Biodiesel

ChE 422 Design Report

Continuous Production of Biodiesel

Conducted by: Daniel Bieber Stuart Houle Kathlene Jacobson

Date Due: April 7, 2008

Department of Chemical Engineering University of Saskatchewan

Date received

i

Executive Summary

The objective of the proposed design project was to produce biodiesel and make it

economically competitive with conventional diesel by utilizing low quality feedstock’s

and a solid acid catalyst.

A Canadian Climate Change Action Plan was produced in 2000 that will be

implemented in 2010 that states conventional diesel must have a 2% blend of biodiesel.

This plan will create a Canadian demand of 500,000,000 litres per year. SFA will

produce 17,000,000 litres per year of biodiesel by converting triglycerides from waste

cooking oil and greenseed canola in a continuous process to meet Saskatchewan’s

demand.

The plant consists of three main sections; a canola crushing plant, a biodiesel

conversion stage and finally a biodiesel/glycerol separation stage. The canola crushing

stage uses a screw press to extract 75% of the oil from the canola seed and a membrane

separator to extract the remaining 25% maximizing oil recovery and also will produce a

valuable meal. The biodiesel conversion stage uses a solid acid catalyst, zinc ethanoate

supported on silica to achieve 95+ conversion in a fixed bed reactor. Finally gravity

separators will be used to separate biodiesel from glycerol.

A conservative economical analysis found that the capital cost for the entire plant

was $2,800,000. Using a 10% discounted rate, the net present value of the plant was

$5,300,000 after 10 years of operation, giving a discounted cash flow rate of return of

49%. This concludes that the plant could produce biodiesel on a scale to be competitive

with conventional diesel.

ii

Acknowledgments We would like to thank Dr. Gordon Hill for his assistance and guidance with this

project, and his patients during the days we abused his open door policy. We would also

like to thank Dr. Richard Evitts for his assistance in solving any issues we had with our

HYSYS simulation, Dr. Mehdi Nemati for his assistance with heat exchangers and

mixers, and Dr. Ding-Yu Peng for his patients and guidance in our design of the hexane

stripper. A great deal of thanks goes Dr. Dalai for giving us the opportunity to work on

this project and for his additional guidance throughout the year.

Additional acknowledgments;

• Zenneth Faye, CEO of Milligan Biotech, for letting us tour their Biodiesel plant

located on campus.

• Blair Harker, Chemical Engineering Student, for helping us with design of our

canola crushing plant by sharing her past work experience.

iii

Table of Contents Executive Summary ............................................................................................................. i Acknowledgments............................................................................................................... ii Table of Contents ............................................................................................................... iii List of Figures ..................................................................................................................... v List of Tables ..................................................................................................................... vi Nomenclature .................................................................................................................... vii 1.0 Introduction ................................................................................................................... 1 2.0 Qualitative Process Description .................................................................................... 8 

2.1.1 Physical Canola Extraction ................................................................................ 9 2.1.2 Solvent Extraction ............................................................................................ 12 2.1.3 Meal Preparation .............................................................................................. 16 2.2 Overview of Biodiesel Production ...................................................................... 18 2.2.1 Gum Removal from Canola Oil ....................................................................... 18 2.2.2 Mixing of Reactants ......................................................................................... 19 2.2.3 Conversion to Biodiesel via a Fixed Bed Reactor ........................................... 19 2.2.4 Methanol Recovery .......................................................................................... 20 2.2.5 Separation of Reaction Products ...................................................................... 21 

3.0 Simulation Results ...................................................................................................... 22 3.1 Process Simulator and Fluid Package Used ........................................................ 22 3.2 Canola Extraction ................................................................................................ 24 3.3 Meal Preparation ................................................................................................. 28 3.4 Biodiesel Conversion .......................................................................................... 29 3.5 Heat Exchangers and Process Heaters ................................................................ 32 3.6 Simulation Summary .......................................................................................... 34 

4.0 Process Alternatives .................................................................................................... 35 4.1 Canola Extractor Using Only the Screw Press ................................................... 35 4.2 Evaporator for Hexane Separation ...................................................................... 36 4.3 Production of Biodiesel via Homogeneous Catalyst and Batch Reactor ............ 38 

5.0 Safety Analysis ........................................................................................................... 41 6.0 Equipment Sizing and Costs ....................................................................................... 43 

6.1 Roll Mill .............................................................................................................. 43 6.2 Rotary Steam Cooker E-111 ............................................................................... 44 6.3 Screw Press C-110 .............................................................................................. 44 6.4 Leacher H-120 .................................................................................................... 44 6.5 Membrane Pump K-122 ...................................................................................... 45 6.6 Membrane Separators H-140 .............................................................................. 45 6.7 Hexane Evaporator R-150 ................................................................................... 45 6.8 Hexane Stripper D-160 ....................................................................................... 46 6.9 Hexane Flash Drum H-180 ................................................................................. 46 6.10 Desolventer/Heater H-130 & E-132 ................................................................. 46 6.11 Heat Exchangers ............................................................................................... 47 6.12 Mixer Design .................................................................................................... 48 6.13 Fixed Bed Reactor............................................................................................. 49 6.14 Flash Drum w/ Mist Eliminator ........................................................................ 49 6.15 Gravity Separators ............................................................................................ 50 

iv

7.0 Economics ................................................................................................................... 51 7.1 Alternative 1: Negating Glycerol Sales .............................................................. 53 7.2 Alternative 2: Using Grade 1 Canola as the Feedstock ...................................... 54 7.3 Alternative 3: Using an Evaporator for Hexane Extraction ................................ 56 7.4 Alternative 4: Selling Glycerol at Current Prices ............................................... 57 7.5 Alternative 5: Selling Glycerol at Current Prices using an Evaporator .............. 58 7.6 Alternative 6: Producing Value Added Products from Glycerol ........................ 59 

8.0 Conclusions ................................................................................................................. 60 9.0 Recommendations ....................................................................................................... 62 10.0 References ................................................................................................................. 63 Appendix A ....................................................................................................................... 65 

A.1. Sample Calculation for Sizing and Cost of Area 100 ....................................... 66 A.2. Sample Calculation for Sizing and Cost of Area’s 200 and 300 ...................... 78 

Appendix B ....................................................................................................................... 82 Appendix C ....................................................................................................................... 98 Appendix D ..................................................................................................................... 138 Appendix E ..................................................................................................................... 139 

v

List of Figures Figure 1.1: Tailpipe Emissions of Biodiesel Relative to Conventional Diesel Figure 1.2: Transesterification Reaction in the Presence of a Base Catalyst Figure 1.3: Esterification Reaction in the Presence of an Acid Catalyst Figure 1.4: Mechanism for the Simultaneous Transesterification and Esterification of FFA and TGs Figure 2.1.1i: Rotary Steam Cooker to Evaporate Moisture for the Canola Seed Figure 2.1.1ii: Schematic of o Typical Screw Press used for Physical Extraction of Canola Seed Figure 2.1.2i: Rotocel Type Extractor for Hexane Leaching of Cake Figure 2.1.2ii: Nanofiltration Membrane Separator Schematic to Recover Hexane Figure 2.1.2iii: Falling Film Evaporator for Hexane Recovery Figure 2.1.3: Marc Desolventer/Toaster Unit to Produce Meal Figure 2.2: PFD for Area 200 – Conversion to Biodiesel Figure 3.2.1: HYSYS PFD of Canola Extraction Facility Figure 3.4.1: HYSYSTM for the Biodiesel Conversion Process Figure 4.2.1: Canola Extraction Facility Using Evaporator Separator (H-150a) Figure 4.3.1: Basic Schematic of Convention Biodiesel Production Utilizing Homogeneous Catalyst in a Batch Reaction Figure 7.1: Cash Flow Analysis Figure 7.1.1: Cash Flow Analysis with Negating Glycerol Figure 7.2.1: Alberta’s Grade 1 Canola Prices Figure 7.2.2: Cash Flow Analysis Using Grade 1 Canola Figure 7.3.1: Cash Flow Analysis Utilizing Our Design’s Prices, But Using an Evaporator Instead of a Membrane Separator Figure 7.4.1: Cash Flow Analysis Using Current Prices and a Membrane Separator Figure 7.5.1: Cash Flow Analysis Using Current Prices and an Evaporator

vi

List of Tables Table 3.2.1: Important Parameters for Physical Extraction of Canola Oil Table 3.2.2: Important Stream Parameters for Chemical Canola Oil Extraction Table 3.4.1: Important Stream Parameters before Biodiesel Conversion Table 3.4.2: Important Stream Parameters after Biodiesel Conversion Table 3.5.1: Parameters of Heat Exchangers Used in Process as Modeled by HYSYSTM Table 3.5.2: Parameters for Process Heaters in Biodiesel Plant Table 3.5.3: Energy Required for the Process as Calculated by HYSYSTM and by the Short-cut Method. Table 4.1.1: Energy Savings Associated with Screw Press Only Operation Table 4.1.2: Comparison of Physical Extraction versus Combined Extraction Table 4.2.1: Cost Comparison between Evaporator and Membrane Separator Table 4.2.2: Energy Comparison between Membrane and Evaporator Separator Table 5.0: Chemical Hazard Information Summary Table 6.11.1: Heat Exchanger Size and Cost for Area 100. Table 6.11.2: Area 100 Unit Sizing and Cost Table 6.14: Sizing and Cost Summary of Fixed Bed Reactor and Flash Drum Table 7.5.1: Summary of Economic Alternatives

vii

Nomenclature A-Area (m2)

ABD – Annual depreciation

ADME – Direct manufacturing expenses

AGE – General expenses

AI - Investment

AIME –Indirect manufacturing expenses

AIT – Income taxes

AME – Manufacturing expenses

ANCI – Net cash income

ANNP – Net annual profit after taxes

ANP – Net annual profit

AS- Sales

ASTM-American Society for Testing and Materials

ATE- Total expenses

BI – Business interruption loss

Cbm – Bare module cost

CFC – Fixed capital

CTC- Total capital investment

CWC – Working capital

D,d-Diameter (m)

DBEP - Discounted breakeven point

DCFRR - Discounted cash flow rate of return

es-Stage efficiency

fD - Discounted factor

F & EI – Dire and explosion index

FFA - Free Fatty Acids

Fsolid-Mass Flowrate of Solids in Leacher (kg/s)

Fsolvent Mass Flowrate of Solvent in Leacher (kg/s)

H-Specific Enthalpy (kJ/kg)

H-Height (m)

viii

Le-Length (m)

m=Mass flowrate (kg/s)

mbiodiesel-Mss flowrate of biodiesel (kg/s)

mglycerol-Mass flowrate of glycerol (kg/s)

nov-Number of stages

NPV - Net Present Value

Nseparators –Number of Separators

PFD-Process Flow Diagram

ppm-parts per million

P-Power (kW)

Q-Volumetric Flowrate (m3/s)

T-Temperature (°C)

TG - triglycerides

U-Heat Transfer Coefficient (kW/m2K)

u-Velcoity (m/s)

V-Volume (m3)

Xn-Mass/mole ratio in nth stage of underflow/liquid steam

xn-Mole fraction in liquid stream in stripper

Yn-Mass/mole ratio in nth stage of overflow/gas stream

yn-Mole fraction in gas stream in stripper

z-Packing Height (m)

Greek Symbols rbiodiesel-Density of biodiesel (kg/m3)

rbulk-Bulk density (kg/m3)

rcatalyst-Density of catalyst (kg/m3)

rg-Density of Gas

(kg/m3)

rglycerol-Density of glycerol (kg/m3)

rl-Density of liquid (kg/m3)

1

1.0 Introduction Sustainable fuel alternatives have recently become a high priority for many

countries and will play a large role in the chemical industry in the near future. Biodiesel

is an alternative fuel for the following reasons; it is biodegradable, non-toxic, has low

emission profiles and is environmentally beneficial (Krawczyk, 1996). Biodiesel fuel has

the potential to reduce the level of pollutants and the level of potential or probable

carcinogens (Krawczyk, 1996). Furthermore, biodiesel is an alternative fuel that can be

used in unmodified engines with the current fuelling infrastructure (US Dep. Of Energy,

2001) and it is made from renewable biological sources such as vegetable oils and animal

fats.

With recent technology, diesel engines have become more powerful and 30-35%

more efficient than a gasoline engine because of a higher compression ratio. Through the

higher compression ratio, diesel engines operate at a higher efficiency, which increases

fuel economy by 40%. In addition, diesel fuel contains approximately 15% more energy

per unit volume compared to gasoline. Diesel engines are considered cleaner because

they burn all excess fuel in its tank enabling it to reduce carbon monoxide and black soot

emissions. Furthermore, low-sulfur diesel has been created to reduce sulfur emissions.

However, if biodiesel replaces or is blended with diesel, sulfur emissions and soot will be

further reduced.

In the 1990 Clean Air Act Amendments, biodiesel was the only alternative fuel to

pass the health effects test. In Figure 1.1 below, it can be seen that a twenty percent

blend of biodiesel will reduce the particulate matter emissions by 10%, and the

hydrocarbon emissions by 20% when compared with conventional diesel. Also, figure

1.1 indicates nitrous oxide emissions are increased in comparison to petroleum diesel

2

because it burns at a higher temperature. However, fuel additives or a catalytic converter

can be used to reduce nitrous oxide emissions.

Figure 1.1: Tailpipe emissions of Biodiesel relative to Conventional Diesel http://www.ucsusa.org/assets/images/trucks_and_buses/biodiesel_graph.gif

Biodiesel is the only alternative fuel to possess an overall positive life cycle

assessment. NRCan analyzed the life cycle of biodiesel on a based on GHG emissions.

Their analysis found biodiesel produces 64-92% fewer emissions compared with

petroleum diesel. Furthermore, while analyzing the fossil fuel consumption, biodiesel

yields as much as 3.2 units of fuel product energy for every unit of fossil energy

consumed in its life cycle, compared to only 0.83 units for petroleum diesel (Kiss et al.,

2005).

In accordance with the Kyoto Accord, Canada must reduce their CO2 emissions to

10% below the 1990 levels. Biodiesel will help achieve this goal as there are no net CO2

emissions as biodiesel is produced from vegetables oils made from the consumption of

CO2 through photosynthesis. Furthermore, the Canadian government has legislated R.F.

3

standards which will implement a 2% biodiesel content in diesel by 2012, resulting in a

Canadian biodiesel demand of 500 million liters.

Biodiesel has not become a vastly popular alternative fuel worldwide due to its

higher cost when compared with traditional petroleum diesel. Currently, the National

Biodiesel Board states that a 20% blend of biodiesel costs $0.20 more per gallon than

pure petroleum diesel. The two major hurdles in the commercialization of biodiesel are

its large production costs and raw material costs.

The production of the alternative fuel consists of alkyl esters derived from either

the transesterification of triglycerides (TGs) or the esterification of free fatty acids

(FFAs) with short-chained alcohols. The chemical reactions for both transesterification

and esterification as shown below.

Figure 1.2: Transesterification reaction in the presence of a base catalyst

Figure 1.3: Esterification reaction in the presence of an acid catalyst

Triglycerides are most commonly available as straight vegetable oils (canola oil)

and free fatty acids are available as a low quality oil or waste cooking oil (WCO). The

use of either can be used to make biodiesel, and both will be discussed as feedstocks in

C H 3 O C O R 1 C H O C O R 2 C H 2 O C O R 3

+ 3 C H 3 O H CH2OHCHOHCH2OH

CH3O C O R 1 CH3O C O R 2 CH3O C O R 3

T r i g l y c e r i d e M e t h a n o l Glycerol Methyl es t e r s ( b i o d i e sel)

+

RCOOH CH3OCOR + CH3OH + H2OFFA Methanol Methyl esters (biodiesel)

4

the following report. WCO is a preferable feedstock over refined vegetable oil because it

is a cheap, low quality oil and will aid in achieving economical feasibility in biodiesel

production. The amount of WCO generated in each country varies depending on the use

of vegetable oil. An estimate of the potential amount of WCO from the collection in the

European Union (EU) is approximately 0.7 to 1.0 Mtonnes per year. The United States

and Canada produce on average, 9 and 8 pounds of yellow grease respectively, per person

(Kulkarni, 2006). Currently, WCO is collected from households and restaurants are

disposed of as either animal feed or an environmental pollutant. Thus, WCO offers

significant potential as an alternative low–cost biodiesel feedstock which could partly

decrease the dependency on petroleum-based fuel.

The production of biodiesel from WCO is challenging due to the presence of

undesirable components such as free fatty acids (FFAs) and water. Usage of

homogeneous alkali catalyst for transesterification of such feedstock suffers from serious

limitation of formation of undesirable side reactions. One such reaction is saponification,

which creates the serious problem of difficult product separation and ultimately lowers

the ester yield. Homogeneous acid catalysts have the potential to replace alkali catalysts

since they do not show measurable susceptibility to FFAs and can catalyze esterification

and transesterification simultaneously. However, slow reaction rate, requirement of high

temperature, high molar ratio of oil and alcohol, separation of the catalyst, serious

environmental and corrosion related problems make their use non-practical for biodiesel

production (Lotero et al, 2005). Currently a dual step process has been used for biodiesel

preparation from high FFA containing WCO. The first step of the process is to reduce

FFA content in the oil by esterification with methanol to methyl esters catalyzed by an

acid (generally sulfuric acid) followed by transesterification process, in which

5

triglyceride (TG) portion of the oil reacts with methanol and base catalyst (usually

sodium or potassium hydroxide) to form ester and glycerol. The current process is not

economical as it involves a number of steps including washing of the esters to remove

acid/alkali catalysts in addition to creating contaminated water disposal issues.

Solid acid catalysts have the strong potential to replace liquid acids, eliminating

separation, corrosion and environmental problems. Through a summer research project,

which this design project was spawned, it was discovered that solid acid catalyst was

capable of producing biodiesel to meet ASTM standards in a small scale batch reactor.

More specifically, research found Zinc Ethanoate supported on silica is capable of

simultaneous transeterification and esterification to produce quality biodiesel (mechanism

can be seen in Figure 4). To our knowledge, there are no reports on the utilization of solid

acid catalysts for the production of biodiesel from WCO in a single step. Therefore, in an

attempt to develop a robust solid acid catalyst that can simultaneously catalyze

esterification as well as transesterification reaction, different types of solid acid catalysts

are synthesized and evaluated for biodiesel preparation from WCO. Also, influence of

various reaction parameters such as molar ratio of WCO to alcohol and catalyst loading

was studied in the present investigation.

6

Figure 1.4: Mechanism for the simultaneous transesterification and esterification of FFA and TGs

The use of cheap low quality feed stocks such as waste cooking oil (WCO) in

combination with green seed canola oil (canola that has not matured due to frost) will

help in improving the economical feasibility of biodiesel. In Saskatchewan, the

availability of these low-cost raw materials is vast. Based upon 8lb/person of WCO

(Kulkarni, 2006) in Canada and the Saskatchewan population, it is possible to maximize

the utilization of WCO and make up the remaining feedstock requirements with green

seed canola. Saskatchewan produces 150 tonnes/year of WCO and 30,000 tonnes/year of

greenseed canola to process 15,000 tonnes/year of biodiesel.

Furthermore, whenever biodiesel is produced, glycerol is produced as a

byproduct. The process detailed above produces 4,750 tonnes of crude glycerol every

year. It should be noted that this process involves a separable solid catalyst and removal

of methanol by flash distillation creating a higher quality crude glycerol. A higher quality

of glycerol is produced because it is not contaminated with catalyst, there will be minimal

L +

R C

O

O HR ' C

O

O R "

L

CR O H+

C H 3 O H

L

O

CR O H

H 3 C O H+

O

L

O

CR O H 2

H 3 C O

+

L

O

C +R

H 3 C O

-H 2 O

R C

O

O C H 3

R ' C

O

L

+O R "

C H 3 O H

L

O

CR ' O R "

H 3 C O H+

L

O

CR ' O H R "

H 3 C O

+

L

O

C +R '

H 3 C O

- R "O H

R ' C

O

O C H 3

E S T E R IF IC A T IO N T R A N S E S T R IF IC A T IO N

E S T E R S

L + = a c id s ite o n c a ta ly s t su r fac eR = a lk y l g ro u p o f a c idR ' = c a rb o n c h a in o f fa tty a c idR "= g ly c e r id e

7

water content (no washing of finished product to remove catalyst), and minimal methanol

content (methanol recovery maximized by distillation of entire product mixture before

separation).

There is much debate over the value and demand of glycerol. However, the

demand for glycerol is high in the United States, China, and many developing countries.

Glycerol has thousands of uses in many industries, such as food, medical, lubrication, and

tobacco industries.

The use of solid acid catalysts benefits the economics of the process further by

reducing the amount of process steps, eliminating water use and toxic waste streams. This

project analyzed the design and economics of continuous biodiesel production utilizing a

solid acid catalyst with the above mentioned feedstock and incorporates innovative

methods of an onsite canola crusher.

8

2.0 Qualitative Process Description Description of converting raw canola seed and pre-treated waste cooking oil into

biodiesel, while creating valuable side products such as meal and glycerol.

2.1Canola Extraction Overview

Since the primary method to obtain triglycerides for biodiesel comes from

Saskatchewan Canola seeds, a process in order to extract this oil needs to be developed.

There are many approaches in extracting the oil out of canola seed. Two examples

include physical extraction to obtain the oil, and other uses a combination of physical and

solvent extraction. Most biodiesel producers use the physical extraction method as it is

inexpensive and easy to operate (Faye 2008). However, the physical extraction method

has a disadvantage as it leaves a large portion of oil in the seed, therefore increasing

operating expenses as more seeds need to be purchased.

Another alternative to acquire canola is to purchase the refined canola from a

crushing facility already in operation. However, purchasing refined canola oil is

prohibitively expensive as that oil can be used for human consumption which makes it of

high value (Faye).

The Salamander Fuels Association (SFA) plant will use a combination of physical

and solvent extraction using hexane. SFA intends to design the facility to maximize the

recovery of oil in the range of 95%-99% (O’Brien, 2000). This will minimize our

consumption of canola seed and our operating cost because the extraction scheme

maximize the oil recovery for the feed.

Through the use of physical and chemical extraction, SFA gains another benefit

by producing a high protein by-product called meal. After the oil has been extracted from

9

the seed, the remaining material can be sold as a high value meal to livestock owners

(O’Brien 2000). The livestock owners demand a feed that is low in fats because any fat

present in the meal will lower the daily intake of solid feed matter by the livestock as

explained by Litherland (2005). According to Racz, as long as the fat/oil content in the

meal is lower than 6% by weight there should be no decrease in solid material

consumption for the livestock (2004).

Given this constraint of low oil content animal meal, SFA must meet these

requirements through physical and solvent extraction. According to O’Brien, this can be

achieved through chemical and physical extraction as there is “<1% residual oil in

extracted material” which meets the criteria set above. Even though chemical extraction

is expensive, there is an added bonus of a value-added meal. In fact, according to O’Brien

“[approximately] 70% of the returns in processing soybeans is due to the sale of meal.”

(2000). This is an added benefit to our extraction and biodiesel process as it will give

SFA an extra cash flow to help offset the high processing costs of biodiesel

(Zhang et al. 2003).

2.1.1 Physical Canola Extraction

The first step in the canola extraction process is to remove the hull or the seed

coat from the canola seed. Removal of the hull is beneficial to the overall process as it

exposes the oil from the seed and minimizes the material in the process, thus reducing

energy requirements. Also, removing the hull increases the protein content of the meal

which is a desirable trait for livestock feed (O’Brien 2000). Roller mills crack the hull

and expose the innards of the seed into a fine cake to make processing easier (O’Brien

2000). Oil is lost from cracking the hull during the milling stage, which is the only main

10

drawback to de-hulling. This oil loss is minimal in comparison to what is gained from the

revenue of high protein meal and energy savings.

The flaked seed is then sent to a rotary steam cooker to reduce the water content

of the seeds down to 9-10% by weight (O’Brien 2000). The water content has to be

brought to this low level because the hexane used in the extraction and the water in the

seed are immiscible. Therefore, a seed with a high water concentration prevents the

penetration of hexane into the seed (Perry’s, 1999). This means the effectiveness of the

oil recovery will be reduced if the moisture in the seed is not removed. Therefore, a

rotary steam cooker, as shown in Figure 2.1.1i, will be employed to cook the seeds, and

evaporate the water until the seeds moisture content is at the desired level.

According the Perry’s, this type of cooker is “the most common type of indirect

heat rotary dryer” (2008) used in industry and thus SFA decided to use this equipment

because of the industry confidence in the said equipment.

Figure 2.1.1i: Rotary Steam Cooker to Evaporate Moisture for the Canola Seed. Taken from Perry’s Handbook (2008), Pg. 12-78, Fig 12-60

11

Next, the seeds are sent to a screw press where 75% of the extracted oil is

recovered by physical extraction (O’Brien 2000). Figure 2.1.1ii shows the inner workings

of a screw press that is utilized in SFA’s canola extraction facility.

A screw press is essentially a continuous screw auger that applies pressure on the

canola seed as it is conveyed through the screw. The combination of the barrel and the

screw form a pocket that is ever decreasing down the length of the screw. The decreasing

size of an individual pocket and the friction caused by the moving screw increases the

pressure on the canola seed. This increased pressure causes the oil in the flaked seed to be

separated from the canola cake. The separated oil is then collected at the bottom of the

screw as shown in figure 2.1.1ii, on the next page. The remaining cake travels down the

screw press and collects at the nozzle on the same plane of the screw. The cake is then

sent to solvent extraction to recover 95-99% of the in-situ oil in the seed (O’Brien 2000).

Figure 2.1.1ii: Schematic of a Typical Screw Press used for Physical Extraction of Canola Seed

Obtained from http://www.ag.ndsu.nodak.edu/abeng/alternativefuels/images/screw

12

2.1.2 Solvent Extraction After the physical extraction, the canola cake is then sent to the solvent extraction

train which uses hexane to leach the oil out of the cake. Hexane was chosen as the desired

solvent because it has “high oil solubility, ease of evaporation, abundant availability, and

historically low price” (O’Brien 1999). Hexane has one major drawback as it can be

harmful to worker safety. According to the MSDS for hexane, health effects may include:

impaired fertility, harmful through inhalation, irritant, CNS depression, and serious health

damage. (MSDS 2005).

Solvent extraction begins with canola cake being sent to a rotocel type extractor

as shown in Figure 2.1.2i below. In this extractor, the canola meal is feed in a counter-

current manner to the hexane solvent to maximize the mass transfer of oil into the hexane

solvent. Maximum mass transfer is achieved in counter-current operation because the

greatest oil concentration gradient in each stage exists when the solvent and the cake are

feed in opposite directions from one another.

Figure 2.1.2i: Rotocel Type Extractor for Hexane Leaching of Cake Taken from Perry’s Handbook (2008), Pg. 18-62, Fig 18-86

13

In the rotocel, solvent is fed into isolated compartments which contain a fixed

amount of solid cake. Mass-transfer of solute to solvent is achieved by percolating the

solvent through the solid in an open screened bottom compartment. After percolating

through the solid, the solute rich solvent runs through a mesh screen into a holding tank

before being pumped into the next stage. Each compartment is rotated around a pivot

point. The solids are subjected at each stage to the solvent spray collected from the

bottoms of the following stage. After the cake has been subjected to the specified number

of stages of solvent washings, the floor of the compartment opens up and the solids are

collected and sent for further processing. The cake that was leached is now referred to as

marc which consists of 30% hexane and 70% meal (O’Brien 2000). Overflow from the

leaching process called micella is loaded with 25% oil and 75% hexane (O’Brien 2000).

After the leaching unit, the oil in the micella needs to be separated from the

hexane and recombined with the oil recovered in physical extraction. The hexane

concentration needs to be reduced down to 200 ppm (O’Brien 2000). The low

concentration can be achieved through separation using either a membrane separator, an

evaporator, or a hexane steam stripper.

Next, the micella from the leacher is sent to a nanofiltration membrane for the

first stage of hexane recovery. A schematic of a nanofiltration membrane separator is

shown in figure 2.1.2ii below, which gives the configurations for permeate and

concentrate (retentate) flow through the membrane system.

14

Figure 2.1.2ii: Nanofiltration Membrane Separator Schematic to Recover Hexane

Taken from Perry’s Handbook (1997), Pg. 22-41, Fig 22-53

Nanofiltration was selected based on the research of Ribeiro et al. They did a study using

different types of membrane separators on typical micella compositions (1:3 w/w) in

order to recover hexane from the micella (2006).

The goal of the membrane separator is to lower the energy requirements of

hexane separation. High energy requirements are of major concern for the canola

crushing industry which currently uses evaporation separation techniques (Ribeiro 2006).

Membrane separation is possible because of the large molecular weight difference

between triglycerides (canola oil) and hexane, 900 g/mol and 86 g/mol respectively.

Through this size difference, a reduction in the pore size will enhance the relative

diffusivity by allowing only the hexane to pass through the membrane (Ribeiro et al.

2006). In the research of polyamide/polysulfone SEPA GH nanofiltration membranes,

Ribeiro et al. found optimum conditions of the membrane separator at 15MPa and 59°C

with a permeate oil concentration of 8.0% wt and retentate oil concentration of 67% wt

15

(2006). This gives comparable results to the evaporator case that concentrates the oil to

60% wt (O’Brien 2000). Given the similar performance between the membrane separator

and the evaporator, the membrane separator was chosen for our design because of the

added energy savings. The permeate hexane is recycled to the leaching unit and the

retentate micella is sent for second stage separation.

The second stage of separation uses steam evaporation to recover more of the

hexane from the micella. Using a long vertical tube falling film evaporator as shown in

figure 2.2.2iii on the next page, the oil in the micella will be concentrated to 95% wt.

The evaporator consists of a vertical single-pass shell and tube heat exchanger

discharging to a very small vapour head. Micella enters into the bottom of the heat

exchanger and comes in contact with countercurrent saturated steam. As the steam

condenses, the hexane is vaporized from the micella and the oil stays in liquid form. The

two phase mixture is sent to a tank where hexane vapour is recovered off the top of the

tank and the liquid oil falls to the bottom of the tank where it is drawn off. The

advantages of this style of evaporator are; its low cost, large heating surface, small floor

space and low liquid hold-up (Perry’s 2008). Thus it is an ideal choice for our process.

The recovered hexane from the evaporator is recycled to the leaching unit, and micella is

sent to the third stage of separation.

16

Figure 2.1.2iii: Falling Film Evaporator for Hexane Recovery Obtained from: < http://www.praj.net>

The third and final stage of hexane recovery is the counter current hexane stripper

which uses high temperature steam at 130°C in contact with the oil to remove the hexane.

After the stripper column, the oil contains less than 200 ppm of hexane and is recombined

with the physically extracted oil. This complete process recovers 99% of the oil present

in the seeds.

2.1.3 Meal Preparation After the oil has been extracted from the seed in the leaching unit, the marc from

the underflow is now a high protein meal with approximately 30% hexane. The marc

contains <1% of the residual oil (O’Brien 2000) initially in the seed. This meal is ideal

for livestock feed as it contains 44% protein and approximately 1% fat (Racz 2004).

Since >97% of the canola meal is used for animal feed, the hexane must be

recovered from the marc and the meal must be subjected heat treatment. The heat

treatment is to inhibit enzymes which allows the livestock to consume more meal thus

increasing their body mass (O’Brien 2000). These two criteria are achieved by using a

17

desolventer and toaster (DT) unit. The DT unit consists of hollow stay bolt tray uppers

and solid dryer trays to remove the hexane from the meal. Figure 2.1.3 shown below

displays a schematic diagram of a typical DT unit that would be used in industry and for

our design.

Figure 2.1.3: Marc Desolventer/Toaster Unit to Produce Meal Taken from :< http://www.fao.org/docrep/t0532e/t0532e25.gif>

The upper section of the DT unit uses steam that travels upward through the

hollow trays while the marc travels downward. The rising steam continually evaporates

the hexane out the meal and recycles it back to the leaching unit. Low pressure steam is

used for this process of evaporating hexane.

The lower portion of the DT unit is used to toast the meal by blowing hot air at

100°C over the meal in the solid trays. The toasting denatures protease inhibitors (trypsin

and chyrmotrypsin inhibitors) and the enzyme urease. This gives the meal a longer shelf

life and improves the digestibility of the meal for the cattle (Racz 2004). The meal is then

ready to be sold to Canadian and American markets.

18

2.2 Overview of Biodiesel Production The canola oil obtained from the canola extraction process previously described is

combined with citric acid and allowed to settle in gravity separators to remove the gums.

From here, it is then combined with 2 million L/year of pre-treated waste cooking oil and

12million L/year of excess methanol. This mixture will be pumped to a pressure of 4240

kPa (g) via reciprocal pump and heated to 200ºC via two heat exchangers and a fired

heater. The high pressure, high temperature feed is then sent to a fixed bed conversion

reactor that is filled to 50% volume with Zinc Ethanoate/Silica. Within the reactor,

simultaneous esterification and transesterification of the TG’s (canola oil) and FFA’s

(waste cooking oil) occur to achieve a minimum conversion of 95% to fatty acid methyl

esters (biodiesel) and glycerol. The effluent mixture is then depressurized to atmospheric

pressure while maintaining the high temperature. This enables the use of a flash drum

with a mist eliminator to capture the methanol as vapor from the product mixture and

recycle it leaving a mixture of desirable products: biodiesel and glycerol. The separated

methanol is recycled back to the original mixers to minimize methanol consumption. The

product mixture is then fed into three gravity settlers to allow for glycerol to be separated

from biodiesel by utilizing their density differences.

2.2.1 Gum Removal from Canola Oil Unit L-163 is three mixing tanks that are used to extract the canola gums from the

canola oil. Once the mixing tank is full a 50% citric acid/water mixture will be added

until the citric acid solution is 2500ppm. Once the mixture is added the solution will be

mixed for 10 minutes. The citric acid is used to flocculate the gums together. Water will

be added until the mixture reaches 2% water by volume, and then it will be mixed for

another 20 minutes. A gum will appear at the top of the tank which will then be sucked

19

off using a wet vacuum. The gum will then be re-distributed to the canola meal for added

protein. Oil will be drained from the bottom of the tank through a strainer to purify the oil

and before being sent to the reactor.

2.2.2 Mixing of Reactants The treated canola oil from SFA’s on-site canola crushing plant is fed to mixer L-

210 to combine the canola oil, pretreated waste cooking oil and methanol. It should be

noted that the waste cooking is treated at both the Regina and Saskatoon processing

plants prior to reception. WCO will be trucked from both processing plants to SFA and

pumped into storage tanks which provide enough oil for two weeks of operation. The

waste cooking oil is filtered to remove particle matter and treated to remove the water

created during the frying process. Furthermore, the methanol fed into the second mixer is

a combination of fresh methanol and make-up methanol recovered from the flash drum to

be discussed in section 2.2.3.

2.2.3 Conversion to Biodiesel via a Fixed Bed Reactor After the reaction mixture is obtained from the second mixing tank, the mixture is

pressurized to 4240 kPa using a reciprocal pump and heated to 200 °C using two heat

exchangers and a fired heater as shown in figure 2.2 on the next page. The high

temperature, high pressure mixture is fed into a fixed bed reactor (P-220) to convert TG’s

and FFA’s into methyl esters (biodiesel) and glycerol. The reaction that facilitates the

conversion is simultaneous transesterification and esterification (mechanism can be

viewed in Figure 1.4). The packing of the fixed bed is Zinc Ethanoate supported on

Silica, which acts as a strong acid site for both reactions. In preliminary laboratory

research, it was observed that this particular catalyst was capable of 95+% conversion to

20

methyl esters in the presence of a 3 wt% catalyst and a 1:18 oil to alcohol molar ratio.

Furthermore, the catalyst was found to be stable and reusable.

Two reactors will be employed by SFA. Catalyst recharging will be performed

once a week by switching the continuous flow to the first reactor to the second reactor.

The first reactor will be drained and recharged using methanol and hexane. The methanol

will remove the polar constituents from the catalyst, where the hexane will remove the

non-polar constituents. During this time, an automated system will allow for continuous

production to continue in the second reactor.

Figure 2.2: PFD for Area 200 – Conversion to Biodiesel

2.2.4 Methanol Recovery The product mixture obtained from the fixed bed reactor is fed to a flash drum, H-

230, where the pressure will be dropped to atmospheric and the temperature will remain

at 200 °C. The flash drum will vaporize the methanol from the biodiesel in the mixture.

This recovered methanol will be directed to mixer L-223 (figure 2.2) to combine with

fresh methanol to act as ‘make-up’ methanol. The hot methanol is cooled in a heat

exchanger that is used to partially heat the reaction mixture before it reaches the fixed

21

bed reactor (E-213B). The methanol recovery stream is further cooled in another shell

and tube heat exchanger using cooling water (E-232). Heat exchanger (213A) removes

heat from the biodiesel/glycerol mixture and adds the energy to the reactant stream before

the reactor. Finally, the biodiesel/glycerol mixture is sent to gravity settlers for

separation.

2.2.5 Separation of Reaction Products The treated product mixture of biodiesel and glycerol are fed into gravity settlers

after being cooled in a heat exchanger in contact with the reaction mixture. Unit H-300

are three gravity settling tanks which will be utilized to ensure continuous production.

Once the tank is full it will be left alone for a minimum of 4 hours for sufficient

separation to occur. Since glycerol has a higher density of 1215kg/m3 and biodiesel has a

lighter density of 880kg/m3, the glycerol will all settle to the bottom and the biodiesel

will float to the top. Once separation occurs the tank will be drained off the bottom

through a density control valve into a glycerol basin. After the glycerol is fully removed

and the solution turns to pure biodiesel the valve shuts. Another hose will be attached to

the bottom of the tank and the biodiesel will be pumped into its own separate basin.

22

3.0 Simulation Results

3.1 Process Simulator and Fluid Package Used HYSYSTM 2006 was the simulator chosen to model the biodiesel process. This

process simulator was chosen because of its capability to incorporate thermodynamic

models to predict the energy requirements of the individual units. This allowed us to

analyze the energy requirements/surpluses of the process and to optimize the system

accordingly. A particular advantage of the energy parameters within the HYSYSTM model

allowed us to design heat exchange systems to maximize the utilization of sensible heat

within the process.

The fluid package used in this simulation was the Non-Random Two Liquid

(NRTL) thermodynamic modeling system. The NRTL thermodynamic model was chosen

as it was the same model used by Zheng et al in 2003 for modeling an alkali catalyst for

biodiesel production.

The NTRL fluid package had adequate modeling parameters for each of the

components. This NRTL method could define most components from the HYSYSTM

component library including: water, hexane, citric acid, glycerol, air, and methanol.

However, biodiesel, which is mainly composed of methyl esters, and canola meal could

not be defined from the NRTL component library and had to be defined from ‘the Hypo

Manager” tool in HYSYSTM. Thus, the simulation had to make some assumptions

regarding these two components. However, the simulation can still be trusted to give

accurate results based as Zheng’s publication and his biodiesel process results.

With the thermodynamic package defined, this model should give a reasonably

accurate representation of how an actual plant would behave. The biggest problem

23

associated with using HYSYSTM as a process simulator, is that it does not take into

account pipeline pressure losses, as there are no pipeline connections for the simulator.

Therefore, before an actual facility is completed, a pressure loss survey should be

conducted to analyze the pipeline friction losses throughout the plant. From this study it

may be determined that the pumping requirements are drastically different than what is

present in the HYSYSTM simulation.

24

3.2 Canola Extraction The Process Flow Diagram (PFD) of the Canola Extraction process and meal

preparation process is shown in figure 3.2.1 below.

Figure 3.2.1. HYSYS PFD of Canola Extraction Facility. Due to the limited unit operations in HYSYS, most separators where modeled as splitters. Material streams are shown in blue and energy streams are in red. Tables 3.2.1 and 3.2.2 give parameters for the material streams

For SFAs’ design, both physical and solvent extraction were used to maximize

the recovery of canola oil from the canola seed.

In physical extraction, the seed is first run through rolls mills to crush the seed

into a fine powder to allow for easier extraction. This process was not modeled in

25

HYSYSTM as it would be a redundant unit in the simulation as HYSYSTM does not model

solid crushing.

The crushed seed is then sent to the rotary steam cooker, E-111, to bring the

water content down to 4% (O’Brien 2000) by heating it to 91°C as shown in Table 3.2.1.

In order to model this unit, SFA used a shell and tube heat exchange model in HYSYSTM

to determine the amount of low pressure steam that is required to heat up the incoming

feed to 91°C. The steam rate was found to be 265 kg/hr at 148°C and 4.5bara.

Table 3.2.1: Important Parameters for Physical Extraction of Canola Oil

Once the canola seed had the right moisture content, it was sent to the screw press

C-110 where 75% of the oil was extracted out of the seed (O’Brien 2000) this was

modeled using a component splitter in HYSYS. The remaining 25% of the canola is

recovered using hexane extraction.

Once 75% of the oil is recovered via the screw press, the remaining canola oil is

extracted using a hexane leaching system capable of 98% canola oil recovery. (O’Brien

2000). The “Oil to Hexane Recovery” stream from figure 3.2.1 is cooled to 59°C using a

10.0°C, 756.0 kg/hr water stream in the shell and tube heat exchanger E-112. This

temperature decrease is needed for optimum recovery of canola oil in the hexane leacher

H-120.

Stream Oil Flowrate (kg/hr)

Meal Flowrate (kg/hr)

Temperature °C

After roll Mill 2080 2080 25.0 Steam 1 (Saturated

Steam) 265 (steam) N/A 148

Steam 2 (Liquid) 265 (water) 147 2 2080 2080 91

Oil Recovered 1726 0 91 Oil to Hexane

Recovery 354.47 2080 91

26

The counter-current leaching operation using a hexane solvent is modeled as a

splitter unit in HYSYSTM . Unfortunately, the splitter was the only unit for leaching

simulation as there is no other suitable unit within HYSYSTM. Thus literature data from

O’Brien (2000) was used to model and specify the leaching operation of splitter H-120 in

the simulation. Staging of the leacher is discussed in appendices A and E of the report.

The oil rich micella is pressurized to 15 MPa using a reciprocating positive

displacement pump K-122 before being sent to the nanofiltration membrane separators to

remove the hexane from the oil. This pressure is needed to accomplish the separation as

explained by Ribeiro et al. (2006). The efficiency of the reciprocating pump was

specified at 75% as shown in Ulrich (2004). The power requirement of K-122 was

determined to be 11 kW from the HYSYSTM simulation.

According to HYSYSTM, the discharge temperature of the micella from the pump

is 71°C. It will then have to be cooled 59.0°C for membrane separation (Ribeiro, 2006)

by using a heat exchanger with a 1260 kg/hr 43°C counter-current water stream.

The nanofiltration membrane, H-140, was modeled using a splitter and specifying

the recovery of hexane in the system. Membrane separation has an advantage over other

forms of separation because of the minimal energy requirement in comparison to

evaporation techniques (Ulrich, 2004). According to Wu, the separation of hexane from

triglycerides is possible due to differences in their molecular weights, 86 to 900

kg/kgmol, respectively (Wu 1999). Also, hexane has a greater diffusivity with respect to

oil when the pore size gets smaller and thus membrane separation is possible (Wu 1999).

When Ribeiro et al. tested the separation power of the nanofiltration membranes, they

found that 68.0% of the oil was retained in the retentate at 59°C and the permeate stream

27

only had 8% oil by weight (Ribeiro 2006). Using the research by Ribeiro et al., H-140

splitter was modeled under the same conditions as shown in table 3.2.2 .

Table 3.2.2: Important Stream Parameters for Chemical Canola Oil Extraction

*Has a pressure of 15MPa for membrane separation as presented by Ribeiro et al.

The oil rich retentate, from the membrane separator, is then sent to an evaporative

unit R-150. In this unit, steam is used to evaporate 95% of the hexane originally in the

micella feed. As shown in table 3.2.2, the amount of power required to accomplish this

separation is 12.5 kW as determined by the Non-Random Two Liquid (NRTL) fluid

package in HYSYSTM. The remaining canola oil and hexane is sent to a steam stripper,

D-160, to increase the recovery of hexane from the micella.

After the evaporative unit R-150, the “hexane evap” stream only contains 1.7% wt

hexane in the stream and the micella will be further purified by running the micella into a

steam stripper. In the stripping unit, a counter-current steam of 130°C and 101.3 kPa,

produced the by the boiler E-182, is used to lower the hexane concentration down to 200

ppm in exiting canola oil stream. The stream rate required to obtain this specification is

223.0 kg/hr. The exiting canola oil with greater 99.98% purity, as shown in table 3.2.2, is

then mixed with the remainder of the canola oil that was extracted via physical separation

Stream Flow Rate (kg/hr)

Oil wt% Hexane wt% Temperature °C

Micella 1331 36 63 59.0 Micella HP LT*

1331 36 63 59.0

Retentate 436.3 75 25 59.0 Permeate 895.1 82 12 59.0 Hexane Evap 1 103 0 1.0 70.0 Micella Evap 334 98 2 70.0 Hexane Steam 228.0 N/A 98 110.0 Hexane make-up

0.72 N/A 1.0 25.0

24 1666 9 91 59.0 14 1666 9 91 59.0

28

in mixer L-161. The hexane rich steam is then cooled to condense the water and hexane

using heat exchangers E-164 and E-165. A gravity separator is used to separate both the

hexane from the water in unit H-180. Hexane is then recycled back to the hexane leacher

H-120.

Hexane recovered from H-140, R-150, D-160, is recycled back to the hexane

leaching unit H-120. A make-up hexane stream of 5200 kg/year is needed to maintain the

source hexane rate of 1.22x107 kg/year which is needed for solvent extraction. As shown

in the hexane recycle stream in figure 3.2.1, a balance was incorporated within the PFD

to reduce the degrees of freedom and make the recycle problem solvable. This

specification requires that the recycle stream for the leaching system must be equivalent

to what was arithmetically determined by mixer L-182. Without this specification, the

inlet hexane rate could not be determined in the simulation and the subsequent flows for

the separation would be zero in the simulation.

3.3 Meal Preparation After hexane extraction of the canola seed in leaching unit H-120, the remaining

seed is further processed to be sold as a protein rich canola meal for animal feed (O’Brien

2000). The underflow of the leaching system consists of solids which contains 30%

hexane and will be purified to make meal. This mixture of hexane and meal is known as

marc. The marc is removed of hexane through the use of a steam desolventer at 70°C and

warm air drying as shown in table 3.2.2 on the previous page. The hexane desolventer is

modeled by using the component splitter H-130 and air exchanger E-132. The energy

requirement for the hexane desolventer is approximately 61 kW and it will require a

steam rate of 80 kg/hr to remove 99% of the hexane in the marc (O’Brien, 2000). The

29

removed hexane is recycled back into the leaching unit H-120. The meal is cooled to

25°C by contacting the meal to air within the bottom part of the desolventer. Through the

use of the air cooled heat exchanger, E-132, the air rate needed for cooling the meal is

approximately 424 tonnes/day of air. The resultant meal will be produced at a rate of

2.1tonne/h and sold as a low cost animal feed.

3.4 Biodiesel Conversion After 95%-99% of the canola has been recovered for the feedstock seed, it then

enters into the process to produce biodiesel and glycerol. A PFD for the biodiesel

conversion process is shown in Figure 3.4.1.

Figure 3.4.1: HYSYSTM for the biodiesel conversion process. Material streams are shown in blue and energy streams are shown in red. Please note that Stream 39 comes from the canola extraction process. Tables 3.4.1 and 3.4.2 show important stream parameters

30

Table 3.4.1: Important Stream Parameters before Biodiesel Conversion Stream Flowrate

(kg/hr) Wt% Oil Wt%

Methanol Temperature (°C)

Pressure (kPa)

32 1887 100 0 40.0 101.3 Inlet Methanol

1330 0 100 43.0 101.3

26 (WCO) 148 100 0 25.0 101.3 31 3365 60 40 34.0 4240 35 3365 60 40 200.0 4240

Table 3.4.2: Important Stream Parameters after Biodiesel Conversion Stream Flowrate

(kg/hr) Wt% Methyl Ester

Wt% Methanol

Wt% Glycerol

Temperature (°C)

Pressure (kPa)

Rxn Product

3365 60 20 20 200.0 4240

33 626 0 100 0 200.0 101.3 30 626 0 100 0 63.0 101.3 34 2739 75 0 25 200.0 101.3 39 2739 75 0 25 55.0 101.3 Biodiesel 2064 1.00 0 0 55.0 101.3 Glycerol 675 0 0 100 55.0 101.3

After the canola extraction process, the canola oil is mixed with 208 tonnes/year

of citric acid to emulsify proteins in the canola feed called gums (O’Brien 2000).

The gums are separated through strainer H-170 where 100% of the gums are removed

from the oil and combined with the meal to increase its protein content. The canola oil

will be combined with 1000 tonnes /year of waste cooking oil (stream 26), and 9000

tonnes/year of methanol (Inlet methanol). Note the make-up methanol is only 5000

tonnes/year as 44% of the methanol is recovered from the reaction.

After mixing the canola oil, waste cooking oil, and methanol; the mixture is then

pressurized to 4240 kPa(g) using a 6kW, 75% efficient, multistage reciprocating pump to

achieve the desired pressure as calculated by HYSYSTM.

31

According to experimental optimization done by Kathlene Jacobson in the

summer of 2008, the reaction occurs best at 200°C. The reactants will be heated to 200°C

via two heat exchangers E-213A, E-213B and a fired heater E-213 which was modeled to

use 391.7 kW of power.

With reactants specified at the correct reaction conditions, they are charged into a

packed bed reactor which is modeled by the conversion reactor P-220. The reaction was

specified in the global reaction set as shown in equation X which is the same reaction as

specified by Zheng et al (2003).

Where the triglyceride is the canola oil and waste cooking oil reactants and FAME is

Free Fatty Methyl Esters or biodiesel. Therefore to specify the reaction in the simulation,

the stoichiometric coefficients were given as -1,-3,1,1 respectively for reaction X. From

Jacobson et al. (2008) the reactor was specified to achieve 95+% conversion using zinc

ethanoate/silica. The reactor was modeled as an adiabatic, isothermal and isobaric

reaction to be consistent with Jacobson et al.

Since the methanol was added in 50% excess, there is 4500 tonnes/year that can

be recycled. A flash tank at atmospheric pressure is used to vaporize the excess methanol

in unit H-203. The methanol leaves the separator at 200°C and is used to heat up the

reaction mixture. Finally, it is sent to a condenser, E-232, to cool it to 63.0°C using 2x103

tonnes/year of water at 10.0°C, before the methanol is recycled back to feed steam 37.

The biodiesel and glycerol, which comes off the bottom of the separator H-230, are

cooled through heat exchangers E-301 and E-213A to lower the temperature of the

mixture before being sent to gravity separators H-300. In the three gravity separators, the

Triglyceride + 3CH3OH → FAME + Glycerol (X)

32

residence time will be approximately 3 hours and 100% separation will occur as observed

in experiments (Jacobson 2008). The production of biodiesel was calculated by

HYSYSTM to be 16,000,000 litres/year, which is within the desired constraints of the

design.

3.5 Heat Exchangers and Process Heaters One advantage to our process is over 700 kW can be recovered via 11 heat

exchangers as shown in Table 3.5.1.

Table 3.5.1: Parameters of Heat Exchangers Used in Process as Modeled by HYSYSTM Unit E-164 E-165 E-124 E-183 E-112 E-121 E-162 E-301 E-213A E-213B E-232 Transfer Energy (kJ/h)

3x104 5x105 3x104 6x104 2x105 6x104 2x105 5x105 2x105 1x105 7x105

∆Tin 30 89 28 27 81 14 83 101 67 145 60 ∆Tout 11 40 9 1 2 8 3 1 0.75 0.04 1 ∆Tlm 19 61 17 8 21 11 24 22 15 18 14

Therefore, with the aid of HYSYSTM SFA was able to maximize the recovery of excess

heat from various process streams. These HYSYSTM values for the heat exchangers were

then compared with the shortcut calculations presented in Ulrich (2004) and were found

to agree within 0.47%. A detailed sample calculation is presented in Appendix A. With

this small error associated between the shortcut and HYSYSTM calculated values, the

HYSYSTM values were taken as reliable and were used to size the individual heat

exchangers.

Through the extensive use of heat exchangers, SFA was able to minimize the

number of process heaters down to only two units, E-182 and E-213. The HYSYSTM

parameters are shown on Table 3.5.2 as presented on the next page.

33

Table 3.5.2: Parameters for Process Heaters in Biodiesel Plant Unit No. Description Fuel Power (kW) E-182 Steam Fire-Tube

Boiler No. 6 Fuel Oil 11

E-213 Thermal Fluid Heater

No. 6 Fuel Oil 392

As shown in table 3.5.2, HYSYSTM was able to calculate the power requirement of the

heaters from the heater simulation unit function. These values were compared to short-cut

methods as shown in appendix A and agreed within 1% of each other. It is interesting to

note that the process heater used to heat the reactant feed is the most energy intensive unit

within the complete process as shown in Table 3.5.3 below. The heater E-213 consumes

over 60% of the energy required for the process. Therefore, SFA recommends that more

research could be direct towards finding a catalyst that does not require the reactants to

be at a high temperature.

Table 3.5.3: Energy Required for the Process as Calculated by HYSYSTM and by the Short-cut Method.

Unit Energy (kW) % of Energy Consumed E-111 156.0 23.9 C-110 2.6 0.4 K-122 10.4 1.6 H-130 61.4 9.4 R-150 12.5 1.9 E-182 11.4 1.7 K-211 5.9 0.9 E-213 391.7 60.1 Total 651.9 100.0

34

3.6 Simulation Summary

Based on the HYSYSTM simulation, our process is operating with correct mass

and energy balances as predicted by the NRTL correlation package. HYSYSTM was an

excellent tool used to find the energy requirements for all the units in the process. Based

on the simulation of the complete process the utility requirement was determined. The

advantage of using HYSYSTM was to configure the heat exchangers to maximize the

sensible heat recovery in the process. The simulation also allowed SFA to optimize

recycle streams such as; methanol and hexane recycle streams. With only mass and

energy balances the recycle and heat recovery would be very difficult to calculate.

Therefore, HYSYSTM made our lives easier because it gave us accurate results and made

the design easier to optimize.

35

4.0 Process Alternatives The alternative processes for canola extraction and traditional biodiesel

production will be discussed in this section.

4.1 Canola Extractor Using Only the Screw Press In this alternative, the solvent recovery of canola oil is not considered and the

extraction of oil is only achieved by the screw press. The advantage of this alternative is

that the energy requirement is minimized. The energy savings associated with this

alternative is $55,000 as shown in Table 4.1.1 below.

Table 4.1.1: Energy Savings Associated with Screw Press Only Operation

Energy Solvent

Extractor Cost

$0.08/kWh

Energy Screw press

Cost $0.08/kWh

Energy Savings

kW $/year kW $/year $/year 254.25 $146,448 158.62 $91,365 $55,083

Table 4.1.2: Comparison of Physical Extraction versus Combined Extraction

Even though the energy saving associated with using only the screw press is quite

significant, there is a disadvantage of using a screw press as it only recovers 75% of the

oil within the seed (O’Brien 1999). With this lower recovery of oil, more seeds will have

to purchased, which will increase operating costs. Also a protein rich canola meal cannot

be produced because it will still contain 25% oil, which is an unwanted attribute for

livestock feeders (Ravel 2004). Therefore, Table 4.1.2 shows the complete cost

comparison for the screw press and the proposed design.

Cbm Screw Press Plant

Cbm Solvent Extraction

Plant Difference Energy Saving

Revenue from Meal

Profit by Using Solvent Extraction in first year

$ $ $ $ $/year $ $1,412,989 $2,771,758 $1,358,769 $55,083 $5,700,000 $4,286,148

36

From table 4.1.2, the combined extraction techniques of a screw press and a

solvent extraction train will generate a high protein meal worth $5.7 million dollars per

year in revenue. In the first year, the high revenue meal pays for the high capital cost and

larger energy requirement in excess of $4.2 million dollars. In other words, the revenue

from the meal pays for the hexane extraction facility in one year of operation. The screw

press extraction was considered, but the combined physical and solvent extraction was

chosen due to the value added meal.

4.2 Evaporator for Hexane Separation

Figure 4.2.1: Canola Extraction Facility Using Evaporator Separator (H-150a)

37

Another alternative to the crushing process is to use a falling-film evaporator

instead of a membrane separator to perform the initial step in the oil/hexane separation

process. The process is almost identical to the membrane separator cases in terms of

extraction process steps as shown in the HYSYSTM process flow diagram (PFD) in figure

4.2.1 on the previous page. This separator alternative does not affect the biodiesel process

in areas 200 and 300, only economics.

The main difference between the steam evaporator and the nanofiltration

separator is the degree of recovery of hexane from the micella. Steam separation achieves

greater relative recovery of hexane at 93% compared to the 87% recovery in the

membrane separator (Ribeiro et al. 2006). Also, the evaporator gives a pure hexane

recycle stream whereas the membrane separator will give a recycle stream that contains

8.0% wt of oil (Ribeiro et al. 2006).

Another advantage to the evaporator is that the individual capital cost for the unit

is far less when compared with the membrane separators. The cost of the membrane

separator is 88% greater than the evaporator as shown in Table 4.2.1 below.

Table 4.2.1: Cost Comparison between Evaporator and Membrane Separator Cost of Evaporator Cost of membrane Separators % Difference

$ Cbm $ Cbm $15,200 $132,000 88.5%

Table 4.2.1 was calculated on the basis of using carbon steel as the construction

material and the shortcut calculation procedure found in Ulrich (2004). A disadvantage of

the membrane separator is the membrane needs to be replaced yearly whereas the

evaporator does not require any major periodic maintenance.

38

The most obvious disadvantage to the evaporator is the energy requirement. The

membrane separator will give $13,000 per year of energy savings as shown in Table 4.2.2

below. Given that the plant will run for 10 years, the energy saving per year will be

enough to justify the high capital cost of the membrane separators in comparison to the

evaporator.

Table 4.2.2: Energy Comparison between Membrane and Evaporator Separator

Energy Evaporator Cost

Energy Membrane Cost

Cost Savings

% Savings

kW $/year kW $/year 87 $ 14,798 12 $ 2,041 $ 12,757 86%

Even though the membrane separator can pay for itself over the lifetime of the

project, it does not give much economic advantage over the steam evaporator. The

membrane separator was chosen due to the Kyoto accord and the Canadian Clean Air

Act that requires carbon dioxide (CO2) emissions to be reduced by half of 1990 levels by

2012 (CBC 2006). The carbon credits will have to be monitored closely to determine

when the membrane separator will be economically feasible. The current price for carbon

credits is $5.50/ton of CO2 (Carbonfund 2007). A further study on the amount of CO2

produced by using the evaporator instead of the membrane separator would have to be

undertaken to find the value of CO2 credits needed to make the membrane more

economically attractive.

4.3 Production of Biodiesel via Homogeneous Catalyst and Batch Reactor In this alternative, conventional production of biodiesel from waste cooking oil is

analyzed. This alternative is considerably more complex in comparison with the given

acid-catalyzed process. In a homogeneous base catalyst process, a two step procedure is

39

needed to convert triglycerides to biodiesel and glycerol in the presence of a short-chain

alcohol. However, an acid catalyzed batch reaction is required to neutralize the FFA’s

from the waste cooking oil in order to prevent the formation of soap through

saponification in an alkali catalytic reaction. This neutralization occurs as an

esterification reaction which produces methyl esters and water. The neutralized waste

cooking oil can proceed to undergo transesterification by the homogeneous base catalyst.

If oil has a free fatty acid (FFA) content greater than 2 wt%, it must be treated

with liquid sulphuric acid (homogeneous acid catalyst) to remove the FFAs. WCO

produced from Saskatchewan based processing plants is typically yellow grease that

contains 5-15% FFA. Therefore, it must be treated with sulphuric acid to prevent

saponification.

In the alkali process, sulfuric acid and methanol are mixed together and fed into a

primary batch reactor with pre-treated waste cooking oil to produce methyl esters and

water. Excess methanol is added to speed up the rate of reaction of oil conversion to free

fatty methyl esters. However, the excess methanol varies in literature from 1:6 – 1:50 on

a molar basis. The neutralized waste cooking oil will be fed into a secondary batch

reactor with methanol and potassium hydroxide. The potassium hydroxide acts as the

homogeneous base catalyst to carry out the transesterification reaction. Also, canola oil

can be added at this stage to increase biodiesel production.

Typically, methanol is added to the neutralized oil in a 1:3, oil to alcohol ratio and

is mixed with 1 wt% potassium hydroxide catalyst for thirty minutes. After thirty

minutes, 80% reaction is achieved and solution is left for ninety minutes. After which the

glycerol and catalyst can be separated from the biodiesel and methanol via gravity

separation. The previous procedure is then repeated to achieve 95% conversion in the

40

second step of transesterification. Once the catalyst and glycerol have been separated

from the product mixture, methanol is removed from the biodiesel by washing with

water. The water is then removed by drying with a desiccant such as; silica or sodium

sulfate. Some processes use distillation to separate the biodiesel, methanol, and glycerol.

This method of methanol removal was seen in Milligan Biotech’s facilities on the

University of Saskatchewan Campus in October 2007. After separation a quality product

of pure biodiesel is obtained. A simplified schematic of the conventional process is

displayed in figure 4.3.1 below.

Figure 4.3.1: Basic Schematic of Convention Biodiesel Production Utilizing Homogeneous Catalyst in a Batch Reaction By comparing the conventional batch alkali process with SFAs’ acid catalyzed

process, it can be concluded that the alkali process more complex and expensive.

41

5.0 Safety Analysis

In order to identify hazards and operability problems a safety analysis was

performed using a HAZOP study and a DOW Fire and Explosion analysis. The purpose

of HAZOP is to determine how the plant might deviate from its designed intent. The Dow

Fire and Explosion Index is used to determine the possible explosion hazards and its

economical impact on the plant. MSDS sheets were collected and attached in Appendix

C. A summary of the MSDS is shown in Table 6 on the next page.

The proposed process for the conversion of oil to biodiesel is a fairly safe process

that has two major concerns. The first major concern is methanol at high pressure and

temperature in the reactor. Since methanol has a low boiling point and a low flash point it

has a high material factor of 16 according to the Dow Fire and Explosion Index in

appendix C. Since the methanol is heated with a fired heater and pressurized to 4240 kPa

creating a very explosive material. The fire and explosion index was calculated to be 243

which is an extremely high value. The full detailed analysis of the fire and explosion

index can be found in exhibit A of Appendix C. The radius of exposure was found to be

243ft which would destroy the entire plant. From the previously described factors it was

concluded that the reactors would be placed in a separate explosion proof room or in its

own separate building outside the plant to ensure safety for the operators and plant

equipment.

If the reactors were to explode the cost of the area exposed went from $3,200,000

to $200,000 once the reactors were isolated. The estimated replacement time for the

reactors would be 123 days giving a business interruption loss of $575,000. Refer to

exhibit B in the appendix C for the unit analysis summary.

42

Table 5.0: Chemical Hazard Information Summary

Component LD-50 or LC-50 (route/species) LEL/UEL TLV/Exposure Limits Reaction Hazards

Hexane C6-H14

Oral toxicity: 25000 mg/kg (rat)

Gas toxicity: 48000 mg/kg (rat)

Lower: 1.15% Upper: 7%

176 mg/m3 TWA – ACGIH

50 ppm TWA-ACGIH 1000 ppm TWA – 500

STEL 3500 ppm TWA – 1760

STEL

Can react vigorously with strong oxidizers

Canola Oil n/a No known toxicological effects from the product Non-flammable For oil mist: 10 mg/m3

None known for Liquid Avoid contact with strong

oxidizers

Citric Acid C6H8O7 Oral: 3000 mg/kg (rat)

Lower: 0.28 kg/m3 (dust)

Upper: 2.29 kg/m3 (dust)

No known toxicological effects from this product

Incompatible with oxidizing agents,

potassium tartrate, alkali, alkali carbonates and

bicarbonates, acetates, sulfides, and metal

nitrates

Methanol CH3OH

Oral: 5628 mg/kg (rat) Dermal: 15800 mg/kg

(rat) Vapour: 64000 mg/kg

(rat)

Lower: 6% Upper: 36.5%

250 ppm – TWA – 200 STEL

Can react vigorously

with oxidizers. Violent reaction with alkyl

aluminum salts, acetyl bromide, chloroform +

sodium methoxide, chromic anydride,

cyanuire chlorite, lead perchlorite, phosphorous

trioxide, nitric acid Exothermic reaction with

chloroform + NaOH

Biodiesel n/a No known toxicological effects from this product None known Not hazardous under the

criteria of OSHA Stable; Incompatible with

strong oxidizers

Glycerol C4H8O3 Oral: 8000 mg/kg

(mouse) None known No known toxicological effects from this product Hygroscopic

43

6.0 Equipment Sizing and Costs This section of the design report gives an overview of the logic behind how each

of the units were sized and priced in the process. For a more detailed quantitative

description of how each of the units were sized please refer to the sample calculations in

appendix A.

After the units were sized, and the material of construction was determined, the

bare module cost was found for each units using Ulrich (2004). According to Ulrich

(2004), these bare module costs represent the purchase price of the unit, the installation

cost, the unit foundation and installing the relative instrumentation. The prices for all the

units in the process were obtained from Ulrich using a CE index of 400 as the text was

published in 2004. However, to account for inflation, a more recent CE index of 528.7

obtained for October 2007 from the Chemical Engineering Magazine. (CE index 2008).

This index value was applied to all the bare module costs of the units to account for

inflation.

All the units were chosen to be constructed of carbon steel because non-corrosive

process fluids. Therefore it will be assumed that carbon steel will be chosen for the unit

material unless otherwise stated.

6.1 Roll Mill The first unit that the canola seed encounters in the extraction process is the roll

mill. The roll mill is where the canola seed is flaked into a thin layer. Roll mill energy

requirement was determined by using the power equation presented in table 4.5a in

Ulrich. The roll mill was designed to have a reduction ratio of four and from the

HYSYSTM simulation the canola feed rate is 1.16 kg/s which gives a power requirement

44

of 1.0 kW (Ulrich 2004). The bare module cost for the roll mill was found to be $42,000

given from Ulrich figure 5.16.

6.2 Rotary Steam Cooker E-111 This unit heats the canola to remove the water in the canola seed to a 1-3%

moisture content by heating to 91°C. Using low pressure steam at 148.5°C and the cooker

was modeled as a heat exchanger with a heat transfer coefficient of 700 W/m2K from

Table 4-15a in Ulrich. The heat transfer area was found to be 1.6 m2 and the bare module

cost for the unit was found to be $15,000.

6.3 Screw Press C-110 One of the most important units of the process is the screw press because the

majority of oil is recovered. In order to size this piece of equipment, the power

requirement of the press was determined from Table 4.23c in Ulrich. By pressing 1.2 kg/s

of canola seed, the power requirement was calculated to be 2.6 kW. The cost for the

screw was determined to be $412,000 using figure 5.58 in Ulrich (2004).

6.4 Leacher H-120 The next unit in the extraction process is the solid-solvent leacher. The type of

leacher used was the Rotocel design as it is simple and widely used in industry (Perry’s

2008). To size this piece of equipment, the number of stages and residence time had to be

determined from the equilibrium relationship of oil-seed-hexane interaction. An

equilibrium relationship for oil/hexane and oil/seed was determined by Zaher et al.

(2004). By using the rigorous stage wise calculation method as shown in appendix A and

E, the number of stages for the leacher was determined to be 11 with a residence time of

6 minutes per stage. The bare module cost for this unit is $152,000 as given in figure 5.53

in Ulrich (2004).

45

6.5 Membrane Pump K-122 The micella is pressured up to the desired pressure of 15 MPa used for the

membrane separators as explained by Ribeiro et al. (2006) using a multistage

reciprocating pump. Since the temperature of the pump is operating below 200°C and the

pressure is relatively high, cast steel was used to construct the multistage reciprocating

pump. From the HYSYSTM simulation and from the shaft work equation in Ulrich (2004)

the pump requires 11 kW of power in order to achieve the desired outlet pressure. The

bare module cost of the pump was determined to be $162,000 using figures 5.49 to 5.51

in Ulrich (2004).

6.6 Membrane Separators H-140 Using the research put forth by Ribeiro et al. (2006), nanofiltration membranes

were designed for the initial hexane separation. By using the maximum experimental flux

of 3.0x10-5m3/s.m2 (Ribeiro 2006) the total area of membranes was found to be 22.6m2.

Since the maximum surface area of an individual membrane separators is 2.4m2 (Ulrich

2004), 10 parallel separators will have to be employed to handle the volumetric flowrate

of micella in this process. For this area of membrane separator, the bare module cost for

these units is $132,000 as determined from figure 5.57a in Ulrich (2004).

6.7 Hexane Evaporator R-150 In the falling film evaporator, hexane will be evaporated out of the micella using

low pressure steam at 148°C. This unit will require approximately 13kW of steam energy

to evaporate the required amount of hexane. The procedure outlined in Ulrich on pages

133-136 (2004) was used to determine the dimensions of hexane evaporator. The

evaporator was determined to have a height of 0.6m and a diameter of 0.3m. The bare

module cost was determined to be $4,500 given by figure 5.22 (Ulrich 2004).

46

6.8 Hexane Stripper D-160 The rigorous calculation method was used to determine the dimensions of the

stripper and was completed by using the stage wise equilibrium integration technique as

shown in appendix E. The hexane steam equilibrium was modeled using Raoult’s law of

ideal liquid-vapor equilibrium. This approximation gave reasonable results for a

preliminary design of the column (O’Brien 2000). Using the stage wise equilibrium

method, the height of the column was determined to be 2.8m. The diameter of the column

was determined by finding the stage efficiency in the column achieved by using 6mm

Raschig Rings and using that value in equation 4.85 in Ulrich (2004). Using this method,

the diameter of the column was found to be 0.44m or 44 cm.

The stripping unit will be made from stainless steel because the steam in the

column is corrosive. Given this material of construction, the stripper will cost $48,000 as

given by figures 5.44-5.47 in Ulrich (2004).

6.9 Hexane Flash Drum H-180 In the flash drum, water and hexane are separated from each other from the

resultant hexane steam mixture in stripping unit D-160. Assuming that the drum can hold

10 minutes of liquid, and using the smallest diameter possible 0.30m (Hill 1999), the

flash drum was calculated to have a height of 0.8m. The cost of the flash drum is $18,000

as determined by figures 5.44-5.46 in Ulrich (2004).

6.10 Desolventer/Heater H-130 & E-132 The desolventer unit was the most difficult unit to size, as it consisted of an

evaporator in the upper layers and an air dryer in the lower layers. Therefore, the upper

portion of the column was modeled as an evaporator and the lower portion was modeled

as extra vapour disengagement stages. Using this criterion, the upper portion modeled as

47

an evaporator was found to have a height of 1.5m and a diameter of 0.7m as calculated

from pg 133-136 in Ulrich (2004). The lower portion of the column was found by adding

two evaporator heights to the bottom of the evaporator. Therefore, the total height for the

desolventer and heater was 2.8m. The cost of the unit was determined to be $100,000 as

shown in figure 5.33 for a Rotary Dryer in Ulrich (2004).

6.11 Heat Exchangers Since all the heat exchangers are operating at temperature below 200°C, the

material used to make the heat exchangers will be carbon steel. Since there are so many

heat exchanger units, Table 6.11.1 shows the area and costs associated with each heat

exchanger in the process. A sample calculation for these values is given in appendix A

for E-124 following the procedure given in Ulrich on pg. 191 (2004) and from Dr.

Nemati’s notes (CHE 324 2006).

Table 6.11.1: Heat Exchanger Size and Cost for Area 100.

HX Unit Heat Transfer Area m2 Bare Module Cost Cbm (2008)

E-112 7.0 $ 11,895.75 E-121 5.5 $ 11,895.75 E-124 1.9 $ 9,120.08 E-162 6.5 $ 11,895.75 E-164 0.3 $ 4,163.51 E-165 2.6 $ 9,913.13 E-183 2.2 $ 9,913.13

E-213A 9.4 $ 13,878.38 E-213B 3.2 $ 4,361.78 E-232 16.8 $ 23,791.50 E-301 5.1 $ 11,895.75

48

Table 6.11.2: Area 100 Unit Sizing and Cost

6.12 Mixer Design

First a volumetric flowrate was needed to design the mixer for the separation of

gum from canola oil. The volumetric flow rate going into the mixers was 41m3/day.

Three mixing tanks were used and rotated continuously. A plant height restriction of 5m

was implemented therefore the unit height was determined to be 4m. Using a 2:1 height

to diameter ratio the diameter was 2m giving a total volume of 12.6m3. The bottom of the

tank is conical shaped to aid in oil drainage. The volume of the cone was determined to

be .5m3 having a height of .5m and a diameter of 2m.The total volume of the tanks are

13.1m3. They will be filled with 10m3 of oil. Once the tank is full, a 50% weight mixture

of citric acid and water will be added until a concentration of 2500ppm is reached. Water

will be added until the total mixture is 2% by weight. A 40 hp sticky vacuum will be used

to suck the gum off the top of the tank. The oil will be drained out of the bottom through

a strainer. Our mixing blade will be .5m in diameter. The rule of thumb power

consumption for mixing was estimated to be .5kW/m3 giving a total daily energy

Unit Name Unit Number Height Diameter Area Power Material Costm m m2 kW $

Roll Mill 1 Carbon Steel $42,000Cooker E-111 1.6 Carbon Steel $15,000

Screw Press C-110 2.6 Carbon Steel $412,000

Leacher H-120 Carbon Steel $513,000Pump K-122 10 Cast Steel $162,000

Membrane H-140 22.6 $132,000Evaporator R-150 0.6 0.3 Carbon Steel $4,500

Stripper D-160 2.8 0.44 Stainless Steel $48,000

49

requirement of 111MJ. Using figure 5.42 in Ulrich the current estimated price was found

to be $55,800 for the three units.

6.13 Fixed Bed Reactor The fixed bed reactor is used for the conversion of triglycerides and free fatty

acids into biodiesel (methyl esters) and glycerol. Due to the high temperature and

pressure of this vessel, a stainless steel material will be used. The fixed bed reactor was

difficult to size because of no kinetic data. Without kinetic data, the size of the reactor

was based on the weight hourly space velocity, as defined by Vartuli et al, to obtain the

residence time of the liquid. Furthermore the reactor was sized based on 50% volume of

catalyst. From laboratory work, it was determined that a 3 wt% catalyst was capable of

achieving the desired conversion. The volume of the reactor was found to be 175 L with a

liquid residence time of 1.3min. The bare modular cost of this particular fixed bed reactor

is $57,600. The remainder of the reactor parameters can be viewed in table 6.14 on the

following page.

6.14 Flash Drum w/ Mist Eliminator This flash drum is used to separate methanol from glycerol and biodiesel,

produced in the fixed bed reactor. A mist eliminator is employed to arrest any liquid that

is carried along with the vaporizing methanol. This drum is designed for a residence time

of ten minutes. Based on this, the flash drum was calculated to have a diameter of 1.94 m

and a height of 2.04 m, giving a total volume of 12.1 m3. It is to be constructed of

stainless steel, to withstand high temperature and pressure and will cost $144,900.

50

Table 6.14: Sizing and Cost Summary of Fixed Bed Reactor and Flash Drum

Unit Number  P‐220  D‐230 Description  Fixed Bed Reactor  Flash Drum Height of Unit (m)  0.96  2.04 Diameter of Unit (m)  0.48  1.94 Residence time (min)  1.32  10 Volume (m^3)  0.173  12.06 

Cbm  $57,600.00  $144,900.00 

6.15 Gravity Separators To design the gravity separators for this plant the first variable needed was the

inlet flow rate of glycerol and biodiesel. The combined flowrate was 71.5m3/day and 3

tanks were designed to hold this volume for 4 to 6 hours of separation time. The

dimensions of the tanks were 5m for the height and 2.5m for the diameter, giving a total

volume of 24.5m3. Table 5.61 in Ulrich was used to determine the bare module cost of

$1,100 for each unit.

51

7.0 Economics

A complete economic analysis proves that SFA has a strong potential for

profitability. Many alternatives were considered, and the plant always sustained enough

income to generate a net positive cash flow.

A detailed cost analysis was performed for the alternative using a membrane

separator for hexane extraction and conservative estimates for the price of diesel, glycerol

and meal at $80/barrel, $.4/kg and $.38/kg respectively. The revenue generated from this

plant was estimated to be 16.4 million dollars. The biodiesel, the meal and the glycerol

produced 8.8, 5.7 and 1.9 million dollars respectively. The total capital investment for the

proposed biodiesel plant was estimated to be $3,300,000. The total manufacturing costs

was estimated to be $14,700,000 per year as shown in the detailed summary in table B-1

in appendix B.

The direct manufacturing expense was estimated to be 13.8 million dollars per

year. The cost for the raw materials, mainly greenseed canola, was estimated to be 11.7

million dollars per year. The cost of greenseed canola is around half the price of regular

canola seed which has increased 65% from November 2007. The conservative price used

for the greenseed canola was $.33/kg. It was estimated that it would take 17 operators to

maintain and run this facility which will cost $820,000 annually. Other main factors

contributing to the direct manufacturing cost are Supervisory and clerical labor, cooling

water, maintenance and repairs and laboratory charges calculated to be $160,000,

$110,000, $170,000 and $120,000 respectively.

The annual cost for indirect manufacturing expenses was determined to be

$590,000 which mainly consisted of overhead, packaging and storage. The total annual

52

general expense was estimated to be $640,000. Depreciation was determined to

be$290,000 annually and income tax was estimated to be $600,000 per year. From all of

these estimates, the after tax rate of return was determined to be a promising 48.6% as

shown in figure 7.1 below.

The discounted breakeven point (DBEP) was 3.5 years; the undiscounted and

10% discounted net present values were estimated to be $11.1 and $5.3 million dollars

after 10 years of operation. In comparing this scenario with the one presented in

alternative 7.3 it can be seen that the evaporator is more economical after ten years of

operation. However, the membrane separator was chosen because it produces less CO2

emissions and requires less energy.

Figure 7.1: Cash flow analysis

53

7.1 Alternative 1: Negating Glycerol Sales

The first alternative assumes the glycerol produced in plant is worth nothing. In order

to produce a net positive cash flow, the price of biodiesel has to be increased to $90 per

barrel, which is still realistic because the current price of diesel is $100 per barrel. This

method produces a discounted cash flow rate of return of 30%. The undiscounted and

10% discounted net present values were determined to be $6,000,000 and $2,500,000

respectively, after 10 years of operation. The discounted breakeven point occurs after 4.5

years of operation. This alternative is the most probable worst case scenario and the plant

still generates a positive net present worth.

Figure 7.1.1: Cash Flow Analysis with Negating Glycerol

54

7.2 Alternative 2: Using Grade 1 Canola as the Feedstock

The second alternative is the worst case scenario because the plant would have to

purchase expensive grade 1 canola seed as the major feedstock instead of the inexpensive

greenseed canola. Since canola seed is the main raw material, the fluctuations in the price

will have an extreme effect on our economics. The price of canola used for the analysis

was $650/tonne.

Figure 7.2.1: Alberta’s grade 1 Canola Oil Prices

In order for this process to still generate a positive net cash flow, biodiesel needs to

sell at the same price as diesel at the current $100/barrel and the glycerol needs to be sold

for $.7/kg. If these criteria are met, the discounted cash flow rate of return would be

18.5%. The undiscounted and 10% discounted net present values would be $3,200,000

55

and $950,000 respectively, after 10 years of operation. The discounted breakeven point

will occur at 6.5 years.

Figure 7.2.2: Cash Flow Analysis Using Grade 1 Canola

56

7.3 Alternative 3: Using an Evaporator for Hexane Extraction

The third alternative studied would be to use an evaporator instead of a membrane

separator for the hexane extraction unit. This method produces a discounted cash flow

rate of return of 50%. The undiscounted and 10% discounted net present values are

estimated at $11,500,000 and $5,500,000 respectively, after 10 years of operation. The

discounted breakeven point requires 3.1 years of operation. Although this alternative

produces a higher discounted cash flow rate of return in comparison to the membrane

case. Even though the evaporator case has favorable economics, it produces more CO2

emissions and consumes more energy than the membrane separator. Therefore, the

membrane separator will be used for our project design.

Figure 7.3.1: Cash Flow Analysis Utilizing Our Design’s Prices, But Using an Evaporator Instead of a Membrane Seperator.

57

7.4 Alternative 4: Selling Glycerol at Current Prices The fourth alternative was estimated using the current price of glycerol at $.7/kg.

This alternative produces a discounted cash flow rate of return of 87.5%. The

undiscounted and 10% discounted net present values are estimated at $22,800,000 and

$11,800,000 respectively, after 10 years of operation. The discounted breakeven point

requires 2.6 years of operation. This is best case alternative, so if glycerol prices remain

high this biodiesel plant could make a fortune.

Figure 7.4.1: Cash Flow Analysis Using Current Prices and a Membrane Seperator

58

7.5 Alternative 5: Selling Glycerol at Current Prices using an Evaporator

The fifth alternative uses the same prices for the products as in alternative 3, but uses an

evaporator instead of the membrane separator for hexane extraction. This alternative

produces a discounted cash flow rate of return of 88.9%. The undiscounted and 10%

discounted net present values are estimated at $23,200,000 and $12,100,000 respectively,

after 10 years of operation. The discounted breakeven point requires 2.6 years of

operation.

Figure 7.5.1- Cash Flow Analysis Using Current Prices and an Evaporator

59

7.6 Alternative 6: Producing Value Added Products from Glycerol

Another alternative considered was to produce value added products from the

glycerol produced. The glycerol would be sent to a fixed bed reactor with a silica alumina

catalyst. The four main products produced from this reaction are acetol, formaldehyde,

acrolein and acetaldehyde. The reaction also produces many side products such as;

Hydrogen gas, carbon monoxide, carbon dioxide, methane, acetone and allyl alcohol.

Since these products have different volatilities the separation process to acquire the four

main products would be too costly.

A summary of the alternatives are displayed in a table below. Table 7.5.1: Summary of Economic Alternatives

Alternative DBEP i=10% (years) 

NPV i=0% ($) 

NPV i=10% ($) 

Rate of Return (%) 

Glycerol Worth Nothing  4.5  6,000,000  2,500,000  30 

Grade 1 Canola Oil  6.5  3,250,000  950,000  18.5 

Regular Glycerol  price             

evaporator  2.6  23,200,000  12,100,000  88.9 

membrane separator  2.6  22,800,000  11,900,000  87.5 

Low Glycerol  Price             

evaporator  3.1  11,500,000  5,500,000  50 

membrane separator  3.5  11,100,000  5,300,000  48.6 

60

8.0 Conclusions 1. SFA will produce 15,000,000kg of biodiesel per year at $.59/kg giving an annual

revenue of $9,000,000.

2. SFA will produce 15,200,000kg of canola meal per year at $.38/kg giving an

annual revenue of $5,700,000.

3. SFA will produce 4,900,000kg of glycerol per year at $.59/kg giving an annual

rate of $1,900,000.

4. Glycerol was deemed to be a viable source of income because the world demand

is 1,600,000 tonnes/year and is increasing with new technology.

5. SFA will consume 30,000,000kg of greenseed canola per year at $.33/kg giving

an annual cost of $9,800,000.

6. SFA will consume 1,100,000kg of waste cooking oil per year at $.15/kg giving an

annual cost of $100,000.

7. SFA will consume 5,100,000kg of methanol per year at $.34/kg giving an annual

cost of $1,700,000.

8. The total capital cost of the plant was determined to be $2,800,000.

9. At a 10% discounted rate the net present value after 10 years was calculated to be

$5,300,000.

10. The after tax rate of return was calculated to be 48.6%.

11. In comparing the membrane scenario to the evaporator scenario for hexane

separation, it was determined that the evaporator was more economical. But the membrane separator was chosen because of reduced CO2 emissions and energy requirements.

12. The acid catalyst, zinc ethanoate supported on silica has 95+ conversion rate.

61

13. The fixed bed reactor was deemed to be a severe hazard and will be isolated from

the plant.

14. The heterogeneous acid catalysis process has fewer process steps in comparison

with a homogeneous alkali catalyst process.

15. By utilizing a solid catalyst and a low quality feedstock, the biodiesel price is

theorized to be competitive with current diesel prices.

62

9.0 Recommendations 1. The biodiesel reactor needs more kinetic data to obtain more accurate reactor

properties and sizing.

2. Membrane separators need to be tested in pilot facility to determine their compatibility with the other units in the process.

3. Pipeline pressure survey should be completed before an actual plant is built to more accurately predict frictional losses and pumping requirements.

4. Supercritical CO2 oil extraction should be investigated as an attempt to lower the solvent extraction energy requirements in the given process.

5. More research should be conducted to find a catalyst that requires a lower reaction temperature.

6. Meal standards may change and physical extraction can be utilized for canola

recovery to minimize energy and unit costs associated with solvent extraction.

7. Propylene glycol can be created as a value added product if glycerol demand decreases or glycerol supply increases from other biodiesel projects.

8. Profitability of project is largely dependent of on feedstock prices. More testing

should be done on feedstocks other than canola seed as canola prices may rise to a level that will make this process uneconomical.

9. Further study on the characterization of the finished product is needed to ensure that produced biodiesel meets standards.

63

10.0 References Carbonfund, Leading Offset Provider Value Chart, Carbonfund.org. 2007. April 3, 2008.

< http://www.carbonfund.org/> CBC News, In Depth Kyoto and Beyond Trading Carbon, CBC News. November 3, 2006. < http://www.cbc.ca/news/background/kyoto/carbon-trading.html> CE Index October 200, Economic Indicators, Chemical Engineering. February 2008.

Pg. 68. Faye Zenneth, CEO Milligan Biotechnology, Foam Lake. Personal Communication. Forge, Frederic. Biodiesel – An Energy, Environmental or Agricultural Policy? Library

of Parliament. 8 February 2007

Litherland N.B. et al., Dry Matter Intake is Decreased More by Abomasal Infusion of Unsaturated Free Fatty Acids than by Unsaturated Triglycerides, J. Dairy Sci. 99 (2005) 632-643.

Hill Gordon. Chemical Engineering Process and Design I, CHE 325, University of

Saskatchewan (1999). Kiss, A.K.; Dimian, A.C.; Rothenberg, G. Solid Acid Catalysts for Biodiesel Production

– Towards Sustainable Energy. Adv. Synth. Catal. 2006, 348, 75-81. Krawczyk, Tom. Biodiesel. INFORM. Vol. 7, No. 8, August 1996. 800 Kulkarni, M.G. and Dalai, A.K. Waste Cooking Oil – An Economical Source for

Biodiesel: A Review. Ind. Eng. Chem. Res. 2006, 45, 2901-2913 Lotero, E.; Liu, Y.; Lopex, D.; Suwannakarn,K.; Bruce, D.A.; Goodwin, J.G. Synthesis

of Biodiesel via Acid Catalysis. Ind. Eng. Chem. Res. 2005, 44, 5353-5363. Nemati Mehdi, Chemical Engineering Heat Transfer, CHE 324, University of

Saskatchewan (2006). O’Brien Richard D., Farr Walter E., Wan Peter J., Introduction to Fats and Oils

Technology, 2 edition. Champaign: AOCS. 2000. Pg. 109-135 & Pg.545. Perry R.H., Green D.W., Maloney J.O., Perry’s Chemical Engineers’ Handbook,

7 edition. Toronto: McGraw-Hill. 1997. Perry R.H., Green D.W., Maloney J.O., Perry’s Chemical Engineers’ Handbook,

8 edition. Toronto: McGraw-Hill. 2008.

64

Racz V., Christensen D.A. , Whole Canola Seed use and Value, Prairie Feed Resource Centre. (2004).

Ribeiro et al., Solvent recovery from soybean oil/hexane miscella by polymeric

membranes, J. of Membrane Science 282 (2006) pg. 328-336. Sea News Circular, Canada Moves on Renewable Fuels, Sea News Circular. Vol: X,

Issue:10. Jan., 2008. Titipong, I.; Kulkarni, M.G.; Dalai, A.K.; Bakhshi, N.N. Production of Biodiesel from

Waste Fryer Grease Using Mixed Methanol/Ethanol System US Department of Energy, Clean Cities – Alternative Fuel Information Series – Fact

Sheet, May 2001 Ulrich G., Vasudevan P. Chemical Engineering Process and Economic A Practical Guide

2nd edition. Process Publishing. Durham. (2004). Zaher F.A., El Kinawy O.S., El Haron D.E., Solvent extraction of jojoba oil from pre-

pressed jojoba meal, Grasas y Aceites. 55, 2, (2004), 129-134. Zheng S., Kates M., Dube M.A, McLean D.D. Biodiesel Production from waste Cooking

Oil 1. Process design and technological assessment, Bioresource Technology 89 (2003) 1-16.

65

Appendix A

Sample Calculations

66

A.1. Sample Calculation for Sizing and Cost of Area 100 Note: All Figures Referred to for the Sample Calculations are from Ulrich (2004)

1. Sample Calculation for Power Requirement of Roll Mill

RmP.

20.0= Equation from Table 4.5a for Roll Crusher of medium hardness

13.635,41$400

7.52810.2000,15$

)()(

))((

10.2000,15$

/16.1925.0

)4)(/16.1(20.0

2004

2008

=

=

=

==

===

Cbm

xxCbm

CECE

FbmCpCbm

FbmCp

skgCapacitykWP

skgP

index

index

2. Sample Calculation for Rotary Steam Cooker E-111

2

2

3

2

.

12

63.1

)5.82(700

)101.85(

)(

2.85)(

))91.91148(

)25148(ln(

)91.91148()25148()(

))()(ln(

)()()(

)(1.85

)1036.7)(7.2119(

ker7.2119

9.274216.623

mA

CCm

WWxA

TTsUPA

CTTsCC

CCCCCCTTs

TTTT

TTTTTTs

TTsUAPkWP

skgxkg

kJP

mHP

CootoSuppliedkgkJH

TablesSteamFromkgkJ

kgkJH

ChangeSteamForHHH

lm

lm

lm

outs

ins

outsinslm

lm

=°

=

−=

=−

−−

−−−=−

−−

−−−=−

−==

=

Δ=

=Δ

−=Δ

−=Δ

°

°

°°

°°

°°°°

−

K

67

Heat Transfer Coefficient Obtain from Table 4-15a, for Condensing Water on Hot side and Low-Viscosity Organics in Cold Side. The Bare module cost for the rotary steam cooker was found using Ulrich from Figure 5.24 for evaporators. Since the unit is at temperatures below 200°C, then carbon steel was selected for materials of Construction. Since the steam cooker was operated at atmospheric pressures Fp=1.0

13.200,15$)()(

3.21

24.55000$

2004

2008

=

×××=

===

CbmCECECpFpFbmCbm

FbmFp

FigCp

index

index

3. Sample Calculation for the Shell-&-tube Heat Exchanger. E-124

2

2

3

11

12

12

21

87.1

)94.0)(7.16)(300(

)104.9(

)(

191.22.494.043.0

)01.7100.43()01.7100.59(

704.)()(

59.0)01.7100.59()04.5000.43(

704.

mA

CCm

WWxA

FTTUQA

pgFigurefromFS

CCCCS

UlrichfromEqtTtt

S

RCCCCR

UlrichfromEqttTT

R

Tlminout

T

=

°°

=

Δ−Δ=

==

−−

=

−−−

=

=−−

=

−−−

=

°°

°°

°°

°°

Same Procedure for all the Heat Exchanger in Area 100, 200, and 300. Results shown on Table 6.11.1

68

The bare module cost associated with heat exchangers was determined by using figures 5.36 to 5.38 in Ulrich (2004). All the heat exchangers will be constructed using carbon steel because all the units will be operating at temperatures below 200°C. A pressure factor of one was used for all heat exchangers because the heat exchangers are all at atmospheric pressure.

80.120,9$)()(

311

00.2300$

2004

2008

=

××=

====

CbmCECE

FbmCpCbm

FbmFp

steelcarbonFmCp

index

index

4. Sample Calculation for Screw Press Unit C-110

kWPs

kgP

mP

62.2

)10.1(5.2

5.2

5.0

5.0.

=

=

=

Duty Equation Obtained from Table 4.23c in Ulrich (2004). Using Figure 5.58, the Bare Module Cost of the screw press was determined. Carbon steel is used to be more economical

00.386,412$)()(

4.2000,130$

2004

2008

=

××=

==

CbmCECEFbmCpCbm

SteelCarbonFbmCp

index

index

69

5. Sample Calculation for Number of Stages and Residence Time of Solid-Liquid Extractor as well as the Bare Module Cost H-120

80.1

/370.0/664.0

=

=

solvent

solid

solvent

solid

FF

skgskg

FF

min611

min601

)'(11

)(1

91.3)(

1/256.0)(

282.0537.0)(

/537.0

)165.0)(255.3(

/165.0165.0)282.0282.0(556.0

)(

1 1*

1

1*

1

1*

1

1*

1

*1

*1

*1

0

0

011

=

==

==

−=

=−

=−

−=−

=

=

=

=+−=

+−=

∫+

+

+

TimeConstactStagesn

TimeTimeContact

hrTimeRulesSimpsonstagesn

dYYY

n

YY

kgSolventkgSoluteYYYY

kgSolventkgSoluteY

Y

mXY

kgSolidskgSoluteXX

LineOperatingXYYFF

X

ov

ov

Y

Yov

nn

nsolid

solventn

n

n

Equilibrium Value from Canola Oil-Hexane Equilibrium Obtained from

Zaher (2004). Bare module cost for leacher determined from figure 5.53.

00.000,152$)()(

3.2000,50$

2004

2008

=

××=

==

CbmCECE

FbmCpCbm

FbmCp

index

index

70

6. Sample Calculation for Power Requirement of Membrane Pump Unit K-122

kWP

PaxsmxP

ePqP

6.1375.0

)1015)(/1079.6( 734

.

=

=

Δ=

−

Efficiency was estimated from Table 4.20 in Ulrich for Reciprocating Pump. Bare Module Cost was determined from figures 5.49 to 5.51. Cast Steel was used as the temperature of the Process was well below 200°C.

38.914,161$)()(

9.4arg101

8.100.000,25

2004

2008

=

××=

=<=

==

CbmCECEFbmCpCbm

FbmbFp

SteelCastFmCp

index

index

7. Sample Calculation for Size and Number of Nanofiltration Membrane Units H-

140

6.936.26.22

36.2

)5.1)(5.0(

5.05.1

6.22/1000.3

/1079.6

/1079.6

2

2

2

2

235

34

34

=

=

=

=

=

===

=

==

=

−

−

−

Separators

Separators

Separator

TotalSeparators

Separator

Separator

Separator

Total

TotalTotal

Total

NmmN

AA

N

mA

mmA

DLAmD

mLmA

smmxsmx

FluxQ

A

smxQ

π

π

Flux Obtained from Ribeiro et al. Length and Diameter of Separator Obtained from Table 4.23b.

71

Bare Module Cost was obtained by using figure 5.57a in Ulrich. Nanofiltration was used for costing as that was the same membrane used by Ribeiro et al. 2006. Cbm=$132,175.00

8. Sample Calculation for Rising Film Evaporator Unit R-150

2

2

2

.

2.

253.0

)5.87(8.0

47.12)(

47.12)/1044.4)(/5.334(

/5.334

/1044.4

mA

KKm

kWkWA

TTUPA

kWPskgxkgkJP

mHP

kgkJH

skgxm

jacket

jacket

lmsjacket

vapvap

vap

vap

=

=

−=

==

Δ=

=Δ

=

−

−

72

( )

mH

mm

mH

DD

AH

mD

mD

AD

mA

smmkgskgA

umA

mlsumkg

mkgmkgu

u

jacket

tioncross

tioncross

tioncross

gtioncross

g

gl

58.0

23.0)23.0(

253.0

23.0

)040.0(4

4

040.0

)/957.0)(/06.3(/117.0

957.0/06.3

/06.3/3.78106.0

06.0

2

2/12

2/1sec

2sec

3sec

.

sec

3

33

=

+=

+=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

⎟⎠⎞

⎜⎝⎛=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

−

−

−

−

π

π

π

π

ρ

ρρρ

Procedure Outlined in Ulrich Pages 133-136 for Short-Cut Design. Physical Properties of Hexane from Perry’s Handbook.

Bare Module Cost was obtained from figure 5.22 in Ulrich. Carbon steel will be used because the temperature is low. The Evaporator will also operate under atmospheric pressure.

04.560,4$)()(

3.21

24.500.500,1$

2004

2008

=

×××=

===

CbmCECECpFpFbmCbm

FbmFp

FigCp

index

index

73

9. Sample Calculation for Hexane Stripper Sizing Using Raschig Rings. Unit D-160 Determine the diameter of the column from short cut methods in Ulrich on Page 240. Figure 4.31.

TraycapBubbleforEfficiencyStageesaxisx

mkgsPakgmolkgaxisx

mMaxisx

l

ll

−==−

⎟⎟⎠

⎞⎜⎜⎝

⎛=−

=−

178.000068.0

/1150.0102.0*/100*756.0

3

ρμ

mz

mx

z

mxx

xxz

HETPNovzmD

mD

UlrichfromEquationesHETPD

mHETPsmmkgmolex

skgmolexHETP

KLHETP

lm

la

81.2

987.01092.7

)0150.0241.0(

987.0)(

44.04.0

)178.0)(987.0(

85.44.0

987.0/10844.5

/1021.9

2

*01

324

5

=

×−

=

×−−

=

×==

=

×=

=

=

=

−

−

−

Mass Transfer Coefficient Obtained from Mass Balance on Column

74

Bare Module Cost for the Stripper was determined using Figures 5.44-5.47. Stainless Steel a there is corrosive materials in the stripper. 6mm Raschig rings were used in the tower.

18.715,47$)()(

5.94

10.560,4$

2.10.800,3$

2004

2008

=

××=

=

=====

CbmCECE

FCbmCbm

FFmFpCbmFbm

pC

index

indexabm

abm

ss

10. Sample Calculation for Desolventor Energy Requirements/Sizing Units H-130 and E-

132 Treat as an evaporator that removes hexane from Meal.

2

2

1

.

1.

6.1

)6.91(8.0

)38.61()(

38.61)/1081.2)(/5.334(

/5.334/1081.2

mA

KKm

kWkWA

TTUPA

kWPskgxkgkJP

mHP

kgkJHskgxm

jacket

jacket

lmsjacket

hexanevap

vap

hexane

=

=

−=

==

Δ=

=Δ=

−

−

75

( )

mH

mm

mH

DD

AH

mD

mD

AD

mA

smmkgskgxA

umA

smumkg

mkgmkgu

u

jacket

tioncross

tioncross

tioncross

gtioncross

g

gl

83.2

)68.0(3)68.0(

6.1

3

68.0

)36.0(4

4

36.0

)/855.0)(/06.3(/1049.9

/855.0/06.3

/06.3/62506.0

06.0

2

2/12

2/1sec

2sec

3

1

sec

.

sec

3

33

=

+⎟⎟⎠

⎞⎜⎜⎝

⎛=

+=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛ ×=

⎟⎠⎞

⎜⎝⎛=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

−

−

−

−

−

π

π

π

π

ρ

ρρρ

Procedure Outlined in Ulrich Pages 133-136 for Short-Cut Design. Physical Properties of Hexane from Perry’s Handbook. Bare Module Cost for the Rotary Dryer was found by using figure 5.33 in Ulrich. Material of construction is carbon steel because of the low temperature of the unit operation.

25.131,99$)()(

5.100.000,50$

2004

2008

=

××=

==

CbmCECE

FbmCpCbm

FbmCp

index

index

76

11. Sample Calculation for Hexane Flash Drum H-180

mDsmmkg

kgmolkgskgmolxD

uVMgD

smumkg

mkgmkgu

u

gsg

gs

gs

g

glgs

02.0)/138.1)(/066.3(/18.86/106169.14

4

/138.1/06.3

/06.3/0.97306.0

06.0

2/1

3

5

2/1

,

,

3

33

,

,

=

⎟⎟⎠

⎞⎜⎜⎝

⎛ ××=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

−

π

πρ

ρρρ

Design to hold 10 minutes of liquid in the bottom of the Separator

mLe

mmxLe

DV

Le

mxV

hourh

mV

hoursmV

liq

liq

liq

liq

42.118

)02.0()1072.3(4

4

1072.3

)167.0(2235.0

)167.0(

2

32

32

3

.

=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

⎟⎟⎠

⎞⎜⎜⎝

⎛=

=

=

=

−

−

π

π

ρ

Too big assume D=0.30m

mHDLe

mLe

m

mxLe

DliqV

Le

83.075.1/

53.0

2)30.0(

)321072.3(4

4

==

=

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛ −=

⎟⎟⎟

⎠

⎞

⎜⎜⎜

⎝

⎛=

π

π

77

Procedure Outlined from Dr. Hill’s CHE 325 notes. Properties for water and hexane obtained from Perry’s Handbook. Bare module cost was determined from figures 5.44 to 5.46 in Ulrich. Carbon Steel was used because there is no corrosive material in the feed.

63.843,17$)()(

5.411

00.000,3$

2004

2008

=

××=

=

===

CbmCECE

FCpCbm

FFpFmCp

index

indexabm

abm

78

A.2. Sample Calculation for Sizing and Cost of Area’s 200 and 300 12. Sample Calculation for the sizing of the canola mixers (L-163)

Approximate volume needed for each tank

Using a rule of thumb of a 2:1 height to diameter ratio (Ulrich 579). A height of 4m and a diameter of 2m was used to calculate the actual volume. The tank is cylindrical with a conical funnel at the bottom. The propeller size was determined to be .5m in diameter which is the average maximum mixing blade size. (Ulrich 578)

Calculating the energy used for mixing assuming the tanks are filled with 12 cubes of canola oil. The rule of thumb for mixing is the power consumption should be between .3 to .7kW/m3 (Ulrich 578). A chosen conservative value of .5kW/m3 was used because canola oil is partially viscous but it is not thick.

79

From figure 5.42 in Ulrich using 6kW as the power consumption, the purchased equipment cost was estimated at 7,000$ and the unit was made of carbon steel giving a FBM of 2.0

Calculation for the bare module cost

Calculation for the current price using 528.7 as the current CE index value

13. Sample Calculation for Fixed Bed Reactor

Determination of the density of catalyst Zn Ethanoate/Silica in the laboratory

mLcatalystg

catalystVolumecatalystweight

catalyst

catalyst

3_1529.2

__

=

=

ρ

ρ

Calculation of weight of catalyst based on 3wt% of oil.

L

mkg

kgghrkgCatalyst

hrkgoiltemassflowra

3.866.717

61*612035*03.0

2035_

3

===

=

80

Using the LHSV definition as layed out by Vartuli et. Al as the weight of liquid reactant (ie. Methanol and oil) contacting a given weight of catalyst in one hour, we can find the LHSV based on the flowrate and the weight of catalyst:

uteshourtimeresidence

hrL

hrLhrLcatalystVolume

liquidofamountLHSV

min3.1022.0_

3.453.86

/1687/2222_

__ 1

===

=+

== −

θ

We then know the reactor volume can be calculated using the volumetric flowrate of the oil multiplied by the residence time plus the volume of the catalyst.

13.86022.0)/1687/2222( 1 =++= − LhrhrLhrLV The diameter and height of the reactor can then be calculated using the rule of thumb that the height should be twice the diameter.

( )

mdH

mVd

ddddrV

96.02

48.02

22

23/1

322

==

=⎟⎠⎞

⎜⎝⎛=

=⎟⎠⎞

⎜⎝⎛==

π

πππ

14. Sample Calculation for Fixed Bed Reactor

Calculation for the Methanol Flash Drum after the reactor uses the same method as the flash drum for hexane. However it should be noted that for the density of liquid, a bulk density was calculated for biodiesel and glycerol using the equation:

)()(biodiesel

biodiesel

glycerol

glycerol

biodieselglycerolbulk mm

mm

ρρ

ρ+

+=

Where m represents the mass flowrate. Furthermore, it should be noted that the bare module cost was calculated for stainless steal due to the high temperature and pressure in place of carbon steal. A mist eliminator with the same diameter as the flash drum was also added to the bare module cost of the flash drum.

81

15. Sample Calculation for the sizing of the three gravity separators (H-300). Calculation for the approximate amount of glycerol/biodiesel mixture for each tank.

Calculation for the actual volume of each tank. The heights of the tanks were 5m and the diameters were 2.5m. The tank was cylindrical with a conical drain at the bottom.

Using figure 5.61 in Ulrich and a volume of 24.9m3 the capital cost was estimated at 750$ per tank. The tank was made of carbon steel giving a FBM of 2.1.

Calculation for the current price using 528.7 as the current CE index value

82

Appendix B

Economics

Please Refer to Disc for More Details

83

84

85

86

87

88

89

90

91

92

93

94

95

96

97

98

Appendix C

HAZOP, MSDS,

DOW FIRE and ExplosionHAZOP/Safety Considerations1: Identify Hazards by considering the following Process Parameters:

i) Flow: - design for a no flow situation (pump cavitation/failure) and purchase (2)

secondary pumps for the biodiesel system - Flowrate to reactor is relatively low, and should not be considered hazardous

ii) Time:

-Because flowrates are low and residence time in the reactor is short, overflow of the tank is not a concern

iii) Frequency: - Consider frequency of loading new catalyst into the reactor or how often the

catalyst need to be recharged with methanol (further research needs to be done on the reusability of the catalyst)

iv) Mixing:

- N/A

v) Pressure: - Design for minimal pressure drop in the reactor across the packed bed.

Reactor operates at 600 psi and could be considered hazardous.

vi) Composition: - Methanol fed to the reactor in a 1:18 oil to alcohol ratio is considered

carcinogenic and caution should be taken if pump fails, catalyst needs to be replaced or recharged.

99

vii) Viscosity: - liquid is used, no slurries or solids, no chance of freezing. Product liquid of

high viscosity, thus pipe fowling should be considered to prevent build-up of pressure

viii) Temperature:

- Reactor operates at high temperature of 200 °C and thus stainless steel will be utilized

ix) pH:

- N/A

x) Separation: - Occurs in the gravity settling tanks after the reactor

xi) Level: - Flowrates and residence time are relatively low, therefore overflow is not of

concern

xii) Speed: - Pumps have moving parts, consider operator maintenance manuals for the

pump attached to the reactor

xiii) Information: - Documentation such as operating manuals, equipment specifications, design

criteria and MSDS should be supplied to new users to communicate details of plant

xiv) Reaction: - Reaction is mild and its exothermic properties are not of concern. All

components of the reaction should be kept away from oxidizers. Refer to MSDS chart below for more information.

xv) Operation: - N/A

100

General Operating Problems:

xvi) Desired conversion is not achieved.: - If the catalysts activity degrades, then the necessary conversion of 95% will

not be obtained. - Consider purging the reactor with methanol and hexane to recharge the

catalyst (methanol to remove polar particles from the catalyst and hexane to remove non-polar particles) or to replace the catalyst since it is silica based and low-cost

xvii) Pressure build-up:

- Because the reactor is operating at a high pressure in order to keep the methanol in it’s liquid phase at 200°C, pressure drop and build-up is of main concern

- Pressure drop should be carefully monitored as well as the pumps performance

xviii) Piping Design2:

- High viscosity may cause fowling of pipes - “advisable to:

1) Avoid the use of screwed fittings whenever practical. 2) Use welded fittings with long radius ells; avoid tees when possible.

101

Plant Safety1:

i) Safe Design: See HAZOP above

ii) Pollution Prevention: - When catalyst is replaced, proper disposal methods will be required. - If spills occurred, MSDS should be consulted for proper clean up techniques

iii) Lifecycle Analysis of Products: - Literature has reported a positive life cycle for the production of biodiesel

(1:3). Further research should be done to establish life cycle for this particular biodiesel process

iv) Inherently Safe Design:

Goal: to eliminate all hazards in the process using the 10 concepts that follow:

1. Intensification: Use very small amounts of hazardous material so that if there is a leak the hazard will be small.

- Our plant is of small capacity, and few reactants are of concern

2. Substitution: Replace hazardous materials with less hazardous ones.

- Our process has eliminated the need to dispose of contaminated water

3. Attenuation: Use hazardous material under the least hazardous conditions.

- N/A

4. Limitation: Limit the effects of failures by equipment design or by change in condition rather than adding on protective equipment.

- N/A 5. Simplification:

Simpler plants provide fewer opportunities for error and less equipment that can fail.

- Design based on using less equipment 6. Knock on Effect:

Design so that a domino effect doesn’t happen. - N/A. There is no inherent domino effect that could occur in the process.

7. Avoid Incorrect Assembly: - Since this is a grassroots plant, assembly will be required

102

8. Status Clear: It should be possible to see, at a glance, if valves are open or shut, if levels are ok, if correctly assembled.

- Process controls will be used on the reactor, mixers, and settling tanks 9. Control:

Control systems should be in place. - Process controls will be used on the reactor, mixers, and settling tanks 10. Survival:

If a hazard occurs personnel should be protected. - Determine a fire, explosion, and emergency evacuation plan

103

Process Safety Management System:

- Employee safety training which includes: o Personal Protective Equipment(PPE) for hearing protection, protective

clothing, eye protection, and proper footwear (steel toed CSA approved)

o Hazardous materials (eg. METHANOL) used on-site, how and where to locate the MSDS sheets on the materials and on-site via National Fire Protection Agency (NFPA) signs

- Documentation such as operating manuals, equipment specifications, design

criteria will be given to new users to communicate details of plant

- Employee training in fire, explosion, and emergency evacuation plans

- Maintenance plans for equipment being installed

104

Chemical Hazard Information and MSDS: Definitions: LD-50/LC-50: Lethal Dose 50. The dose that kills half (50%) of the animals tested. LEL/UEL: Lower/Upper Explosive Limit. Is the limiting/maximum concentration (in air) that is needed for the gas to ignite and explode. TLV: Threshold Limit Value. The reasonable level to which a worker can be exposed without adverse health effects. IDLH: Immediately Dangerous to Life or Health. The exposure to airborne contaminants that is “likely to cause death or immediate or delayed permanent adverse health effects or prevent escape from such an environment.”

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

Appendix D

Simulation Results

Please Refer to Disc for More Information

139

Appendix E

Rigorous Calculations For Leaching and Stripping Units

Please Refer to Disc for More information

140

141

142

143

144

145

146