
Cleanup and Maintenance commercial plant was made.
The cleanup process for the retention network worked
well. The valve at the bottom of the network at the pipe
junction allowed flow to continue from the network after the
Table 3. Estimated capital cost of equipment for constructing
continuous flow system CF–2 as tested.[a]
Table 2. Laboratory analysis of process sensitive parameters for Item Cost ($)
biodiesel produced in continuous flow process configuration CF–2. Centrifuge 14,790
Analysis ASTM Limit Run 1 Run 2 Run 3 FMI ethanol/catalyst metering pump 915
Free glycerol (%) 0.02 0.036 0.021 0.016 Oil flow regulating pump and speed control 604
Total glycerol (%) 0.24 1.49 0.267 0.236 Centrifugal mixing pump 300
Viscosity @ 40_C (mm2/s) 1.9–6.0 5.81 5.9 6.0 Static mixers 78
Potassium (ppm) 170 16 27 Miscellaneous 500
Ethanol (%) 0.12 0.09 0.09 Total 17,187
[a] No labor charges are included.

10 APPLIED ENGINEERING IN AGRICULTURE

CONCLUSIONS REFERENCES
The continuous flow reactor successfully produced ester AOCS. 1989. AOCS Official Method Ca 14–56. Total, free and
from canola oil during one run. The glycerol content of the combined glycerol. Sampling and analysis of commercial fats
ester, as shown by results of one test, was potentially within and oils. Champaign, Ill.: AOCS.
the limits of the ASTM PS121–99 (ASTM, 2001) standard ASAE Standards, 47th Ed. 2000. EP552. Reporting of fuel
properties when testing diesel engines with alternative fuels
specification for biodiesel. It is anticipated that refinements
derived from biological materials. St. Joseph, Mich.: ASAE.
to the system, such as computer controls and optimization of ASTM. 2001. PS121–99. Provisional Specification for biodiesel
feedstock flow rates, could further improve the final product. fuel (B100) blend stock for distillate fuels. West Conshohocken,
Production of biodiesel at a rate three times the existing batch Pa.: ASTM.
system as currently operated would result from 24–hour Bam, N., D. C. Drown, R. Korus, D. S. Hoffman, T. G. Johnson,
operation of this continuous flow system. Capital cost of the and J. M. Washam. 1995. Method for purifying alcohol esters.
system was reasonable with the major expense being the U.S. Patent No. 5424467.
costs of the centrifuge and metering pumps. Bradin, D. S. 1996. Biodiesel fuel. U.S. Patent No. 5578090.
Lowe, G. A., C. L. Peterson, J. C. Thompson, J. S. Taberski,
P. T. Mann, and C. L. Chase. 1998. Producing HySEE Biodiesel
from used french fry oil and ethanol for an over–the–road truck.
RECOMMENDATIONS ASAE Paper No. 98–6081. St. Joseph, Mich.: ASAE.
Additional testing of the continuous flow reactor should Noureddini, H., D. Harkey, and V. Medikondura. 1996. A
include studies of the following: Continuous Process for the Conversion of Vegetable Oils into
S alcohol vapors coming from the centrifuge should be Biodiesel, Liquid Fuels and Industrial Products from Renewable
Resources, 83–94, St. Joseph, Mich.: ASAE.
recovered to reduce cost, improve safety, and reduce the
Peterson, C. L., M. Feldmen, R. Korus, and D. L. Auld. 1991.
negative environmental impact; Batch type transesterification process for winter rape oil. Applied
S a computer monitoring and control system for sensing Engineering in Agriculture 7(6): 711–716.
flow of oil and control of the alcohol/catalyst mixture Peterson, C. L., and D. L. Reece 1996. Emissions characteristics of
injection rate; ethyl and methyl ester of rapeseed oil compared with low sulfur
S further studies on ester flow rate and centrifuge diesel control fuel in a chassis dynamometer test of a pickup
adjustments on ester purification; truck. Transactions of the ASAE 39(3): 805–816.
S a method for recovering the excess ethanol from the ester _____. 1996b. Emissions testing with blends of esters of rapeseed
and glycerol phases and an apparatus to capture the oil fuel with and without a catalytic converter. SAE Technical
ethanol vapor leaving the centrifuge; and, Paper No. 961114. Warrendale, Pa.: SAE.
Peterson, C. L., D. L. Reece, J. C. Thompson, S. M. Beck, and C.
S secondary treatment including evaluation of a second
Chase. 1996. Ethyl ester of rapeseed used as a biodiesel fuel ––
centrifuge step for further purification. A case study. Biomass and Bioenergy 10(5/6): 331–336.
S operation of the continuous flow system over a longer time Peterson, C. L., Casada, M. E., L. M. Safley Jr., and J. D. Broder.
to assess product quality variations over the duration of a 1995. Potential production of agriculturally produced fuels.
long run. Applied Engineering in Agriculture 11(6): 767–772.
S development of a real time assay to monitor the quality of Peterson, C. L., D. L. Reece, B. L. Hammond, J. Thompson, and S.
the product and to provide for real time flow rate M. Beck. 1997. Processing, characterization, and performance of
modifications as required to maintain high quality. eight fuels from lipids. Applied Engineering in Agriculture
13(1): 71–79.
Peterson, C. L., and T. Hustrulid. 1998. Carbon cycle for rapeseed
ACKNOWLEDGEMENTS
oil biodiesel fuels. Biomass and Bioenergy 14(2):91–101.
The authors express appreciation to the Idaho Department Zhang, X, C. Peterson, D. Reece, R. Haws, and G. Moller. 1998.
of Water Resources Energy Division for sponsoring the Biodegradability of biodiesel in the aquatic environment.
development of the continuous flow reactor for biodiesel Transactions of the ASAE 41(5): 1423–1430.
production. Project manager was John Crockett, Bioenergy
Specialist. Advice and assistance was also obtained from
Craig Chase, consultant and Jeffrey James, manager,
Department of Energy, Office of Transportation
Technologies, Pacific Northwest and Alaska Regional
Bioenergy Program.

Vol. 18(1): 5–11 11

12 APPLIED ENGINEERING IN AGRICULTURE