Professional Documents
Culture Documents
Signature
: ................................................
th
May 2006
: 5...............................................
May 2006
I declare that this project report entitled Biodegradation of Oil and Grease in Upflow
Anaerobic Sludge Blanket Reactor for Palm Oil Mill Effluent Treatment is the result of
my own research except as cited in the references. The project report has not been
accepted for any degree and is not concurrently submitted in candidature of any other
degree.
Signature : ....................................................
Name
Date
ACKNOWLEDGEMENT
In preparing this project report, I was in contact with many people, researchers,
academicians and practitioners. They have contributed towards my understanding and
thoughts. In particular, I wish to express my sincere appreciation to my main project
report supervisor, Professor Ir. Dr. Zaini Ujang, for encouragement, guidance, critics and
friendship. Without his continued support and interest, this project report would not have
been the same as presented here.
I am also indebted to Universiti Teknologi Malaysia (UTM) for provide facilities
for my master project. Librarians at UTM, also deserve special thanks for their assistance
in supplying the relevant literatures.
My fellow postgraduate students should also be recognized for their support. My
sincere appreciation also extends to all my colleagues and others who have provided
assistance at various occasions. Their views and tips are useful indeed. Unfortunately, it is
not possible to list all of them in this limited space. I am grateful to all my family
members.
ABSTRACT
ABSTRAK
Pencerna lapisan enap cemar anaerobik alir-naik (UASB) adalah satu sistem di
mana bahan pemula akan bergerak melalui satu lapisan kembangan enap cemar
terpendam yang mengandungi biojisim yang tinggi kepekatannya. Biojisim yang ada
dalam pencerna boleh wujud dalam bentuk butiran atau gumpalan. Kebanyakan
pencernaan bahan pemula berlaku dalam biojisim terpendam dan baki bahan pemula
bergerak melalui satu lapisan yang kurang tumpat, iaitu lapisan kembangan enap cemar.
Pencerna lapisan enap cemar anaerobik alir-naik adalah salah satu pencerna anaerobik
berkadar tinggi yang mampu membiodegradasi minyak dan lemak (O&G) yang
terkandung dalam effluen kilang minyak sawit (POME). Kajian ini bertujuan untuk
mengkaji biodegradasi bagi minyak dan lemak yang terkandung dalam effluen kilang
minyak sawit mentah dengan menggunakan pencerna lapisan enap cemar anaerobik alirnaik. Objektif-objektif bagi kajian ini adalah menentukan kecekapan biodegradasi bagi
minyak dan lemak dan pengurangan permintaan oksigen secara kimia (COD) dalam
rawatan effluen kilang minyak sawit dan juga menaksir prestasi pencerna berhubung
dengan pelbagai kawalan parameter-parameter operasi. Ciri-ciri bagi 5 sampel berlainan
effluen kilang minyak sawit mentah dari Bukit Besar, Kulai ditentukan. Heksana
digunakan sebagai pelarut untuk pengekstrakan minyak dan lemak dengan menggunakan
corong pemisah. Eksperimen-eksperimen biodegradasi minyak dan lemak dan
pengurangan permintaan oksigen secara kimia direka secara faktoria penuh. Faktor-faktor
yang dipilih adalah masa penahanan hidrolik (HRT), kekuatan influen berdasarkan
permintaan oksigen secara kimia and pH influen. Keadaan mantap bagi pencerna tersebut
dicapai selepas 26 hari penyesuaian dengan pengurangan permintaan oksigen secara
kimia malar pada 62%. Seterusnya, eksperimen-eksperimen yang direka mula dijalankan
dan peratusan bagi biodegradasi minyak dan lemak serta pengurangan permintaan
oksigen secara kimia sebagai reaksi-reaksinya. Kombinasi parameter-parameter operasi
yang terbaik adalah HRT 12.9 jam, kekuatan influen 5,000 mg/L COD dan pH 7 di mana
ia berjaya mengurangkan 62% COD dan sebanyak 63% minyak dan lemak
terbiodegradasi. Analisis secara statistik digunakan untuk mengkaji pretasi pencerna
berhubung dengan pelbagai kawalan parameter-parameter operasi. Faktor-faktor utama
yang mempunyai kesan nyata dan penting kepada biodegradasi minyak dan lemak dan
pengurangan permintaan oksigen secara kimia telah dikenalpasti secara statistik.
TABLE OF CONTENTS
CHAPTER
TITLE
PAGE
DECLARATION
ii
DEDICATION
iii
ACKNOWLEDGEMENT
iv
ABSTRACT
ABSTRAK
vi
TABLE OF CONTENTS
vii
LIST OF TABLES
LIST OF FIGURES
xii
LIST OF ABBREVIATIONS
xiv
LIST OF SYMBOLS
xvii
INTRODUCTION
1.1
1.2
1.3
1.4
Industry
11
1.5
Problem Statements
12
1.6
Scopes of Study
13
1.7
Objectives of Study
13
LITERATURE REVIEWS
2.1
14
2.2
17
2.3
2.4
2.2.1
18
2.2.2
21
23
2.3.1
Process Description
24
2.3.2
Microbiology
26
2.3.3
30
2.3.4
Growth Kinetics
31
2.3.5
32
2.3.6
34
35
2.4.1
Design Considerations
37
2.4.2
41
2.4.3
42
RESEARCH METHODOLOGY
3.1
49
3.2
51
3.3
52
3.4
52
3.5
Design of Experiments
53
55
4.2
56
4.3
Design of Experiments
59
4.3.1
61
4.3.2
65
DISCUSSION OF RESULTS
5.1
70
5.2
72
5.3
74
Conclusion
76
6.2
Recommendations
76
REFERENCES
78
APPENDICES
Appendix A
81
Appendix B
84
Appendix C
85
LIST OF TABLES
TABLES NO.
1.1
TITLE
PAGE
1.2
1.3
2.1
15
2.2
16
2.3
17
2.4
19
2.5
19
2.6
2.7
33
2.8
38
2.9
39
2.10
39
40
2.11
41
3.1
49
3.2
53
3.3
3.4
3.5
53
54
54
4.1
55
4.2
57
4.3
59
4.4
59
4.5
60
4.6
61
4.7
62
4.8
63
4.9
66
4.10
66
4.11
67
LIST OF FIGURES
FIGURE NO.
1.1
TITLE
The top four palm kernel oil production and exports
countries.
1.2
PAGE
3
1.3
1.4
10
2.1
2.2
2.3
25
26
28
2.4
43
2.5
46
2.6
2.7
47
47
2.8
48
3.1
50
3.2
51
3.3
51
4.1
4.2
period.
57
58
4.3
61
4.4
64
4.5
65
4.6
4.7
biodegradation.
68
69
LIST OF ABBREVIATIONS
(CH3)3N
Methylamine
2D
Two dimensions
ABR
AMBR
Am-N
Ammoniacal nitrogen
BOD
Carbon
Ca(OH)2
Calcium hydroxide
Ca5OH(PO4)3 -
CaCO3
Calcium carbonate
CB
Cocoa butter
CH3COOH
Acetic acid
CH3OH
Methanol
CH4
Methane
CNO
Coconut oil
CO
Carbon monoxide
CO2
Carbon dioxide
COD
CPO
DS
Dissolved solids
ECP
Extracellular polymer
EQA
FAD
Fe
Ferum
FELCRA
FELDA
FeS
Ferrous sulfide
FFB
GSS
Gas-solid separator
H2
Hydrogen
H2O
Water
HA
Homoacetogen
HCOOH
Formic acid
HPKS 35
HRT
IV
Iodine value
MARDI
Mn
Manganese
MPOB
NAD
NaHCO3
Natrium bicarbonate
NaOH
Natrium Hydroxide
NBD
NH3
Ammonia
NH4(HCO3)
Ammonium bicarbonate
NO3-
Nitrate
NRB
O&G
O2
Oxygen
OHPA
PKO
PKS
PO
Palm oil
POME
PORIM
PORLA
R&D
RBD
RISDA
RM
Ringgit Malaysia
sCOD
Sdn. Bhd.
Sendirian Berhad
SFC
SRB
SRT
SS
Suspended solids
SVI
TN
Total nitrogen
TPAD
TVS
UASB
VFA
LIST OF SYMBOLS
Percent
<
Less than
>
More than
atm
Atmosphere
cm
Centimeter
cm2
Centimeter square
Day
Din
Inner diameter
Gram
g/L.d
Height
hr
Hour
Kd
kg/m3.d
Liter
Meter
m2
Meter square
m3
Cubic meter
m3/d
mg/L
mL/min
mm
Millimeter
Mp
Melting point
Degree of Celsius
Influent flowrate
Tonne
Vr
Reactor volume
Vw
Working volume
Synthesis yield
CHAPTER 1
INTRODUCTION
1.1
original palms introduced from West Africa to the Bogor Botanical Gardens, Indonesia in
1848, their seeds soon arrived on Malaysian shores in 1871 (Basiron and Chan, 2004).
Over the next four decades, the rubber companies in Malaysia saw their planters learning
how to grow the crop in the country. The R&D undertaken soon showed the potential of
the new crop. Following this effort, the first commercial planting was done in 1911 at
Tenammaran Estate, Kuala Selangor. There was the success of the crop that the area
expanded quickly, the most rapid increases occurring during the 1930s, 1970s and 1980s.
The growth in area during the various decades of the last century in Malaysia is shown in
Table 1.1. At the end of 2000, the area stood at 3.376 million hectares, producing 10.842
million tonnes of palm oil, 3.162 million tonnes of palm kernel, 1.384 million tonnes of
palm kernel oil and 1.639 million tonnes of palm kernel meal.
Table 1.1: Area of oil palm planting and growth in the decades of the last century
(Basiron and Chan, 2004).
Years in decades
1870-1910
1920
1930
1940
1950
1960
1970
1980
1990
2000
Hectares
<350
400
20 600
31 400
38 800
54 638
261 199
1 023 306
2 029 464
3 376 664
% Growth
14.2
5050.0
52.4
23.5
40.8
378.0
291.8
98.3
66.3
In the early 1960s, the returns from oil palm were found to be better than rubber
and most of the plantation companies soon had a mix of both crops as their core business.
It was Tun Abdul Razak Hussein, the then Deputy Prime Minister of Malaysia, who
called for greater diversification into oil palm. With diminishing returns from the then
two major commodities of the country, tin and rubber, oil palm should be used as the
vehicle to eradicate rural poverty. The governments three rural development agencies,
Federal Land Development Authority (FELDA), Federal Land Consolidation and
Rehabilitation Authority (FELCRA) and Rubber Industry Smallholders Development
Authority (RISDA) were responsible for planting oil palm with large areas of land that
were rehabilitated or newly opened (Basiron and Chan, 2004).
Landless people were placed as settlers in the newly opened land schemes.
Malaysian government provided them housing and infrastructure including community
halls, schools, health clinics, shops and roads. Initially, the government supported their
livelihoods until the oil palm matured when the income from the crop was sufficient to
pay off their loans. In doing so, the government was able to alleviate rural poverty using
the oil palm as the vehicle to do so.
Initial R&D into the crop was carried out by the government sector, the
Department of Agriculture. Together with the private research companies of the major
plantation groups, the work included collecting breeding materials and experimentation in
breeding, agronomy and palm oil chemistry. It was in 1969 when the Malaysian
Agricultural Research Development Institute (MARDI) was established that the mandate
for oil palm research was taken over from the Department of Agriculture. The task was
later handed to the Palm Oil Research Institute of Malaysia (PORIM) following its
establishment in 1979. On 1 May 2000, PORIM was merged with the Palm Oil
Registration and Licensing Authority (PORLA) to form the Malaysian Palm Oil Board
(MPOB) (Basiron and Chan, 2004). The mission of MPOB is to support the well-being
of the oil palm industry in Malaysia in all aspects of its activities through research,
development and services.
The oil palm only grows well in tropical climates and so all the palm kernel
producing countries are in Southeast Asia, Sub-Saharan Africa and South America
(Pantzaris and Ahmad, 2001). Figure 1.1 shows the production and exports of the top
four producing countries. The largest producer by far is Malaysia, which currently
accounts for more than 50% of world production. While two countries, Malaysia and
Indonesia together, account for about 80% of production and 88% of exports. No other
country produces more than 7% or exports more than 3% of the world total.
Figure 1.1: The top four palm kernel oil production and exports countries (Pantzaris and
Ahmad, 2001).
Until the mid 1970s, Nigeria was the worlds largest producer of palm kernels
while Europe did most of the crushing and was effectively the worlds largest palm kernel
oil producer. But now all the crushing is done in the producing countries and Europe
does no palm kernel crushing at all. From 1977, Malaysia overtook both Nigeria and
Europe to become worlds biggest producer of palm kernels and of palm kernel oil.
However, in the last few years, her oleochemical industry has been absorbing very large
and increasing quantities of the oil and her lead in exports have been reduced. In fact,
Indonesias exports were higher than Malaysias in 2000 (Pantzaris and Ahmad, 2001).
The oil palm industry worldwide has provided the fastest increase in global oils
and fats supplies over the last four decades. World palm oil production increased 20-fold
from a mere 1.2 million tonnes in 1962 to 25.0 million tonnes in 2002 (Basiron et al.,
2004). The share of palm oil production in the world oils and fats complex has increased
markedly by five-fold from 4% in 1962 to 20.8% in 2002, as compared to the only twofold increase experienced by soybean oil during the same period. The spiky increase in
palm oil output was mostly triggered by continued worldwide expansion of the oil palm
planted area and the mature area coming into production as well as growing world
demand for vegetable oils as in Table 1.2.
1962
(000 t)
30 779
1 234
3 432
1 163
2 294
12 040
%
Share
4.0
11.2
7.5
3.4
39.1
2002
(000 t)
120 477
25 034
29 748
13 326
7 611
22 588
%
Share
20.8
24.7
11.1
6.3
18.7
40yrs Average
Growth p.a. (%)
3.5
7.8
5.5
6.3
3.0
1.6
The status of palm oil as it is today in the world market is without doubt due to the
significant contribution by the Malaysian palm oil industry (Basiron et al., 2004). Both
Malaysia and Indonesia continue to remain the largest producers of palm oil, accounting
for 84% of the world production in 2002. In fact, the country has become a role model
for many other palm oil producing countries in their plans to urge economic development
in the agricultural sector as well as to gain foreign exchange through exports of surplus
The raw materials used in the manufacture of oleochemicals were mainly tallow,
coconut oil or palm and palm kernel oils. Malaysia being the worlds largest producer of
palm oil and palm kernel oil is undeniably in a particularly favourable position to become
a major supplier of raw materials for both the local and overseas oleochemical industries
(Ooi and Yeong, 2000). The oleochemicals that produced from palm oil and its products
are widely used in lubricants, plastics, resins, soaps, surfactants, emulsifiers, cosmetics,
toiletries and textile chemicals.
Over the last three decades, the Malaysian palm oil industry has grown to become
an important agricultural based industry. Malaysian palm oil accounted for about 52% of
the world palm oil outputs and this industry generated RM 13 billion in export earnings
for the country (Ahmad et al., 2005). The palm oil industry faces the challenge of
balancing the environmental protection, its economic viability and sustainable
development. There is an urgent need to find a way to preserve the environment while
keeping the economy growing (Ahmad et al., 2003).
1.2
recent years as the demand for vegetable oils increases (Borja and Banks, 1994). The
female bunch bears about 2,500 to 3,000 fruits borne on 100 to 120 spikelets attached to a
peduncle from the axil of a frond. The fruits produce two main products, palm oil from
the outer mesocarp and palm kernel oil from the kernel within the nut (Basiron and Chan,
2004). This tree is generally believed to have originated in the jungle forests of East
Africa and there is some evidence that palm oil was used in Egypt at the time of the
Pharaohs, some 5000 years ago, but now its cultivation is confined mostly to Southeast
Asia. The variety cultivated in nearly all the worlds plantations is the hybrid Tenera, the
cross between Dura and Pisifera, which gives the highest yield of oil per hectare of any
crop (Pantzaris and Ahmad, 2001).
Generally, the oil palms in Southeast Asia yield about 4 tonnes of palm oil, 0.5
tonnes of palm kernel oil, and 0.5 tonnes of palm kernel meal, with the income equivalent
to more than 4.5 tonnes of oil. Nearly for every 8 tonnes of crude palm oil produced at
the mill, about 1 tonnes of palm kernel oil is produced. The palm fruit looks like a plum.
The outer fleshy mesocarp gives the palm oil, while the kernel, which is inside a hard
shell, gives the palm kernel oil and it is rather strange that the two oils from the same fruit
are entirely different in fatty acid composition and properties. Unfortunately, the two oils
had often been confused by nutritionists in earlier days.
In palm oil, most of the fatty acids are C16 and higher, while in palm kernel oil,
they are C14 and lower. Palm oil has iodine value (IV) 50 minimum, while palm kernel
oil has 21 maximum. Semi-solid in temperate climates, palm kernel oil can be
fractionated into solid and liquid fractions known as stearin and olein respectively. These
are then physically refined, bleached and deodorized or chemically neutralized, bleached
and deodorized to give the RBD and NBD grades used in the food industry (Pantzaris and
Ahmad, 2001). The process of fractionation can be carried out either before or after the
refining, according to conditions.
The major fatty acids in palm kernel oil are C12 (lauric acid) about 48%, C14
(myristic acid) about 16% and C18:1 (oleic acid) about 15%. No other fatty acid is
present at more than 10% and it is this heavy preponderance of lauric acid, which gives
palm kernel oil and coconut oil, their sharp melting properties, meaning hardness at room
temperature combined with a low melting point (Pantzaris and Ahmad, 2001). This is the
outstanding property of lauric oils, which determines their use in the edible field and
justifies their usually higher price compared with most other oils. Because of their low
unsaturation, the lauric oils are also very stable to oxidation. Table 1.3 shows the fatty
acid composition of palm kernel oil, its similarity to coconut oil and their differences
from palm oil, the co-product of palm kernel oil and typical non-lauric fat. Even after full
hydrogenation, the melting point of palm kernel oil does not rise much above mouth
temperature and fractionation gives a stearin which is even sharper melting.
Sharp melting fats leave a clean, cool, non-greasy sensation on the palate,
impossible to match by any of the common non-lauric oils. Figure 1.2 shows the melting
behavior in terms of solid fat content (SFC) values of palm kernel oil (PKO), palm kernel
stearin (PKS) and hydrogenated palm kernel stearin of melting point 35C (HPKS 35),
together with cocoa butter (CB) and palm oil (PO) for comparison.
Table 1.3: Percentage of fatty acid composition in palm kernel, coconut and palm oil
(Pantzaris and Ahmad, 2001).
Fatty acids
C6
C8
C10
C12
C14
C16
C18
C18:1
C18:2
Others
Notes:
Figure 1.2: Solid fat content of cocoa butter, palm kernel oil product and palm oil
(Pantzaris and Ahmad, 2001).
Palm oil contains about 1% minor components. The major constituents are
carotenoids, vitamin E and sterols (Basiron and Chan, 2004). The carotene concentration
is around 500 to 700 ppm. Carotene has been concentrated from palm oil successfully.
The concentrate is rich in pro-vitamin A which is normally destroyed during processing.
The major carotenes in the carotenoid concentrate are alpha and beta-carotenes and they
can be diluted to various concentrations, from 1% to 30%. The vitamin E content in palm
oil is unique in that it is about 600 to 1,000 ppm. It is present as tocotrienols (70%) rather
than tocopherols (30%). It confers on the oil a natural stability against oxidation and a
longer shelf-life as well as a potent ability to reduce low density lipoprotein-cholesterol
and anti-cancer properties. Palm oil also contains 250 to 620 ppm sterols. Beta-sitosterol
is the major constituent at 60%. It is potentially hypocholesterolemic.
1.3
The fresh fruit bunches (FFB) are harvested in bunches and sent to the mill for processing
as show in Figure 1.3. Each FFB consists of hundreds of fruitlets each containing a nut
surrounded by a bright orange pericarp which contains the palm oil (Borja and Banks,
1994). These bunches are steam sterilized at a pressure of 3 bar where the fruits soften
and are easily detached from the stalk. These detached fruits are further softened with
steam in digesters. The digester mash is then passed to the screw press where oil together
with the juice from the fruits is expressed. The crude oil slurry which is expressed may
contain approximately 48% oil, 45% water and 7% solids (Chow and Ho, 2000). Some of
the water in this slurry is actually steam condensate from the sterilization, digestion and
screw pressing where steam is injected into the respective machinery to maintain the high
temperature required throughout the milling process.
Hot water is further added to the crude oil slurry to reduce the viscosity so that the
oil will cream to the surface which assists in its separation in large clarification tanks
(Borja and Banks, 1994). The underflow from the lower section of the clarification tank
is centrifuged to remove as much of the heavier phase consisting of solids and water.
This watery phase or sludge is discharged and any oil found here constitutes oil loss as it
is discharged as effluent. The lighter phase from the centrifuge, which consists of oil and
water, is recycled to the clarification tank. The creamed palm oil from the surface of the
clarification tank is then skimmed and further purified, dried and sold as crude palm oil
(CPO) to the refinery for further processing.
Figure 1.3: Flow diagram of palm oil extraction (Chow and Ho, 2000).
In the refinery, the CPO is processed to remove most of the undesirable impurities
thus making the oil bland, colourless and chemically stable according to trade
specifications and consumer requirements. There are basically two types of refining
practiced by the Malaysian palm oil refiners, chemical and physical refining (Chow and
Ho, 2000). The two processes differ in treatment of the oil and result in differently
labeled oils as illustrated in Figure 1.4. In the mill, as the CPO is extracted there is no
continuous on-line monitoring of quality but the impurities present are only of botanical
origin from the palm fruits. In the final quality assessment when sold to the refineries,
only certain contractual specifications are measured. They are free fatty acid, moisture,
peroxide value, and impurities which determines the degree of oxidation.
Correspondingly, in the trading of refined palm oil, the same sets of contractual
specifications are required with the additional requirement of colour. These parameters
are used to assess not only the initial quality of CPO but also the amount of bleaching
earth required which is one of the major costs incurred in CPO refining. No form of
continuous monitoring is known in palm oil refining.
Figure 1.4: Refining processes for crude palm oil (Chow and Ho, 2000).
1.4
natural forests to monoculture are primarily ecological. There are also the environmental
implications and typical environmental problems associated with plantation agriculture
such as soil erosion and loss of soil fertility during land preparation, water pollution due
to application of fertilizers and pesticides, and agricultural runoff and others. However, it
is in the processing of the oil palm crop or fresh fruit bunch (extraction of crude palm oil)
that this agro-industry was notable in the 60s and 70s for its adverse impact of extensive
pollution of the countrys surface water (Department of Environment, 1999).
Palm oil mills use the palm fiber and shell as solid boiler fuel to co-generate
needed steam and electricity. In the past, palm oil mills also typically employed an
incinerator to burn the empty bunches and recover the residual potash for use as fertilizer
in the plantation. Poor control of the air emissions from these facilities often caused
localized problems of air pollution (Department of Environment, 1999). However, the oil
palm industry is presently using the empty fruit bunches waste for mulching and POME
as fertilizer. During replanting, the trunks and fronds are chipped and left in the interrows as mulch under the zero-burn practice (Basiron and Chan, 2004).
Palm oil can be burnt directly as boiler fuel or as diesel for power generation and
vehicle propulsion. Besides, palm oil can be emulsified to palm diesel or methyl diesel.
The biomass from the mill, such as EFB, fiber and shell, can be used for electricity
generation. More than 10 out of the countrys 360 mills are applying to supply electricity
to Tenaga Nasional Berhad. New technologies are now available to harness the biogas
from effluent ponds for power generation. The total value of biogas energy available
from the mills is estimated to be RM 1 billion (Basiron and Chan, 2004). It is also
estimated that if all the biogas is used for the mill operation, then all the fiber and shell
can be freed for generating electricity for sale.
1.5
Problem Statements
The process to extract the oil requires significantly large quantities of water for
steam sterilizing the palm fruit bunches and clarifying the extracted oil. It is estimated
that for 1 tonne of crude palm oil produced, 5 to 7.5 tonnes of water are required, and
more than 50% of the water will end up as palm oil mill effluent (Ahmad et al., 2003).
Thus, while enjoying a most profitable commodity, the adverse environmental impact
from the palm oil industry cannot be ignored.
The oil and grease (O&G) content of POME is an important consideration in the
handling and treatment. O&G are singled out for special attention as their poor solubility
in water. Oil in wastewaters has to be removed in order to prevent interfaces in water
treatment units, reduce fouling in process equipment, avoid problems in biological
treatment stages and comply with water discharge requirements (Ahmad et al., 2005).
Hence the removal of residue oil from process or waste effluent becomes environmentally
important.
POME is an important source of inland water pollution when released into local
rivers or lakes without treatment. With increased cultivation and production of palm oil
in the region, the disposal of the processing waste is becoming a major problem that must
be appropriately addressed. Thus, the challenge of balancing the POME into a more
environmental friendly waste requires a sound and efficient treatment and disposal
approach.
1.6
Scopes of Study
The laboratory scale upflow anaerobic sludge blanket (UASB) reactor being used
was the existing acrylic material made cylindrical reactor without three phases separator.
The sludge was taken from the previous UASB microbial study. The sludge volume in
the reactor was controlled between 4.0 to 4.5 L levels. The raw POME samples were
collected only from the Felda Palm Industries Sdn. Bhd., Kilang Sawit Bukit Besar, in
Kulai, Johor Darul Takzim. Tap water was used for raw POME samples dilutions. Large
and bulky materials in the raw POME samples were removed before the samples dilutions.
1.7
Objectives of Study
This study aimed to evaluate the biodegradation of oil and grease content in raw
POME using UASB reactor. The objectives of this study were as follows:
i.
To determine the oil and grease (O&G) biodegradation for the treatment of raw
POME using UASB reactor,
ii.
iii.
CHAPTER 2
LITERATURE REVIEW
2.1
from the fresh fruit bunch. About 50% of the water results in POME, the other 50%
being lost as steam, mainly through sterilizer exhaust, piping leakages, as well as wash
waters. The POME comprises a combination of the wastewaters which are principally
generated and discharged from the following major processing operations (Department of
Environment, 1999):
i.
ii.
iii.
The typical quality characteristics of the individual wastewater streams from the 3
principal sources of generation are presented in Table 2.1. In most mills, all three
wastewater streams amounting to about 3 tonnes per tonne of palm oil produced, are
combined together resulting in a viscous brown liquid containing fine suspended solids
(Borja and Banks, 1994). A well managed palm oil mill with very good housekeeping
practices will generate about 2.5 m3 of POME per tonne of CPO produced; in term of
FFB this amounts to about 0.5 m3 of POME per tonne of FFB processed. However, the
national average is about 3.5 m3 of POME per tonne of CPO, or 0.7 m3 per tonne FFB.
This shows that much water can be saved through good milling and housekeeping
practices (Department of Environment, 1999).
Steriliser
Condensate
pH
5.0
Oil & Grease (O&G)
4,000
Biochemical Oxygen Demand
23,000
(BOD; 3 days, 30 C)
Chemical Oxygen Demand (COD)
47,000
Suspended Solids (SS)
5,000
Dissolved Solids (DS)
34,000
Ammoniacal Nitrogen (Am-N)
20
Total Nitrogen (TN)
500
Note: * All parameters units in mg/L except pH.
Oil
Clarification
Wastewater
4.5
7,000
Hydrocyclone
Wastewater
29,000
5,000
64,000
23,000
22,000
40
1,200
15,000
7,000
100
100
300
POME is a colloidal suspension of 95% to 96% water, 0.6% to 0.7% oil and 4% to
5% total solids including 2 to 4% suspended solids originating from the mixture of a
sterilizer condensate, separator sludge and hydrocyclone wastewater. It is thick brownish
in color liquid and discharged at temperature between 80 and 90 C and fairly acidic
(Ahmad et al., 2003). The typical quality characteristics of the raw combined POME are
illustrated in Table 2.2.
It can be seen that the BOD : COD ratio of raw POME is approximately 1 : 2,
which means that POME is considered to be suitably treated by biological processes. It
can also be seen that the range of the general parameters are large, which varies
depending on the efficiency and modes of the mill operations. However, the typical
BOD : N : P ratio of 139 : 4 : 1 indicates the limitation of nutrient, which is required for
bacterial growth and metabolic requirements of biomass, to obtain optimum biological
processes under aerobic conditions, which requires 100 : 5 : 1. Nutrient deficiency can
lead to increasing the population of filamentous bacteria (Lim and Ujang, 2004).
pH
Oil & Grease (O&G)
Biochemical Oxygen Demand
25,000
10,000-44,000
(BOD; 3 days, 30 C)
Chemical Oxygen Demand (COD)
50,000 16,000-100,000
Total Solids (TS)
40,500
11,500-79,000
Suspended Solids (SS)
18,000
5,000-54,000
Total Volatile Solids (TVS)
34,000
9,000-72,000
Ammoniacal Nitrogen (Am-N)
35
4-80
Total Nitrogen (TN)
750
80-1,400
Note: * All parameters units in mg/L except pH.
Magnesium
615
Calcium
Boron
Iron
Manganese
Copper
Zinc
440
7.6
47
2.0
0.9
2.3
POME is the fiber-free non-oil components obtained from the clarification zone of
oil mill. POME consists of various suspended components including cell walls,
organelles, short fibers, a spectrum of carbohydrates ranging from hemicellulose to
simple sugars, a range of nitrogenous compounds from proteins to amino-acids, free
organic acids and an assembly of minor organic and mineral constituents (Ugoji, 1997).
Table 2.3: Prevailing effluent discharge standard for crude palm oil mills (Laws of
Malaysia, 2003).
Parameters
Unit
Parameter Limits
(Second Schedule)
2.2
Remarks
100
*
*
400
50
150
200
5-9
45
Filtered sample
Filtered sample
Malaysia. POME contains about 4000 to 6000 mg/l of oil and grease (Ahmad et al.,
2005). The oil droplets of POME can be found in two phases. They either suspend in the
supernatant or float on the upper layer of the suspension. The residue oil droplets in
POME were solvent extractable. The extract of the oil droplets consists of 84 wt%
neutral lipids and 16 wt% of complex lipids (6 wt% glygolipids and 10 wt%
phospholipids). The neutral lipids consist of 74.7% triglycerides, 8% diglycerides, 0.5%
monoglycerides and 0.8% free fatty acids. POME also distributes a high concentration of
surface active compounds like phospholipids (10 wt%) and glycolipids (6 wt%).
O&G in POME can be defined as residual oil that contained in the wastes that
generated from the process of palm oil extraction. The O&G in POME are produced
from combination of O&G in sterilizer effluent, hydrocyclone effluent, and centrifuge
effluent. Under the Environmental Quality (Prescribed Premises) (Crude Palm Oil) Order,
1977, and the Environmental Quality (Prescribed Premises) (Crude Palm Oil) Regulations,
1977, the concentration of O&G in POME for standard discharge limit by the Malaysian
Department of Environment is 50 mg/L (Laws of Malaysia, 2003).
Better understanding of fats and oils in palm oil is relatively important to the study
of biodegradation of O&G in POME. Thus, it will help to achieve the success in POME
treatment. In chemistry, fats, oils, and waxes are all esters. Fats and oils are esters of the
trihydroxy alcohol, glycerol, while waxes are esters of long chain monohydroxy alcohols.
All serve as food for humans as well as bacteria, since they can be hydrolyzed to the
corresponding fatty acids and alcohols.
2.2.1
triacylglycerols. Dietary triacylglycerols are composed of fatty acids that may vary in
their chain length, degree of unsaturation, isomeric orientation of double bonds and
position within the triacylglycerol molecule (Edem et al., 2003). The glycerides of fatty
acids that are liquid at ordinary temperatures are called oils and those that are solids are
called fats (Sawyer et al., 2003). Chemically, they are quite similar. The oils have a
predominance of short chain fatty acids or fatty acids with a considerable degree of
unsaturation, such as linoleic or linolenic.
The fatty acids in a given molecule of a glyceride may be all the same or they may
all be different. The principal acids composing the glycerides of fats and oils are shown
in Table 2.4. The relative amounts of the major fatty acids contained in various fats and
oils are shown in Table 2.5.
Palm oil, a triglyceride with a melting point 34.2 C contains approximately 44%
palmitic acid and is 97% digestible (Edem et al., 2003). This oil is one of the richest
sources of -carotene, is obtained from the fleshy orange-red mesocarp of the fruits.
Refined palm olein is the liquid fraction obtained from the refining, bleaching and
deodorization of the crude red palm oil. This fractionation brings about enrichment with
the monounsaturated oleic acid in addition to the concomitant reduction of palmitic acid,
the major saturated fatty acid.
Table 2.4: Acids of the fats and oils (Sawyer et al., 2003).
Name
Butyric
Caproic
Caprylic
Capric
Lauric
Myristic
Palmitic
Stearic
Arachidic
Behenic
Oleic
Erucic
Linoleic
Linolenic
Formula
C3H7COOH
C5H11COOH
C7H15COOH
C9H19COOH
C11H23COOH
C13H27COOH
C15H31COOH
C17H35COOH
C20H40O2
C22H44O2
C18H34O2
C22H42O2
C18H32O2
C18H30O2
Mp, C
-5.7
-3
16.3
31.9
43.2
53.9
63.1
69.6
76.5
81.5
13.4
34.7
-12
-11
Source
Butter
Butter, coconut oil
Palm oil, butter
Coconut oil
Coconut oil, spermaceti
Nutmeg, coconut oil
Palm oil, animal fats
Animal & vegetable fats, oils
Peanut oil
Ben oil
Animal & vegetable fats, oils
Rape oil, mustard oil
Cottonseed oil
Linseed oil
Table 2.5: Acid content of fats and oils in percent (Sawyer et al., 2003).
Name
Butter
Mutton tallow
Castor oil
Olive oil
Palm oil
Coconut oil
Peanut oil
Corn oil
Cottonseed oil
Linseed oil
Soybean oil
Tung oil
Oleic
Linoleic
Linolenic
Stearic
Myristic
Palmitic
Arachidic
27.4
36.0
9.0
84.4
38.4
5.0
60.6
43.4
33.2
5.0
32.0
14.9
4.3
3.0
4.6
10.7
1.0
21.6
39.1
39.4
48.5
49.3
-
34.1
2.2
-
11.4
30.5
3.0
2.3
4.2
3.0
4.9
3.3
1.9
4.2
1.3
22.6
4.6
Trace
1.1
18.5
0.3
-
22.6
24.6
6.9
41.1
7.5
6.3
7.3
19.1
6.5
4.1
0.1
3.3
0.4
0.6
0.7
-
Like all oils, triacylglycerols are the major constituents of palm oil. Over 95% of
palm oil consists of mixtures of triacylglycerols that are glycerol molecules, each
esterified with three fatty acids. During oil extraction from the mesocarp, the
hydrophobic triacylglycerols attract other fat or oil soluble cellular components (Sundram,
2004). These are the minor components of palm oil such as phosphatides, sterols,
pigments, tocopherols, tocotrienols and trace metals. Other components in palm oil are
the metabolites in the biosynthesis of triacylglycerols and products from lipolytic activity.
These include the monoglycerols, diglycerols and free fatty acids (FFAs). The fatty acids
are any of a class of aliphatic acids, such as palmitic (16:0), stearic (18:0) and oleic (18:1)
in animal and vegetable fats and oils. The major fatty acids in palm oil are myristic
(14:0), palmitic, stearic, oleic and linoleic (18:2).
Palm oil has saturated and unsaturated fatty acids in approximately equal amounts.
Most of the fatty acids are present as triacylglycerols. The different placement of fatty
acids and fatty acid types on the glycerol molecule produces a number of different
triacylglycerols. There are 7% to 10% of saturated triacylglycerols, predominantly
tripalmitin. The fully unsaturated triacylglycerols constitute 6% to 12% (Sundram, 2004).
The Sn-2 position has specificity for unsaturated fatty acids. Therefore, more than 85%
of the unsaturated fatty acids are located in the Sn-2 position of the glycerol molecule.
The triacylglycerols in palm oil partially define most of the physical characteristics of the
palm oil such as melting point and crystallization behavior.
Fats and oils undergo three types of chemicals reactions of interest, hydrolysis,
addition and oxidation. Since fats and oils are esters, they undergo hydrolysis with more
or less ease. The hydrolysis may be induced by chemical means, usually by treatment
with NaOH, or by bacterial enzymes that split the molecule into glycerol plus fatty acids.
Hydrolysis with the aid of NaOH is called saponification (Sawyer et al., 2003).
Hydrolysis by bacterial action may produce rancid fats or oils and renders them
unpalatable. Rancid butter and margarine are notorious for their bad odor.
The fats and oils containing unsaturated acids add chlorine at the double bonds, as
other unsaturated compounds do. This reaction is often slow because of the relative
insolubility of the compounds. It may represent a significant part of the chlorine demand
of some wastes, and chlorinated organics will be produced during chlorination, some of
which may be of health concern.
Oils that contain significant amounts of oleic and linoleic acids may be converted
to fats by the process of hydrogenation (Sawyer et al., 2003). In this process hydrogen is
caused to add at the double bonds, and saturated acids result. Thus, low priced oils such
as soybean and cottonseed can be converted into margarine, which is acceptable as human
food. Many cooking fats or shortenings are made in the same manner. The
hydrogenation can be controlled to produce any degree of hardness desired in the product.
The oils with appreciable amounts of linoleic and linolenic acids or other highly
unsaturated acids, such as linseed and tung oil, are known as drying oils. In contact with
the air, oxygen adds at the double bonds and forms a resin like material. The drying oils
are the major component in all oil based paints.
2.2.2
The first step is hydrolysis, with the production of glycerol and fatty acids. The
free fatty acids derived from the hydrolysis of fatty materials, the deaminization of amino
acids, carbohydrate fermentation, and omega oxidation undergo further breakdown by
oxidation. Oxidation is believed to occur at the beta carbon atom in accordance with
Knoops theory, sometimes called the beta oxidation theory (Sawyer et al., 2003).
According to this theory, oxidation proceeds in a series of steps. Coenzyme A is known
to be active in these transformations. Oxidation is accomplished by enzymatic hydrogen
and electron removal and water addition, which are facilitated by the electron carriers,
FAD and NAD. The pathway of beta oxidation to a fatty acid molecule is showed from
Reactions (2.1) to (2.5).
CH2
CH2
OH
+ HSCoA
enzyme
CH2
CH2
OH
CH2
SCoA
(2.1)
+ H 2O
Coenzyme A
carbon
CH2
SCoA
SCoA
+ FAD +
+ H2O
enzyme
enzyme
SCoA
+ NAD +
enzyme
OH
(2.2)
SCoA + FADH + H+
(2.3)
SCoA
SCoA
(2.4)
+ NADH + H+
H
O
SCoA + 2H2O
enzyme
OH
(2.5)
+ CH3COOH + HSCoA
In the ultimate step, rupture of the molecule occurs, with formation of one
molecule of acetic acid, and the original molecule of acid appears as a new acid derivative
with two less carbon atoms. Thus, by oxidation at the beta carbon atom, long chain fatty
acids are whittled into fragments consisting of acetic acid. During this oxidation four
hydrogen atoms, with corresponding electrons are removed for each acetic acid unit
produced. These electrons are contained in NADH and FADH. For example, hydrolysis
of palm oil may yield palmitic acid (C15H31COOH). Beta oxidation of this acid would
proceed by chopping off two carbons at a time in the form of acetic acid. Since the acid
has a total of 16 carbons, a total of 8 mol of acetic acid will be formed for each mole of
palmitic acid (Sawyer et al., 2003). By this stoichiometry, the net result would be same
as the Reaction (2.6) below:
(2.6)
Under aerobic conditions, the electrons now carried by FADH and NADH are
used to reduce molecular oxygen to water and produce energy. Under anaerobic
condition, however, it is not possible for the bacteria to rid themselves of the electrons in
this fashion, and another scheme must be used.
In aerobic systems, oxygen is the terminal electron acceptor and is reduced while
organic and inorganic electron donors are being oxidized. In the absence of oxygen, other
materials such as nitrate, Fe(III), Mn(IV), sulfate, and carbon dioxide may become
electron acceptors (Tchobanoglous et al., 2003). The use of sulfate and carbon dioxide
requires strictly anaerobic conditions. Nitrate can be used, however, by facultative
organisms living under intermediate conditions referred to as anoxic, which are
characterized by end products of carbon dioxide, water, and nitrogen gas. All are
inoffensive as opposed to products of methane and H2S formed under strictly anaerobic
conditions. Additionally, Fe(II) and Mn(II) formed from reduction of Fe(III) and Mn(IV)
represent potential water quality problems such as staining.
The reactions of sulfate reduction and methanogenesis can be carried out only by
highly specialized groups of anaerobic bacteria and cannot generally proceed in the
presence of O2, Fe(III), NO3-, or Mn(IV). Methanogenesis is the major process by which
wastes are stabilized in anaerobic digestion and is an important part of the stabilization
occurring in septic tank, lagoons, and sanitary landfills (Sawyer et al., 2003). The
methane gas released can be used as a fuel. About 70 percent of the methane resulting
from the complete methane fermentation of complex wastes results from fermentation of
acetic acid formed as an intermediate in the anaerobic fermentation of carbohydrates,
proteins, fats, and other organic compounds.
2.3
Anaerobic fermentation and oxidation processes are used primarily for the
treatment of waste sludge and high strength organic wastes. Nevertheless, applications
for dilute waste streams have also been demonstrated and are becoming more common.
Anaerobic fermentation processes are advantageous because of the lower biomass yield
and because energy, in the form of methane can be recovered from the biological
conversion of organic substrates.
For treating high strength industrial wastewater, anaerobic treatment has been
shown to provide a very cost effective alternative to aerobic processes with savings in
energy, nutrient addition, and reactor volume. Because the effluent quality in not as good
as that obtained with aerobic treatment, anaerobic treatment is commonly used as a pretreatment step prior to discharge to a municipal collection system or is followed by an
aerobic process.
2.3.1
Process Description
Three basic steps are involved in the overall anaerobic oxidation of a waste:
Figure 2.2: Carbon and hydrogen flow in anaerobic digestion process (Tchobanoglous et
al., 2003).
2.3.2
Microbiology
The anaerobic microbial conversion of organic substrates to methane is a complex
ii.
iii.
iv.
v.
vi.
vii.
viii.
ix.
Other
occur, one solution is to add iron at controlled amounts to form iron sulfide precipitate.
Sulfate reducing bacteria, obligate anaerobes of the domain bacteria, are morphologically
diverse, but share the common characteristic of being able to use sulfate as an electron
acceptor and are divided into one of two groups depending on whether they produce fatty
acids or use acetate (Tchobanoglous et al., 2003). Group I sulfate reducers can use a
diverse array of organic compounds as their electron donor, oxidizing them to acetate and
reducing sulfate to sulfide. A common genus found in anaerobic biochemical operations
is Desulfovibrio. Group II sulfate reducers oxidize fatty acids, particularly acetate to
carbon dioxide, while reducing sulfate to sulfide. A bacteria commonly found in this
group is in the genus Desulfobacter.
2.3.3
reactions defined as carbon dioxide, and methyl group type reactions are shown as
follows, involving the oxidation of hydrogen, formic acid, carbon monoxide, methanol,
methylamine, and acetate, respectively in Reactions (2.7) to (2.12).
4H2 + CO2
4HCOO- + 4H+
4CO + 2H2O
4CH3OH
4(CH3)3N + H2O
CH3COOH
CH4 + 2H2O
CH4 + 3CO2 + 2H2O
(2.7)
(2.8)
CH4 + 3CO2
(2.9)
(2.10)
(2.11)
(2.12)
A COD balance can be used to account for the changes in COD during
fermentation. Instead of oxygen accounting for the change in COD, the COD loss in the
anaerobic reactor is accounted for by the methane production. By stoichiometry the COD
equivalent of methane can be determined. The COD of methane is the amount of oxygen
needed to oxidize methane to carbon dioxide and water (Tchobanoglous et al., 2003).
CH4 + 2O2
CO2 + 2H2O
(2.13)
From the Reaction (2.13) above, the COD per mole of methane is 2(32 g O2/mole)
= 64 g O2/mole CH4. The volume of methane per mole at standard conditions (0 C and 1
atm) is 22.414 L, so the CH4 equivalent of COD converted under anaerobic conditions is
22.414/64 = 0.35 L CH4/g COD.
2.3.4
Growth Kinetics
In anaerobic processes two rate limiting concepts are important, the hydrolysis
conversion rate and the soluble substrate utilization rate for fermentation and
methanogenesis. The hydrolysis of colloidal and solid particles does not affect the
process operation and stability but does affect the total amount of solids converted. In
anaerobic digestion processes used for municipal waste sludge, greater than 30 days
detention time is needed to approach full conversion of solids (Tchobanoglous et al.,
2003). The soluble substrate utilization kinetics is of great concern to develop a stable
anaerobic process.
Because of the relatively low free energy change for anaerobic reactions, growth
yield coefficients are considerably lower than the corresponding values for aerobic
oxidation. Typical synthesis yield and endogenous decay coefficients for fermentation
and methanogenic anaerobic reactions are Y = 0.10 and 0.04 g VSS/ g VSS and kd = 0.04
and 0.02 g VSS/ g VSS.d, respectively.
The process is more stable when the volatile fatty acid (VFA) concentrations
approach a minimal level, which can be taken as an indication that a sufficient
methanogenic population exists and sufficient time is available to minimize hydrogen and
VFA concentrations. The rate limiting step is the conversion of VFAs by the
methanogenic organisms and not the fermentation of soluble substrates by the fermenting
bacteria. Thus, the methanogenic growth kinetics is of most interest in anaerobic process
designs. Appropriate system solid retention times (SRTs) are selected based on kinetics
and treatment goals. At 20, 25, and 35 C, the washout or SRTmin values for
methanogenesis are 7.8, 5.9, and 3.2 d, respectively (Tchobanoglous et al., 2003).
Methanogenesis is the rate limiting step due to the nature of the methanogens. In
the design of anaerobic digester units, due consideration must be given to this terminal
phase. The methanogens being a slow growing group must be retained within the system.
System optimization requires the highest possible SRT and therefore some means must be
found to recover the active methanogens from being lost from the system (Shaaban, 1990).
2.3.5
advantages over the conventional aerobic process. The rationale for and interest in the
use of anaerobic treatment processes can be explained by considering the advantages and
disadvantages of these processes. The principal advantages and disadvantages of
anaerobic treatment are listed in Table 2.6.
Advantages
Less energy required
Less biological sludge production
Fewer nutrients required
Methane production, a potential
energy source
Smaller reactor volume required
Elimination of off gas air
pollution
Rapid response to substrate
addition after long periods
without feeding
Disadvantages
Longer startup time to develop
necessary biomass inventory
May require alkalinity addition
May require further treatment
with an aerobic treatment process
to meet discharge requirements
Biological nitrogen and
phosphorus removal is not
possible
Much more sensitive to the
adverse effect of lower
temperatures on reaction rates
May be more susceptible to upsets
due to toxic substances
Potential for production of odors
and corrosive gases
ii.
iii.
Temperature of wastewater
iv.
v.
vi.
vii.
viii.
As we know, anaerobic digestion processes offer great potential for the treatment
of many types of wastewaters. For that reason, anaerobic treatment presently is accepted
as a grown up technology. However, in many other countries this unfortunately still is
not the case. The reasons that obstinate the reluctance are (Malina and Pohland, 1992):
i.
ii.
Lack of knowledge of the basis principles of the process in the established world
of sanitary engineering
iii.
iv.
v.
vi.
vii.
Acceptance problems for a technology which was not developed in their own
institute or company
2.3.6
challenges. With the present situation where there are some mills still failing to comply
with the standard discharge limit even after they have applied the available treatment
system (Ahmad et al., 2003).
Little research has been reported regarding the application of modem high rate
anaerobic digester technologies such as upflow or downflow filters, fluidized beds,
UASB systems or upflow floc digesters for the disposal of POME. In 1994, the use of
UASB reactors for treating POME has been reported, which revealed the potential of
smaller and more efficient reactors technology in POME treatment (Borja and Banks,
1994). In that research study, the start-up of the UASB reactor was monitored and the
treatment efficiency in terms of COD reduction was evaluated. Over 80% COD was
removed at loading rates up to 10.6 g/L.d COD, and at the highest influent concentration
(42.5 g/L) reactor instability was observed. Later in 1996, anaerobic treatment of POME
in a two-stage UASB system using a pair of upflow anaerobic reactors to evaluate the
effect of a staged treatment on sludge granulation (Borja et al., 1996). The results
showed that it is possible to split the anaerobic digestion process into two distinct stages
in a pair of UASB reactors and granulation was observed in both reactors but their
morphological characteristics were distinct.
2.4
systems the potential user obviously will prefer the most economical system, both in
terms of investment but particularly also in operation and maintenance. One of the most
notable developments in anaerobic treatment process technology was the UASB reactor
in the late 1970s in the Netherlands by Lettinga and his coworkers (Tchobanoglous et al.,
2003). Considering the present application of anaerobic treatment, apparently reactors
based on the UASB principle look most favourable. In the meantime the feasibility of the
UASB concept for treating mainly soluble wastewaters has been sufficiently
demonstrated at full scale, at demonstration scale, and at pilot plant scale (Malina and
Pohland, 1992).
The principal types of anaerobic sludge blanket processes include the original
UASB process and modification of the original design, the anaerobic baffled reactor
(ABR) and the anaerobic migrating blanket reactor (AMBR). Of these sludge blanket
processes, the UASB process is used most commonly, with over 500 installations treating
a wide range of industrial wastewaters (Tchobanoglous et al., 2003).
The influent wastewater is distributed at the bottom of the UASB reactor and
travels in an upflow mode through the sludge blanket. Critical elements of the UASB
reactor design are the influent distribution system, the gas-solid separator, and the effluent
withdrawal design. Modifications to the basic UASB design include adding a settling
tank or the use of packing material at the top of the reactor. Both modifications are
intended to provide better solids capture in the system and to prevent the loss of large
amounts of the UASB reactor solids due to process upsets or changes in the UASB sludge
blanket characteristics and density. The use of an external solids capture system to
prevent major losses of the system biomass is recommended.
The demand for high-rate anaerobic processes is still increasing and several labscale studies are being carried out on the application, design, and operation of UASB
reactors. Advantages for the UASB process are the high loadings and relatively low
detention times possible for anaerobic treatment and the elimination of the cost of
packing material. Limitations of the process are related to the wastewater that are high in
solids content or where their nature prevents the development of the dense granulated
sludge (Tchobanoglous et al., 2003).
2.4.1
Design Considerations
In the engineering design of anaerobic bioreactor, factors to consider include
loadings, SRT, mixing and solid-liquid separation (Shaaban, 1990). Organic and
volumetric loadings rates will decide on the suitable operational range in order to avoid
failures due to organic and hydraulic overloads. A minimum SRT of 10 days is required
for a stable operation but in most cases SRT well in excess of 10 days is provided to
ensure trouble free operation and better system performance. Mixing can form an integral
part of a digestion system and it helps to bring about uniformity within the reactor and
since either mechanical impeller type or gas/sludge recirculation requires a high energy
input, serious thought must be given at the preliminary design stage. Effective solidliquid separation is critical in the recovery and recirculation of active biomass to obtain
high SRT values and hence maximum process efficiency.
Wastewaters that contain substances that can adversely affect the sludge
granulation, cause foaming, or cause scum formation are of concern. Wastewaters with
higher concentrations of protein and/ or fats tend to create more of the above problems.
The fraction of particulate versus soluble COD is important in determining the design
loadings for UASB reactors as well as determining the applicability of the process
(Tchobanoglous et al., 2003). As the fraction of solids in the wastewater increases, the
ability to form a dense granulated sludge decreases. At a certain solids concentration
(greater than 6 g TSS/L) anaerobic digestion and anaerobic contact processes may be
more appropriate.
Wastewater
COD, mg/L
Fraction as
particulate
COD
Flocculent
sludge
1000-2000
0.10-0.30
0.30-0.60
0.60-1.00
0.10-0.30
0.30-0.60
0.60-1.00
0.10-0.30
0.30-0.60
0.60-1.00
0.10-0.30
0.30-0.60
0.60-1.00
2-4
2-4
na
3-5
4-8
4-8
4-6
5-7
6-8
5-8
na
na
2000-6000
6000-9000
9000-18000
2-4
2-4
na
3-5
2-6
2-6
4-6
3-7
3-8
4-6
3-7
3-7
8-12
8-14
na
12-18
12-24
na
15-20
15-24
na
15-24
na
na
Design HRT values are also given for the treatment of domestic wastewater in
Table 2.9 based on pilot plant experience. The HRT value needed is longer than that used
in aerobic processes for secondary treatment for BOD removal. In addition, an aerobic
polishing step would likely be needed. The economic benefits of energy savings and
lower sludge production would have to be sufficient to justify the higher capital costs for
liquid treatment with a UASB process.
Temperature, C
15
2-4
3
20
4-6
5
25
6-12
6
30
10-18
12
35
15-24
18
40
20-32
25
* Average sludge concentration is 25 g/L.
2-3
2-4
4-8
8-12
12-18
15-24
2
3
4
10
14
18
Table 2.9: Applicable hydraulic retention times for treatment of raw domestic wastewater
in a 4 m high UASB reactor (Lettinga and Hulshoff Pol, 1991).
Temperature, C
Average HRT, hr
16-19
22-26
>26
10-14
7-9
6-8
7-9
5-7
4-5
The upflow velocity based on the flowrate and reactor area, is a critical design
parameter that will affect the efficiency of upflow reactors. The upflow velocity affects
the sludge retention as it is based on the settling characteristics of sludge aggregates. The
upflow velocity should be high enough to provide good contact between substrate and
biomass. Higher upflow velocity is believed to facilitate the separation of gas bubbles
from the surface of biomass (Mahmoud et al., 2003).
v=
Q
A
Eq. (2.1)
Table 2.10: Upflow velocities and reactor heights recommended for UASB reactors
(Tchobanoglous et al., 2003).
Wastewater type
COD nearly 100% soluble
COD partially soluble
Domestic wastewater
1.00-3.00
1.00-1.25
0.80-1.00
1.5
1.0
0.7
Reactor height, m
Range
Typical
6-10
3-7
3-5
8
6
5
The main physical features requiring careful consideration are the feed inlet, gas
separation, gas collection, and effluent withdrawal (Tchobanoglous et al., 2003). The
inlet and gas separation designs are unique to the UASB reactor. The feed inlet must be
designed to provide uniform distribution and accomplish an optimal contact between
sludge and wastewater. Channelling of the wastewater through the sludge bed and
formation of dead zones in the reactor should be avoided (Lettinga and Hulshoff Pol,
1991). The avoidance of channelling is more critical for weaker wastewaters, as there
would be less gas production to help mix the sludge blanket.
A number of inlet feed pipes are used to direct flow to different areas of the
bottom of the UASB reactor from a common feed source. Access must be provided to
clean the pipes in the event of clogging (Tchobanoglous et al., 2003). Guidelines for
determining the area served by the individual inlet feed pipes as a function of the sludge
characteristics and organic loading are provided in Table 2.11. Particularly in treating
partially soluble wastewaters one should realize that clogging of the nozzles may
represent a serious problem (Lettinga and Hulshoff Pol, 1991). Every inlet system should
be easy to clean, as always after some time of operation inlet pipes tend to become
partially clogged. This will result in an uneven distribution of the wastewater over the
reactor bottom.
Table 2.11: Guidelines for sizing the area served by inlet feed pipes for UASB reactor
(Tchobanoglous et al., 2003).
Sludge type
Dense flocculent sludge,
> 40 kg TSS/m3
< 1.0
1-2
>2
0.5-1
1-2
2-3
< 1-2
>3
1-2
2-5
1-2
2-4
>4
0.5-1
0.5-2
>2
Taking into consideration the slow growth rate of many anaerobic microorganisms
particularly methanogenic, the main objectives of the efficient reactor design must be
high retention time of bacterial cells with very little loss of bacteria from the bioreactor.
The technological challenge to improve the anaerobic digestion lies in enhancing the
bacterial activity together with good mixing to ensure a high rate of contact between the
cells and their substrate (Faisal and Unno, 2001).
2.4.2
UASB reactor is crucial equipment for retaining as much viable sludge as possible. The
GSS device can be simply and inexpensively designed (Bae and Shin, 1998). Relatively a
lot of research has been spent in recent years for developing new and more exclusive
systems. The last word on the design of the GSS device certainly has not yet been said.
It is an area of continuous innovation (Lettinga and Hulshoff Pol, 1991).
The GSS is designed to collect the biogas, prevent washout of solids, encourage
separation of gas and solid particles, allow for solids to slide back into the sludge blanket
zone, and help improve effluent solids removal (Tchobanoglous et al., 2003). A series of
upside down V-shaped baffles is used next to effluent weirs to accomplish the above
objectives. Guidelines for the GSS design are summarized below:
i.
The slope of the settler bottom, the inclined wall of the gas collector should be
between 45 and 60
ii.
The surface area of the apertures between the gas collectors should not be smaller
than 15% to 20 % of the total reactor surface area
iii.
The height of the gas collector should be between 1.5 to 2 m at reactor heights of
5 to 7 m
iv.
v.
The overlap of the baffles installed beneath the apertures should be 100 to 200
mm to avoid upward flowing gas bubbles entering the settler compartment
vi.
Generally scum layer baffles should be installed in front of the effluent weirs
vii.
The diameter of the gas exhaust pipes should be sufficient to guarantee the easy
removal of the biogas from the gas collection cap, particularly in the case where
foaming occurs
viii.
In the upper part of the gas cap, antifoam spray nozzles should be installed in the
case where the treatment of the wastewater is accompanied by heavy foaming
The capability of the GSS device to separate biomass from rising gas bubbles can
be described by the difference of volatile suspended solid (VSS) concentration in the
blanket and effluent (Bae and Shin, 1998).
2.4.3
include the granulation process in the reactor, the characteristics of the wastewater to be
treated, the selection of inoculum material, the influence of nutrients and several other
environmental factors. Among these parameters, the granulation process is believed to be
the most critical one (Show et al., 2004).
The formation of anaerobic granular sludge can be considered as the major reason
of the successful introduction of the UASB reactor concept for anaerobic treatment of
industrial effluents. Figure 2.4 shows the anaerobic granules from the UASB reactor.
The granular sludge has superior settling characteristic, settling velocities of granular
sludge of approximately 60 m/hr are common, whereas the superficial upflow velocities
in UASB reactors are usually kept below 2 m/hr, in practice (Hulshoff Pol et al., 2004).
This allows an extreme uncoupling of the hydraulic retention time from the sludge age.
SRTs of over 200 days can be achieved at HRTs of only 6 hr.
Figure 2.4: Anaerobic granules from the UASB reactor (Hulshoff Pol et al., 2004).
The granular sludge performs high specific methanogenic activities. It could be
demonstrated that high volumetric loading rates of over 50 kg COD/m3.d could be well
accommodated under mesophilic conditions, with specific methanogenic activities of
more than 2 kg COD/kg VSS.d. Studies on the micromorphology of the granules
demonstrated that colonies of acetogenic bacteria are closely linked with micro-colonies
of hydrogenotrophic methanogenic archaea allowing an efficient interspecies hydrogen
transfer and as a result, high degradation rates.
The UASB process involves the anaerobic degradation of organic wastes using a
biomass which is not attached to a support medium but which aggregates, under
favourable conditions, to produce particles with good settlement characteristics. These
particles are known as granules, and their formation commonly termed granulation,
generally enhances the efficiency of the process by producing high biomass retention
times. Microbial granulation involves different trophic bacterial groups, and physico-
chemical and microbiological interactions (Yu et al., 2001). Sludge granules are dense,
multi-species, microbial communities, and none of the individual species in the granular
ecosystem is capable of degrading complex organic wastes (Liu et al., 2003).
The key feature of the UASB process that allows the use of high volumetric COD
loadings compared to other anaerobic processes is the development of a dense granulated
sludge. Because of the granulated sludge floc formation, the solids concentration can
range from 50 to 100 g/L at the bottom of the reactor and 5 to 40 g/L in a more diffuse
zone at the top of the UASB sludge blanket. The granulated sludge particles have a size
range of 1.0 to 3.0 mm and result in excellent sludge thickening properties with sludge
volume index (SVI) values less than 20 mL/g. Several months may be required to
develop the granulated sludge, and seed is often supplied from other facilities to
accelerate the system startup. Variations in morphology were observed for anaerobic
granulated sludge developed at 30 and 20 C, but both exhibited similar floc size and
settling properties (Tchobanoglous et al., 2003).
The presence of other suspended solids in the sludge blanket can also inhibit the
density and formation of granulated sludge. The formation of dense granulated sludge
floc particles is favoured under conditions of near neutral pH, a plug flow hydraulic
regime, a zone of high hydrogen partial pressure, a nonlimiting supply of NH4-N, and a
limited amount of the amino acid cysteine. With a high hydrogen concentration and
sufficient NH4-N, the bacteria responsible for granulation may produce other amino acids,
but their synthesis is limited by the cysteine supply. Some of the excess amino acids that
are produced are thought to be secreted to form extracellular polypeptides which, in turn,
will bind organisms together to form the dense pellets or floc granules (Tchobanoglous et
al., 2003).
It has been shown that divalent metal ions, such as Ca2+ and Fe2+, enhance the
granulation. Divalent ions were reported to play an important role in microbial
aggregation. It was found that extracellular polymers prefer to bind multivalent metals
due to the formation of stable complexes. It was observed that distinct improvement in
the biomass settleability and specific activity after replacing NaHCO3 as a neutralizing
agent by Ca(OH)2. At concentrations of 100 to 200 mg/L, calcium was found to exert a
positive impact on granulation (Yu et al., 2001). Mineral precipitates such as CaCO3,
Ca5OH(PO4)3, and FeS are formed, as a result of metabolic activities and physicochemical reactions, and are accumulated either inside or on the surface of granules.
However, there are also contradictory reports about the role of Ca2+ in granulation. Some
researchers have argued that Ca2+ does not induce granulation, or is not a key element in
granules, or is even detrimental to granule formation together with phosphate.
One major drawback of the UASB reactor is its extremely long start-up period,
which generally requires 2 to 8 months for the development of anaerobic granular sludge
(Liu et al., 2003). In order to reduce the space-time requirements of bioreactors and
eventually lead to a cheaper treatment of high strength wastes, strategies for expediting
granular formation are highly desirable. For achieving such a purpose, the mechanisms
for anaerobic granulation should be thoroughly understood.
Figure 2.5: Schematic representation of the inert nuclei model (Liu et al., 2003).
Figure 2.7: Schematic representation of the polymer or filament bonding model (Liu et
al., 2003).
Based on the microscopic observations, a multi-layer model for anaerobic
granulation was initially proposed. According to this model, the microbiological
composition of granules is different in each layer. The inner layer mainly consists of
methanogens that may act as nucleation centers necessary for the initiation of granule
development. H2-producing and H2-utilizing bacteria are dominant species in the middle
layer, and a mixed species including rods, cocci and filamentous bacteria takes
predominant position in the outermost layer as illustrated in Figure 2.8.
Figure 2.8: Schematic representation of the multi-layer model (Liu et al., 2003).
CHAPTER 3
RESEARCH METHODOLOGY
3.1
distributor and six outlets. The dimension of UASB reactor was measured then the
reactor volume and working volume could be determined. Table 3.1 showed the
dimension of the UASB reactor. No gas-solid separator (GSS) device is installed in the
UASB reactor as the constraint of the existing reactor design and dimension. Then, the
UASB reactor was connected to the pumps and other tanks or containers with proper
pipelines. No leaking from the reactor and pipelines connections had been confirmed
before conducting any experiments. The schematic diagram of the UASB reactor was
demonstrated in Figure 3.1.
Measurement value
18.5 cm
52 cm
268.8 cm2
14 L
12.5 L
Effluent
Recycle
Pump 2
Pump 1
Influent
Tank
3.2
Kilang Sawit Bukit Besar, in Kulai. The cooperation from the palm oil mill and the
people in-charge, Mr. Razak Jaafar lead to smooth sampling frequency and time. POME
samples were freshly collected and preserved at 4C in the cool room. Large and bulky
materials in the raw POME samples were removed before the dilution. Raw POME
samples were diluted using tap water. Figures 3.2 and 3.3 showed the front view of the
palm oil mill and the sampling point of raw POME.
3.3
biochemical oxygen demand (BOD5), chemical oxygen demand (COD), soluble chemical
oxygen demand (sCOD), total solids (TS), total suspended solids (TSS), volatile
suspended solids (VSS), ammoniacal nitrogen (Am-N), and nitrate (NO3-) were
determined according to the Standard Methods for the Examination of Water and
Wastewater (APHA, 2000) and DR/4000 Spectrophotometer Procedures Manual. There
were 5 respective samples of raw POME were collected freshly and determined in the
laboratory. Hexane was used as the organic solvent for the O&G extraction by using
separatory funnel. Appendix A showed the procedures of O&G determination. The
characteristics of these 5 raw POME samples were illustrated in the range of minimum
and maximum detected concentrations. The characteristics of the POME were compared
to the given data as in Appendix B.
3.4
preparation of diluted POME samples as influent for the reactor startup were controlled at
5000 1000 mg/L COD. To produce such strength of influents, about 2 L of raw POME
was diluted into 30 L. The pHs of influents were adjusted to neutral (pH 7 0.10). The
working volume (Vw) for the UASB reactor was 12.5 L and the hydraulic retention time
(HRT) 12.9 hrs, the flow rate of Pump 1 could be determined. The influents were
continuously flowed into the reactor daily. Parameters such as pH, COD, TSS and VSS
of the both influent and effluent were determined. The analysis methods for the desired
parameters were referred to standard methods. The startup of the reactor was proceeding
until the steady state condition. Steady state condition was defined when the COD
removal was constant (not more than 5% difference). The controlled operating
parameters were illustrated in Table 3.2.
3.5
Desired Value
5000 mg/L
7
12.9 hrs
9.3 g/L.d COD
12.5 L
0.97 L/hr
15.0 L/hr
0.59 m/hr
Design of Experiments
The important factors such as strength of wastewater (based on COD), influent pH
and HRT were selected which may affect the response variables as in Table 3.3. As full
factorial design was chosen, the combination treatments of the factors in one repeat were
8 (23) as shown in Table 3.4. The experiments were conducted in three replicates and the
steady state should be achieved before the next replicate is started. The percentages of
O&G biodegraded and COD removed in effluents were the response variables. Table 3.5
showed the low and high levels of various control operating parameters.
Table 3.3: Full factorial design of experiment with three factors in two levels.
Factors
A: HRT
B: Influent COD
C: Influent pH
Level (+)
12.9 hrs
15,000 mg/L
7
Level (-)
3.3 hrs
5,000 mg/L
3
HRT, hrs
(A)
3.3
12.9
3.3
12.9
3.3
12.9
3.3
12.9
Influent pH
(C)
3
3
3
3
7
7
7
7
Treatment
Combination
1
a
b
ab
c
ac
bc
abc
Desired Value
5,000 @ 15,000 mg/L
3@7
3.3 @ 12.9 hrs
12.5 L
3.79 @ 0.97 L/hr
15.0 L/hr
0.70 @ 0.59 m/hr
The analysis of data from the 3 levels 2 factors full factorial designed experiments
(COD removal and O&G biodegradation) were conducted to investigate the main effects
which have large impacts on these response variables. The method and formula were
applied from Design and Analysis of Experiments (Montgomery, 2005).
CHAPTER 4
4.1
were determined. The determined parameters include pH, oil and grease, biochemical
oxygen demand (5 days), chemical oxygen demand, soluble chemical oxygen demand,
total solids, total suspended solids, volatile suspended solids, ammoniacal nitrogen and
nitrate. The results of 5 respective raw POME were analyzed and presented in Table 4.1
below.
Concentration (mg/L)
4.5 - 5.0
8,000 - 11,000
18,000 - 21,000
68,000 - 80,000
33,000 - 41,000
65,000 - 70,000
35,000 - 45,000
31,000 - 37,000
280 - 460
300 - 510
The raw POME samples that collected from Bukit Besar, Kulai were the mixed
effluent from sterilizer condensate, clarification sludge and hydrocyclone discharged.
The detected parameters in each stage are given in Appendix B. The characteristic of the
raw POME above was considered the mixtures of the effluent from those three sources. It
was match with the data given in Appendix B.
From the range of the data obtained the raw POME samples had slightly different
compared to the reported typical quality of POME. This POME had higher strength than
the typical. However, the detected maximum and minimum concentrations do not vary
too much. Hence, it was able to control the strength of the influents for the digestion of
UASB reactor in this study.
4.2
mg/L COD, pH 7 and HRT 12.9 hrs) continuously. The influents were freshly prepared
while the effluents were collected on time. The parameters such as pH, COD, TSS and
VSS of the influents and effluents were determined daily. The data collected were
tabulated in the Table 4.2.
From the data in Table 4.2, the COD removal, TSS removal and VSS removal of
the reactor could be determined. The correlation of the COD removal and effluent pH
within startup period was plotted as in Figure 4.1. The TSS and VSS removal within the
startup period was also plotted as in Figure 4.2. Then, the steady state condition of the
reactor could be defined.
Table 4.2: Data collected for the influents and effluents during the reactor startup period.
pH
Days
1
2
3
4
8
9
10
11
15
16
17
18
19
22
23
24
25
26
In
7.04
7.03
7.01
7.02
7.00
7.08
6.98
7.03
7.10
7.08
7.07
7.05
7.06
7.05
7.01
7.01
7.03
7.07
Out
4.91
4.97
5.04
5.03
5.02
5.09
5.10
5.14
5.13
5.15
5.14
5.14
5.15
5.16
5.16
5.17
5.17
5.16
COD (mg/L)
In
Out
5390
3340
5210
3060
5640
3070
5270
2820
5750
3130
5720
2860
5670
2680
5340
2420
5810
2980
5070
2480
4970
2350
5460
2310
5350
2050
5210
1980
5180
1960
5090
1920
5200
1950
5060
1910
TSS (mg/L)
In
Out
1850
850
2060
740
1710
700
1630
600
2420
490
1810
430
1860
360
2540
480
2830
640
2220
690
2120
700
2710
530
1940
460
1920
430
1890
430
2010
490
2130
530
2080
500
VSS (mg/L)
In
Out
1770
690
1850
630
1590
570
1580
540
1980
420
1730
380
1800
340
2020
400
2020
490
1870
560
1700
610
2100
470
1850
380
1710
360
1700
340
1970
400
2000
470
1990
450
5 .2 0
C O D R e m o v a l, %
5 .1 5
5 5 .0
5 .1 0
5 0 .0
5 .0 5
4 5 .0
5 .0 0
4 0 .0
4 .9 5
3 5 .0
Effluent pH
COD Removal, %
Ef f lu e n t p H
6 0 .0
4 .9 0
1
9 10 11 15 16 17 18 19 22 23 24 25 26
D ay
Figure 4.1: COD removal and effluent pH within the startup period.
Removal, %
8 0 .0
7 5 .0
7 0 .0
T S S R e m o v a l, %
V S S R e m o v a l, %
6 5 .0
6 0 .0
5 5 .0
5 0 .0
1
9 10 11 15 16 17 18 19 22 23 24 25 26
D ay
Figure 4.2: TSS and VSS removal within the startup period.
The steady state condition was achieved after 26 days acclimatization with the
maximum COD removal varies in the range of 62% to 63%, which varies less than 5%.
The effluent pH presented the similar trend as the COD removal and the pH of the
effluents varies from 5.15 to 5.17 in the steady state condition. In these range of pH
values, the process of acidogenesis was more dominant than methanogenesis. In the
process of acidogenesis, O&G was hydrolyzed and fermented to produce acetate,
hydrogen, carbon dioxide, and propionate and butyrate. The propionate and butyrate
were fermented further, to also produce hydrogen, carbon dioxide, and acetate as
precursors for methanogenesis.
The TSS and VSS removal showed the same trends within the startup period. The
UASB reactor was able to remove the TSS and VSS of the influents about 75% to 78%
and 77% to 80% respectively at the steady state condition. The solids and particulate
organic matters did not accumulated in the reactor but they were hydrolyzed and
fermented into soluble form as no significant increment of sludge volume was observed.
The sharp drops at day 3 and 17 were caused by the washout of the filamentous sludge
and particulate of the scum layer at the top of reactor.
4.3
Design of Experiments
The data of COD, pH and O&G for the influents and effluents in three replicates
were collected and tabulated in Tables 4.3, 4.4 and 4.5. The percentages for COD and
O&G removal could also be calculated.
HRT
(hrs)
1
2
3
4
5
6
7
8
3.3
12.9
3.3
12.9
3.3
12.9
3.3
12.9
COD (mg/L)
In
Out
pH
O&G (mg/L)
Removal, %
In
Out
34.8
43.7
30.8
44.0
57.5
62.1
56.6
60.0
3.02
3.00
3.01
3.02
7.02
7.01
7.03
7.00
3.78
3.32
3.28
3.22
5.76
5.13
5.48
5.12
5090 3320
5080 2860
15340 10620
15120 8470
5740 2440
5720 2170
14640 6350
14840 5940
In
Out
Biodegradation, %
561 373
541 299
1713 1134
1643 955
540 268
529 196
1589 798
1623 610
33.5
44.7
33.8
41.9
50.4
63.0
49.8
62.4
1
2
3
4
5
6
7
8
HRT
(hrs)
3.3
12.9
3.3
12.9
3.3
12.9
3.3
12.9
COD (mg/L)
In
Out
5070 3330
5010 2800
14860 10170
15180 8580
5670 2420
5370 2050
15320 6640
14460 5930
pH
O&G (mg/L)
Removal, %
In
Out
34.3
44.1
31.6
43.5
57.3
61.8
56.7
59.0
3.03
3.01
3.03
3.05
7.05
7.06
7.04
7.03
3.95
3.33
3.29
3.24
5.70
5.15
5.52
5.16
In
Out
542 358
533 292
1682 1108
1701 924
538 263
498 187
1613 805
1654 630
Biodegradation, %
33.9
45.2
34.1
45.7
51.1
62.4
50.1
61.9
1
2
3
4
5
6
7
8
HRT
(hrs)
3.3
12.9
3.3
12.9
3.3
12.9
3.3
12.9
COD (mg/L)
In
Out
5200 3420
5130 2880
14980 10320
15290 8590
5350 2290
5410 2100
14890 6550
15050 6080
pH
O&G (mg/L)
Removal, %
In
Out
34.2
43.9
31.1
43.8
57.2
61.2
56.0
59.6
3.10
3.02
3.01
3.05
7.00
7.09
7.01
7.04
3.99
3.30
3.26
3.25
5.72
5.17
5.50
5.13
In
Out
529 341
538 301
1699 1111
1665 931
551 271
531 195
1605 817
1633 644
Biodegradation, %
35.5
44.1
34.6
44.1
50.8
63.3
49.1
60.6
From the results in these three replicates, the optimum combination of operating
parameters was HRT 12.9 hrs, influent 5,000 mg/L COD and influent pH 7 which success
to remove 61% to 62% COD and biodegrade 62% to 63% O&G in experiment Run 6.
However, the minimum COD removal was 31% and O&G biodegradation was 34% in
experiment Run 3, under operating parameters HRT 3.3 hrs, influent 15,000 mg/L COD
and influent pH 3. Other combinations of treatments responded to the intermediate COD
removal and O&G biodegradation.
The COD removal was considered linear to the O&G biodegradation as illustrated
by the scatter plot in Figure 4.3. The R2 is 0.8931, this show that the close relation
between COD removal and O&G biodegradation. Also, from the regression equation the
percent of O&G biodegradation could be predicted from the percent of COD removal or
vice versa.
COD Removal, %
60
y = 1.0271x - 0.5107
2
R = 0.8931
50
40
30
20
10
0
0
10
20
30
40
50
60
70
O& G B iodegradation, %
Figure 4.3: Scatter plot of COD removal and O&G biodegradation in three replicates.
Influent
COD (B)
5,000
5,000
15,000
15,000
5,000
5,000
15,000
15,000
Influent pH
(C)
3
3
3
3
7
7
7
7
COD Removal, %
Average
34.4
43.9
31.2
43.8
57.3
61.7
56.4
59.5
Effect Estimate
A
B
C
AB
AC
BC
ABC
Pure Error
Total
7.38
-1.62
20.43
0.47
-3.65
0.08
-1.10
-
Sum of
Squares
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
1.99
2938.43
Percent
Contribution
11.1311
0.5336
85.2540
0.0446
2.7205
0.0014
0.2471
0.0677
-
The largest effects on COD removal were influent pH (C=20.43), HRT (A=7.38)
and HRT-influent pH interaction (AC= -3.65). The effect of AC was negative, this
suggested that increasing the HRT-influent pH interaction would decrease COD removal.
The influent COD (B= -1.62) and HRT-influent COD-influent pH interaction (ABC= 1.10) did not appear to have as large impacts on COD removal deviation as the main
effects. The effect of B and ABC were negative, this suggested that increasing the
influent COD and HRT-influent COD-influent pH interaction would decrease COD
removal. The main effects C, A and AC dominated the COD removal accounting for
99% of the total variability, whereas the B and ABC interaction accounting for 0.8%.
The analysis of variance might be used to confirm the magnitude of these effects.
The degree of freedom, mean squares, Fo and P-Value for the factors and their
interactions were determined and listed in Table 4.8.
Sum of
Squares
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
1.99
2938.43
Degree of
Freedom
1
1
1
1
1
1
1
16
23
Mean
Square
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
0.12
-
Fo
2725.67
130.67
20876.08
10.92
666.17
0.33
60.50
-
P-Value
<0.0001
<0.0001
<0.0001
0.0045
<0.0001
0.5736
<0.0001
-
From Table 4.8, we note that the main effects (C, A and AC) on COD removal
were highly significant as all have very small P-Value. The other effects (B and ABC
interaction) on COD removal were also significant as their small P-Value.
The main effects of HRT (A), influent pH (C) and HRT-influent pH (AC)
interaction were plotted in Figure 4.4 (a, b and c). The A and C effects were positive, if
considered only these main effects, we would run these factors at the high level to
maximize the COD removal. However, it was always necessary to examine any
interactions that are important. The main effects did not have much meaning when they
were involved in significant interactions.
Figure 4.4: Main effects and interaction plots for COD removal.
From the interaction (AC) plot, the two lines were not parallel so they would meet
and cross at higher level HRT. This indicated that the interaction of HRT-influent pH had
significant large impact on COD removal. The HRT effect was small when the influent
pH was at the low level (black line) but large when the influent pH was at the high level
(red line), with the best results obtained when HRT (A) and influent pH (C) were at the
high level.
This could be confirmed by the 2D contour plot in Figure 4.5. Notice that the
contours were curved lines as the model contains an interaction, HRT-influent pH (AC)
interaction. This contour plot could be used as a guide to control the factors, HRT (A)
and influent pH (C) to achieve optimum COD removal.
HRT (hrs)
3.2
3.6
4.0
4.4
4.8
5.2
5.6
6.0
6.4
6.8
12
12
11
11
10
10
4
3.2
3.6
4.0
4.4
4.8
5.2
Influent pH
5.6
6.0
6.4
COD
< 35
35 - 40
40 - 45
45 - 50
50 - 55
55 - 60
> 60
6.8
4.3.2
Influent
COD (B)
5,000
5,000
15,000
15,000
5,000
5,000
15,000
15,000
Influent pH
(C)
3
3
3
3
7
7
7
7
O&G Biodegradation, %
Average
34.3
44.7
34.2
43.9
50.8
62.9
49.7
61.6
Effect Estimate
A
B
C
AB
AC
BC
ABC
Pure Error
Total
11.05
-0.82
16.98
-0.20
1.00
-0.37
0.12
-
Sum of
Squares
732.62
4.00
1730.60
0.24
6.00
0.81
0.08
13.37
2487.72
Percent
Contribution
29.4495
0.1608
69.5657
0.0096
0.2412
0.0326
0.0032
0.5374
-
The largest effects on O&G biodegradation were influent pH (C=16.98) and HRT
(A=11.05). The HRT-influent pH interaction (AC=1.00) and influent COD (B= -0.82)
did not appear to have as large impacts on O&G biodegradation deviation as the main
effects. The effect of B was negative, this suggested that increasing the influent COD
would decrease O&G biodegradation. The main effects C and A dominated the O&G
biodegradation accounting for 99% of the total variability, whereas the AC interaction
and B accounting for 0.4%.
The analysis of variance might be used to confirm the magnitude of these effects.
The degree of freedom, mean squares, Fo and P-Value for the factors and their
interactions were determined and listed in Table 4.11.
Sum of
Squares
732.62
4.00
1730.60
0.24
6.00
0.81
0.08
13.37
2487.72
Degree of
Freedom
1
1
1
1
1
1
1
16
23
Mean
Square
732.62
4.00
1730.60
0.24
6.00
0.81
0.08
0.84
-
Fo
872.17
4.76
2060.24
0.29
7.14
0.96
0.10
-
P-Value
<0.0001
0.0444
<0.0001
0.5976
0.0167
0.3418
0.7559
-
From Table 4.11, we note that the main effects (C, and A) on O&G
biodegradation were highly significant as all had very small P-Value. The AC interaction
and B was significant at about the 10 percent level, thus there was some mild interaction
between HRT and influent pH as well as influent COD on O&G biodegradation.
The main effects of HRT (A) and influent pH (C) as well as their interaction were
plotted in Figure 4.6 (a, b and c). The A and C effects were positive, if considered only
these main effects, we would run these factors at the high level to maximize the O&G
biodegradation. However, it was always necessary to examine any interactions that were
important.
Figure 4.6: Main effects and interaction plots for O&G biodegradation.
From the interaction (AC) plot, the two lines were approximately parallel so they
would not meet and cross. This indicated that there was lack of interaction between
HRT-influent pH on O&G biodegradation. This could be confirmed by the 2D contour
plot in Figure 4.7. Notice that the contours were parallel straight lines as the model did
not contain an interaction. This contour plot could be used as a guide to control the
factors, HRT (A) and influent pH (C) to achieve optimum O&G biodegradation.
HRT (hrs)
3.2
3.6
4.0
4.4
4.8
5.2
5.6
6.0
6.4
6.8
12
12
11
11
10
10
4
3.2
3.6
4.0
4.4
4.8
5.2
Influent pH
5.6
6.0
6.4
6.8
O&G
< 35
35 - 40
40 - 45
45 - 50
50 - 55
55 - 60
> 60
CHAPTER 5
DISCUSSION OF RESULTS
5.1
be treated by UASB reactors. From the data obtained, the ratio of BOD to COD was 0.26
and the fraction of insoluble COD was 50% which indicated that raw POME was
categorized as less biodegradable or complex wastewater.
The average O&G concentration of 9,500 mg/L was detected in raw POME, this
reflected that the unrecovered hydrophobic pollutants from the palm oil extraction
process. High concentration of O&G would lead to the formation of scum layers
consisting of floating substrate ingredients together with entrapped or attached active
sludge.
Particularly the presence of suspended fats and lipids will heavily promote the
tendency for sludge flotation, both flocculent and granular. This scum layer formation
may result in a more significant washout of active matter, as well as in the production of
considerable quantities of poorly stabilized excess scum layer sludge. Heavy foaming in
the reactor can only be avoided by operating the system under moderate loading rates and
maintaining good contact between sludge and wastewater (Lettinga and Hulshoff Pol,
1991).
On top of that, the average TSS concentration of 40,000 mg/L was detected in the
raw POME. High concentration of suspended solids (potentially precipitating matter)
presented in the wastewater would reduce the specific methanogenic activity of the sludge
in the case the suspended solids were poorly or non biodegradable and when it
accumulated in the sludge bed, either by a mechanism of mechanical entrapment or of
physical adsorption. Also, the attachment of newly generated bacterial matter to the
surface of suspended particles would, in the case when flocculent seed sludge was used,
slow down or even might completely counteract the formation of granular sludge.
Apart from the characteristics of the TSS, also the concentration of the dispersed
matter is of big importance. It will be evident that beyond a certain TSS concentration,
depending on the characteristics of the TSS, an anaerobic treatment system like the
UASB reactor will become less feasible (Lettinga and Hulshoff Pol, 1991).
The characteristics of the raw POME would guide to the UASB reactor designs.
By knowing the high strength of the raw POME was not possible directly apply to the
UASB reactor, dilution of this raw POME was necessary in order to reduce the strength
of the original raw POME. The strength of the wastewater was depending on the COD
concentration.
5.2
digestion. Its main task is to develop a highly active settleable sludge as quickly as
possible. While as anaerobic bacteria are slow growing microorganisms, major problems
encountered with UASB are the typical long reactor startup and spontaneous development
of biogranulation. Startup times for UASB reactors with digested sewage sludge usually
take several months (Show et al., 2004).
With the short startup period, the granular sludge is not able to be formed in the
UASB reactor. It generally requires 2 to 8 months for the development of anaerobic
granular sludge. In order to reduce the space-time requirements of bioreactors and
eventually lead to a cheaper treatment of high strength wastes, strategies for expediting
granular formation are highly desirable (Liu et al., 2003).
For the reactor startup, the steady state condition was achieved after 26 days
acclimatization with the maximum COD removal varies in the range of 62% to 63%.
Although the steady state was achieved, the startup period was relatively short. This had
lead to the used of flocculent sludge in this study.
Flocculent sludge performs less treatment efficiency than granular sludge for
UASB reactor (Hulshoff Pol et al., 2004). With the strength of influent COD 5,000 mg/L
and organic loading 9.3 g/L.d COD, these parameters were slightly higher than the
recommended in UASB design. For the strength of COD between 2,000 to 6,000 mg/L,
fraction of insoluble COD 30% to 60%, the applicable organic loading rate is 4 to 8 g/L.d
(Tchobanoglous et al., 2003). In this case, the COD removal had to be satisfied below
63% at steady state condition.
The effluent pH varies from 5.15 to 5.17 in the steady state condition was
observed. Anaerobic processes are sensitive to pH and inhibitory substances. A pH value
near neutral is preferred and below 6.8 the methanogenic activity is inhibited
(Tchobanoglous et al., 2003). In addition, a pH of around 6.0 appears to be the optimum
for acidogenic activity. If the pH is not controlled, the acidogenic biomass tends to buffer
itself to a pH that depends on the other environmental conditions in the reactor but mainly
on the nature of the wastewater (Borja et al., 1996). Therefore, the process of
acidogenesis was more dominant than methanogenesis in this UASB reactor. The gas
produced in the acidogenic reactor is primarily carbon dioxide and hydrogen
(Tchobanoglous et al., 2003).
The predominant volatile fatty acid found during this period is acetic acid.
Propionic, butyric and valeric acids are also detectable. These volatile fatty acids as
products of the acidogenesis can diffuse into the cells of anaerobic bacteria and ionize to
decrease the pH. This is generally recognized as inhibiting acidogenesis. The alkalinity
of the influent allows the neutralization of the free volatile fatty acids, preventing the fall
in pH from its optimum value both in the reactant influent and in the cells, while the
volatile fatty acids in the acidogenic UASB increased (Borja et al., 1996). In this study,
the influents that supplied into the UASB reactor were adjusted to pH 7 for the
neutralization purpose during the startup period.
In this study, the removals of TSS from 75% to 78% and VSS from 77% to 80%
were achieved at the steady state condition. The acidogenic reactor although tolerant to
an influent TSS concentration of 5,400 mg/L, showed a sign of reduced solids
liquification when the loading was increased and the TSS concentration was 10,800 mg/L
(Borja et al., 1996).
However, the concentration of TSS in the influents that supplied into the UASB
reactor during startup period varies from 1,630 to 2,830 mg/L. As a result, the solids and
particulate organic matters did not accumulated in the reactor but they were hydrolyzed
and fermented into soluble form as no significant increment of sludge volume was
observed.
For the wastewater that contains O&G substances, it can cause foaming and scum
formation (Tchobanoglous et al., 2003). The sharp drops of TSS and VSS removals at
day 3 and 17 were caused by the washout of the filamentous sludge and particulate of the
scum layer at the top of reactor.
5.3
it is necessary to study the joint effect of the factors on a response. There are several
assumptions for factorial designs, the factors are fixed, the designs are completely
randomized and the usual normality assumptions are satisfied. The 2k design is
particularly useful in the early stages of experimental work, when there are likely to be
many factors to be investigated.
It provides the smallest number of runs with which k factors can be studied in a
complete factorial design. Consequently, these designs are widely used in factor
screening experiments. As there are only two levels for each factor, we assume that the
response is approximately linear over the range of the factor levels chosen. In many
factor screening experiments, when starting to study the process or system, this is often a
reasonable assumption and discuss what action to take if it is violated (Montgomery,
2005).
The operation of UASB reactor is not restricted by the microbial capacity but is
limited by hydraulic factors. As the upflow velocity in the settler may not exceed value
of 1.5 m/hr, the minimum retention time is 4 to 6 hrs (Weiland and Rozzi, 1991).
Shorter HRT will lead to operational problems like sludge bed flotation, excessive
foaming at the gas-liquid interface, as well as accumulation of undigested ingredients.
Then, the treatment efficiency deteriorates (Mahmoud et al., 2003). From the
experiments, the COD removal and O&G biodegradation were increasing from low level
of HRT (3.3 hrs) to high level of HRT (12.9 hrs). Longer HRT will provide longer
contact between wastewater and the biomass. Because of this reason, HRT plays as
major role to the COD removal and O&G biodegradation.
Since O&G are esters, they undergo hydrolysis with more or less ease. The
hydrolysis may be induced by chemical means, usually by treatment with NaOH, or by
bacterial enzymes that split the molecule into glycerol plus fatty acids. Hydrolysis with
the aid of NaOH is called saponification (Sawyer et al., 2003). Anaerobic processes are
sensitive to pH and inhibitory substances. A pH value near neutral is preferred for
optimum bacterial enzymes activities (Tchobanoglous et al., 2003).
From the experiments, the COD removal and O&G biodegradation were
increasing from low level of influent pH (3) to high level of influent pH (7). To achieve
the high level of pH, the influent pH was adjusted to neutral by the addition of NaOH. So,
the O&G was hydrolysed by chemical means (saponification) and bacterial enzymes
(anaerobic sludge in UASB). For that reasons, the influent pH was considered main
effect to COD removal and O&G biodegradation.
Although high level of HRT and influent pH would increase the COD removal
and O&G biodegradation, they might have the constraints to go higher levels. Higher
level of pH might be toxic to the biomass (acidogenesis and methanogenesis). This
would cause the low bioactivity and treatment efficiency. Then, for higher level of HRT
the UASB reactor might able to biodegrade the slowly biodegraded matters and achieve
better treatment efficiency. However, it might not be considered as a high rate anaerobic
treatment system.
CHAPTER 6
6.1
Conclusion
The following conclusions have been proved in the study:
i.
The steady state of the reactor was obtained at about 62% of COD removal after
26 days acclimatization.
ii.
Under the designed operating parameters, the maximum O&G biodegradation was
63% and COD removal was 62% for the treatment of POME using UASB.
iii.
The main factors that had statistically significant effect on COD removal were
influent pH, HRT and HRT-influent pH interaction.
iv.
While, the main factors that had statistically significant effect on O&G
biodegradation were influent pH and HRT.
Thus it would be concluded that the objectives of the study had been achieved.
6.2
Recommendations
Some recommendations that can be suggested for the future study are as follows:
i.
ii.
Reactor design should comply with the recommended UASB reactor design
considerations.
iii.
iv.
v.
vi.
vii.
Last but not least, the advanced detection method of O&G should be applied in
order to have better understanding of O&G biodegradation.
REFERENCES
Ahmad, A. L., Bhatia, S., Ibrahim, N. and Sumathi, S. (2005). Adsorption of Residual Oil
from Palm Oil Mill Effluent Using Rubber Powder. Brazilian Journal of Chemical
Engineering. 22(3): 371-379.
Ahmad, A. L., Ismail, S. and Bhatia, S. (2003). Water Recycling from Palm Oil Mill
Effluent (POME) Using Membrane Technology. Desalination. 157: 87-95.
Ahmad, A. L., Sumathi, S. and Hameed, B. H. (2005). Adsorption of Residue Oil from
Palm Oil Mill Effluent Using Powder and Flake Chitosan: Equilibrium and
Kinetic Studies. Water Research. 39: 2483-2494.
APHA, AWWA and WEF, (2000). Standard Methods for the Examination of Water and
Wastewater, 20th Edition. APHA, AWWA and WEF. Washington, DC.
Bae, B. U. and Shin, H. S. (1998). Performance of An Inner Tube-type Gas-Solid
Separator Device in a UASB Reactor. Bioresource Technology. 63: 23-27.
Basiron, Y. and Chan, K. W. (2004). The Oil Palm and Its Sustainability. Journal of Oil
Palm Research. 16(1): 1-10.
Basiron, Y., Balu, N. and Chandramohan, D. (2004). Palm Oil: The Driving Force of
World Oils and Fats Economy. Oil Palm Industry Economic Journal. 4(1): 1-10.
Borja, R. and Banks, C. J. (1994). Anaerobic Digestion of Palm Oil Mill Effluent Using
an Up-flow Anaerobic Sludge Blanket Reactor. Biomass and Bioenergy. 6(5):
381-389.
Borja, R., Banks, C.J. and Sanchez, E. (1996). Anaerobic Treatment of Palm Oil Mill
Effluent in A Two-Stage Up-flow Anaerobic Sludge Blanket (UASB) System.
Journal of Biotechnology. 45: 125-135.
Chow, M. C. and Ho, C. C. (2000). Surface Active Properties of Palm Oil With Respect
To the Processing of Palm Oil. Journal of Oil Palm Research. 12(1): 107-116.
Department of Environment (1999). Industrial Processes & the Environment (Handbook
No. 3) Crude Palm Oil Industry. Malaysia: Aslita Sdn Bhd.
Edem, D. O., Eka, O. U., Umoh, I. B., Udoh, A. P. and Akpan, E. J. (2003). Effect of Red
Palm Oil and Refined Palm Olein on Nutrient Digestion in the Rat. Pakistan
Journal of Nutrition. 2(5): 271-278.
Faisal, M. and Unno, H. (2001). Kinetic Analysis of Palm Oil Mill Wastewater Treatment
by a Modified Anaerobic Baffled Reactor. Biochemical Engineering Journal. 9:
25-31.
Hulshoff Pol, L. W., de Castro Lopes, S. I., Lettinga, G. and Lens, P. N. L. (2004).
Anaerobic Sludge Granulation. Water Research. 38: 1376-1389.
Laws of Malaysia. (2003). Environmental Quality Act and Regulations. Malaysia: MDC
Publishers Sdn Bhd.
Lettinga, G. and Hulshoff Pol, L. W. (1991). UASB-Process Design for Various Types of
Wastewater. Wat. Sci. Tech. 24(8): 87-107.
Lim, L. P. and Ujang, Z. (2004). The Operating Strategy for Treatment of Digested Palm
Oil Mill Effluent (POME) in a Flat Sheet Aerobic Membrane Bioreactor Under
Tropical Conditions. ASIA WATER: 1-8
Liu, Y., Xu, H. L., Yang, S. F. and Tay, J. H. (2003). Mechanisms and Models for
Anaerobic Granulation in Upflow Anaerobic Sludge Blanket Reactor. Water
Research. 37: 661-673.
Mahmoud, N., Zeeman, G., Gijzen, H. and Lettinga, G. (2003). Solids Removal in
Upflow Anaerobic Reactors, a Review. Bioresource Technology. 90: 1-9.
Malina, Jr. J. F. and Pohland, F. G. (1992). Design of Anaerobic Processes for the
Treatment of Industrial and Municipal Wastes. Pennsylvania: Technomic
Publishing Company, Inc.
Montgomery, D. C. (2005). Design and Analysis of Experiments. Sixth Edition. USA:
John Wiley & sons, Inc.
Ooi, T. L. and Yeong, S. K. (2000). Studies on Factors Affecting the Colour Stability of
Some Commercial Palm Fatty Acids. Journal of Oil Palm Research. 12(2): 63-72.
Pantzaris, T. P. and Ahmad, M. J. (2001). Properties and Utilization of Palm Kernel Oil.
Palm Oil Developments. 35: 11-23.
Sawyer, C. N., McCarty, P. L. and Parkin, G. F. (2003). Chemistry for Environmental
Engineering and Science. 5th ed. New York: McGraw-Hill Companies, Inc.
Shaaban, M. G. (1990). Performance Comparison of Anaerobic Bioreactors: Initial Study
of Upflow Anaerobic Sludge Blanket (UASB) Process in the Treatment of Palm
Oil Mill Effluent (POME). Bulletin IEM: 5-10.
Show, K. Y., Wang, Y., Foong, S. F. and Tay, J. H. (2004). Accelerated Start-Up and
Enhanced Granulation in Upflow Anaerobic Sludge Blanket Reactors. Water
Research. 38: 2293-2304.
Sundram, K. (2004). Palm Oil: Chemistry and Nutrition Updates. Malaysian Palm Oil
Board (MPOB): 1-22.
Tchobanoglous, G., Burton, F.L. and Stensel, H.D. (2003). Wastewater Engineering:
Treatment and Reuse, Fourth Edition. New York: McGraw Hill Companies, Inc.
Ugoji, E. O. (1997). Anaerobic Digestion of Palm Oil Mill Effluent and Its Utilization As
Fertilizer for Environmental Protection. Renewable Energy. 10(2): 291-294.
Weiland, P. and Rozzi, A. (1991). The Start-Up, Operation and Monitoring of High Rate
Anaerobic Treatment System: Discussers Report. Wat. Sci. Tech. 24(8): 257-277.
Yu, H. Q., Fang, H. H. P. and Tay, J. H. (2001). Enhanced Sludge Granulation in Upflow
Anaerobic Sludge Blanket (UASB) Reactors by Aluminum Chloride.
Chemosphere. 44: 31-36.
Yu, H. Q., Tay, J. H. and Fang, H. H. P. (2001). The Roles of Calcium in Sludge
Granulation during UASB Reactor Start-Up. Water Research. 35(4): 1052-1060.
APPENDIX A
APPENDIX B
APPENDIX C
2k Factorial Design Examples-Chemical Oxygen Demand Removal
Runs HRT
1
3.3
2
12.9
3
3.3
4
12.9
5
3.3
6
12.9
7
3.3
8
12.9
Factors
Inf. COD Inf. pH
5000
3
5000
3
15000
3
15000
3
5000
7
5000
7
15000
7
15000
7
R1
34.8
43.7
30.8
44.0
57.5
62.1
56.6
60.0
R2
34.3
44.1
31.6
43.5
57.3
61.8
56.7
59.0
R3
34.2
43.9
31.1
43.8
57.2
61.2
56.0
59.6
Total
Total
103.3
131.7
93.5
131.3
172.0
185.1
169.3
178.6
1164.8
Factors
(1)
a
b
ab
c
ac
bc
abc
Using the totals under the treatment combinations shown in table above, we may
estimate the factor effects as follows:
1
[a (1) + ab b + ac c + abc bc], where n = nos of replicate
4n
1
=
[131.7 103.3 + 131.3 93.5 + 185.1 172.0 + 178.6 169.3]
4(3)
1
= [88.6]
12
= 7.38
A=
1
[b + ab + bc + abc (1) a c ac], where n = nos of replicate
4n
1
=
[93.5 + 131.3 + 169.3 + 178.6 103.3 131.7 172.0 185.1]
4(3)
1
= [ 19.4]
12
= 1.62
B=
1
[c + ac + bc + abc (1) a b ab], where n = nos of replicate
4n
1
[172.0 + 185.1 + 169.3 + 178.6 103.3 131.7 93.5 131.3]
=
4(3)
1
= [245.2]
12
= 20.43
C=
1
[ab a b + (1) + abc bc ac + c], where n = nos of replicate
4n
1
=
[131.3 131.7 93.5 + 103.3 + 178.6 169.3 185.1 + 172.0]
4(3)
1
= [5.6]
12
= 0.47
AB =
1
[(1) a + b ab c + ac bc + abc], where n = nos of replicate
4n
1
[103.3 131.7 + 93.5 131.3 172.0 + 185.1 169.3 + 178.6]
=
4(3)
1
= [ 43.8]
12
= 3.65
AC =
1
[(1) + a b ab c ac + bc + abc], where n = nos of replicate
4n
1
=
[103.3 + 131.7 93.5 131.3 172.0 185.1 + 169.3 + 178.6]
4(3)
1
= [1.0]
12
= 0.08
BC =
1
[abc bc ac + c ab + b + a (1)], where n = nos of replicate
4n
1
[228.8 191.9 266.6 + 198.7 188.7 + 184.6 + 229.9 191.8]
=
4(3)
1
= [3.0]
12
= 0.25
ABC =
The largest effects are for HRT (A=7.38), Influent pH (C=20.43) and HRTInfluent pH interaction (AC=-3.65). The sums of squares are calculated as below:
SS =
(88.6)
SS A =
SS B =
8(3)
= 327.08
( 19.4)2
8(3)
= 15.68
(Contrast )2
8n
SSC
SS AB
2
(
245.2 )
=
8(3)
2
(
5 .6 )
=
8(3)
SS AC =
SS BC
= 1.31
( 43.8)2
8(3)
2
(
1 .0 )
=
SS ABC =
= 2505.13
8(3)
= 79.94
= 0.04
( 13.2)2
8(3)
= 7.26
The total sum of squares is SST = 2938.43 and by subtraction, SSE = 2.00. Table
below summarizes the effect estimates and sum of squares. The column labeled percent
contribution measures the percentage contribution of each model term to the total sum of
squares.
Factors
A
B
C
AB
AC
BC
ABC
Pure Error
Total
Factors
(A)
(B)
(C)
AB
AC
BC
ABC
Pure Error
Total
Effect Estimate
7.38
-1.62
20.43
0.47
-3.65
0.08
-1.10
Sum of
Squares
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
1.99
2938.43
Sum of Squares
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
1.99
2938.43
Degrees of
Freedom
1
1
1
1
1
1
1
16
23
Mean
Square
327.08
15.68
2505.13
1.31
79.94
0.04
7.26
0.12
-
Percent Contribution
11.1311
0.5336
85.2540
0.0446
2.7205
0.0014
0.2471
0.0677
F0
P-Value
2725.67
130.67
20876.08
10.92
666.17
0.33
60.50
-
<0.0001
<0.0001
<0.0001
0.0045
<0.0001
0.5736
<0.0001
-