Journal of Cleaner Production 134 (2016) 633e641

Contents lists available at ScienceDirect

Journal of Cleaner Production

journal homepage: www.elsevier.com/locate/jclepro

Material flow management and cleaner production of cassava

processing for future food, feed and fuel in Thailand
Napat Jakrawatana a, *, Prus Pingmuangleka a, Shabbir H. Gheewala b, c
a
School of Energy and Environment, University of Phayao, Phayao, Thailand
b
The Joint Graduate School of Energy and Environment, King Mongkut's University of Technology Thonburi, Bangkok, Thailand
c
Centre for Energy Technology and Environment, Ministry of Education, Thailand

a r t i c l e i n f o a b s t r a c t

Article history: This research applied Material Flow Cost Accounting (MFCA) to identify the costs of material and energy
Received 20 October 2014 loss, as well as the opportunity for technology improvement to increase productivity in starch and
Received in revised form ethanol productivity in Thailand. The results showed that ethanol production incurred more loss cost
15 June 2015
than starch production because it entails several conversion processes. Scenario for technology
Accepted 28 June 2015
Available online 3 July 2015
improvement was evaluated and expanded to the broader scale of the whole country to assess the
possibility of increasing cassava feedstock in order to meet the AEDP target. The result shows that the
Very High Gravity (VHG) and Simultaneous Saccharification and Fermentation (SSF) processes can in-
Keywords:
Cassava
crease plant capacity and efficiency. Ethanol production from cassava pulp can also play an important
Starch role in meeting the Renewable and Alternative Energy Development Plan (AEDP) and utilizing by-
Ethanol product from starch processing, increasing its value and offsetting loss cost from starch production
Material Flow Cost Accounting plant. Those improved options in the scenario can help to reduce cassava feedstock amount required to
Food produce ethanol for the AEDP target. Yield improvement can satisfy feedstock requirement without the
Fuel need of land for plantation expansion, reduce cassava export and allocation of more molasses for ethanol
production.
© 2015 Elsevier Ltd. All rights reserved.

1. Introduction feedstocks such as sugarcane and cassava. The proportion between

Thailand is one of the world's top exporters of cassava products.
Yield of cassava production and cassava processing efficiency in
Thailand has been improved considerably over the last decade.
Cassava is used mainly for food, feed and fuel. The use of cassava
can lead to a possibility of feedstock competition. Cassava pro-
duction in Thailand in 2012 was 26.6 million tons per year (MTPY).
The main utilization was starch production at 14.2 MTPY or 54%
(domestic use 4.8 MTPY, export 9.4 MTPY). Starch 11.9 MTPY or 44%
was used for producing chips and pelleted cassava for animal feed
(domestic use 3.4 MTPY, export 8.5 MTPY) and only 2% (0.5 MTPY)
was available for ethanol production (OAE, 2013).
The Renewable and Alternative Energy Development Plan
(AEDP 2012e2021) aims to expand ethanol production from 1.3
million litres per day (MLPD) (in 2012) to 9.0 MLPD. The plan leads
to the challenge in increasing demand of ethanol production
* Corresponding author. Tel.: þ66 87 513 8282; fax: þ66 54 466 704.
E-mail address: napat_j@hotmail.com (N. Jakrawatana).
sugarcane (molasses) and cassava to be used in ethanol production
depends on many factors including the crop production potential,
existing capacity and configuration of available ethanol plants and
the new upcoming ethanol plants. Sugarcane production in
Thailand has remained stable because the yield is generally lower
than other main sugar producing countries. In Thailand, molasses
and sugar production are strictly controlled as they must comply
with the regulations. The quota of sugar production for domestic
and export market is controlled by the Office of the Cane and Sugar
Board (OCSB) in order to control supply and demand and for sta-
bility of the sugar price (FAO, 2010; Sriroth et al., 2010; Nguyen
et al., 2007). Moreover, molasses is used in manufacturing alco-
holic beverages and is of high competitive use for both domestic
and export spirits and alcohol industry. In 2013, sugarcane pro-
duction in Thailand was 98 million tons, resulting in molasses
production of 4.4 million tons, out of which 1.4 million tons was
used in the domestic spirit and alcohol industry, whereas 1 million
tons was exported. There were 2 million tons left to be used for
ethanol industry (Bank of Thailand, 2014).

http://dx.doi.org/10.1016/j.jclepro.2015.06.139
0959-6526/© 2015 Elsevier Ltd. All rights reserved.
634 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641
On the other hand, cassava has potential to be used as the
feedstock for ethanol production in the future because of several
advantages. Firstly, cassava can tolerate all harsh environmental
conditions especially drought as also low input requirements in
planting and harvesting. Secondly, unlike feedstock from sugar-
based industries (molasses) that available seasonally, cassava can
grow, and can be harvested all year round. Thirdly, it can be con-
verted into dried chips easily and stored for long periods and is easy
to transport. Fourthly, the average yield in Thailand is high
compared to the global average and there is technology potential to
increase the yield even further in the future. The higher demand of
food, feed, and fuel has significant impact on the rising of cassava
price. For this reason, the harvested area of cassava tends to expand
continuously (FAO, 2010; Sriroth et al., 2010; Nguyen et al., 2007).
This makes cost of ethanol production from cassava slightly higher
than ethanol produced from molasses. However, feedstock material
use efficiency and land utilization of cassava seems to be better
than those of sugarcane (molasses) (Silalertruksa and Gheewala,
2009).
At present, even though over half of the ethanol plants in
Thailand are configured to use molasses as feedstock, a number of
new plants equipped cassava ethanol facilities are under con-
struction. Therefore, The Office of Agricultural Economics (OAE)
and the Ministry of Energy (MoE) set the plan for the share of
ethanol targets by using either cassava or molasses feedstock. The
proportion of molasses:cassava feedstock for ethanol production in
Thailand in 2018 would be 20:80 (FAO, 2010). Therefore, ethanol
produced from cassava would be 7.2 MLPD or 2160 MLPY (14 MTPY
cassava feedstock is required) (DEDE, 2014).
However, cassava feedstock for ethanol production is expected
to increase from 2.6 MTPY to over 14 MTPY between 2010 and 2022
(DEDE, 2014; FAO, 2010). If we spare 11.9 MTPY of cassava feedstock
for feed and 14.2 MTPY for starch production the same as in 2012,
without reduction of export, total requirement of cassava would be
40.1 MTPY. Therefore, cassava production has to be increased
considerably. The increase in cassava production can come largely
from yield improvements and expansion of the land area (FAO,
2010). Cassava plantation area in Thailand in 2013 is 1.2 million
ha. The yield has been targeted to increase from 21 ton/ha to 31 ton/
ha. This can increase the cassava production to 37 MTPY (DEDE,
2014). Therefore, yield improvement would not be sufficient to
cope with this challenge. Moreover, not only yield improvement
depends on a number of factors, but it also requires considerable
efforts and is difficult to achieve.
The expansion of the land area is considered as another effi-
cient way to increase production. However, the increase in the
harvested area of cassava may result in the reduction of land for
other food crops with a consequent reduction in food production.
Deforestation may also be increased if agricultural land en-
croaches directly or indirectly on forests. Development of tech-
nologies to improve eco-efficiency and productivity of cassava
supply chain (for example: good agricultural practices, water
recycling, technology modification for reduction of starch and
alcohol lost and material and energy recovery from by-product)
can be a positive alternative to assist Thailand in sustainably
reaching the AEDP target with minimum burden on the envi-
ronment (Blottnitz and Curan, 2007; Liewa et al., 2014; Nguyen
and Gheewala, 2008).
Material Flow Analysis (MFA) can be used to trace the flow of
materials (cassava and sugarcane feedstock and all different cassava
and sugarcane products and semi-products, by-products and
waste) through all the related processes. It provides a useful con-
ceptual framework to plan for regional sustainable resource and
waste management and to evaluate the application of resource
management (Birkeland and Schooneveldt, 2003). This research
will apply Material Flow Analysis (MFA) model as the main tool and
economic analysis will also be covered by Material Flow Cost Ac-
counting (MFCA) model. The loss cost and opportunity for tech-
nology improvement in a starch production and ethanol
production plant in Thailand will be evaluated. This research will
use data collected in 2012 as the base case and will generate sce-
narios to assess the possibility to meet the Thai AEDP Biofuel target
in 2022.
2. Material and methods
This research will apply Material Flow Analysis (MFA) model as a
main tool. The economic analysis will also be covered by Material
Flow Cost Accounting (MFCA) model in order to evaluate oppor-
tunities, effectiveness and potential of technology improvement for
cost reduction. METI (2007) has defined Material Flow Cost Ac-
counting (MFCA) as “one of the environmental management ac-
counting methods aimed to reduce both environmental impact and
costs at the same time by reducing costs through waste reduction,
thereby improving business productivity”.
Material loss in the production refers to non-productive or non-
efficient use of materials extracted from nature. MFCA is a tool that
can identify the cost of production in each sub-process as well as
categorize positive and negative products cost. For example, some
company that just applies MFCA found that its positive product is
70 percent whereas negative product can be up to 30 percent. If the
company can improve material efficiency or reduce their material
loss, their negative cost will reduce (meaning increase in profit).
This information can assist the decision making process of the top
management of the company on whether is it cost effective to
invest in reducing material loss and can help reduce negative
product cost.
MFCA tracks and quantifies the flow and stock of “materials”,
which include raw materials, parts and components in the
manufacturing process using the mass balance principle. Here, part
of raw materials that become waste, defective products and emis-
sions associated with material transformation, can be identified as
“material loss”.
Four categories of the cost information including material costs,
system cost (labor, depreciation etc.), energy cost and waste
treatment cost can be compiled as the quantity data based on
material flow. The MFCA results allocate product costs into 2 types:
“positive product cost” and “negative product cost”. “Positive
product costs” are the costs put into process products (positive
products) released to the next process. “Negative product costs” or
“loss costs” are costs put into wasted or recycled items (negative
products). Material loss cost can be simply calculated by multi-
plying individual quantities (kg) of waste by their material pur-
chased unit price, or by using raw material cost multiplies by
percent of material loss by weight. Energy or system loss cost can
be calculated by using mass allocation method: cost of energy input
multiplied by percent of material loss by weight in each sub-
process (out of material input in each sub-process) (METI, 2007).
MFCA considers the main material flows in all processes as well
as sub materials, except auxiliary, materials. Main materials, sub
materials and auxiliary materials are defined as follows (METI,
2007; Chompu-inwai et al., 2015):
 Main materials: The principal materials in the initial process and
the work-in-process from the previous process.
 Sub materials: Materials added to the main materials to form
part of the company products in each process
 Auxiliary materials: Materials that are used in each process but
do not form part of the company products;
 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641
MFCA involves several steps (adapted from (METI, 2007;
Chompu-inwai et al., 2015) including;
1. Preparation (Set system boundary and determine relevant ma-
terial, flow and processes)
2. Data collection & compilation (Collect and compile the data of
material flow quantities and material loss quantities, collect and
compile relevant cost data)
3. MFCA calculation (Do material balance and add all the cost data
to the material flow (linking physical and monetary flow).
4. Identifying improvement requirements (Identify and list re-
quirements for improvement, including material loss and cost
reduction)
2.1. Preparation: setting system boundary and determining relevant
material, flow and processes
The system boundary of the model is shown in Fig. 1. This
research will focus on the flow of materials and waste through
starch production and ethanol production using cassava. In starch
production, the scope will cover all processes including washing,
rasping, extraction, separation, dewatering (two-stage separation)
and drying. In ethanol production, the scope will cover raw mate-
rials preparation, liquefaction, fermentation, distillation and mo-
lecular sieve dehydration.
2.2. Data collection and compilation
This research considered only material cost and energy cost.
Auxiliary materials, such as chemicals, were not included because
they account for less than 1% of mass flow and do not form part of
the company's products. System cost and waste treatment cost
were not included due to confidentiality and limited availability of
Fig. 1. System boundary of the MFCA model in this research.
635
data. Data used in the MFCA calculation in this study come from
interviewing the plant manager and engineers in 2013.
In Thailand, there are 99 starch plants with total capacity of
34,141 ton starch/day. The average capacity of each plant is 350 ton
starch/day. This research collected the data from interviewing the
plant managers and engineers from 6 starch plants from all over
Thailand covering every region. Moreover, data from literature
(Chavalparit and Ongwandee, 2009) collected from 8 starch plants
in Thailand were also used for the analysis. All the data was
compiled and used to determine the average and range which will
be used in MFCA model. The calculation was based on 100 ton
cassava input for both factories for comparison. The results show
that productivity from different plants is similar (the range is ±7%
from the average), except for water use that has a high uncertainty
of about ±50%.
There are 10 cassava ethanol plants in Thailand under operation
with a total capacity of 1.9 MLPD. One plant is going to cease cas-
sava line and use only molasses line. Capacity of the plants range
between 130,000e400,000 L/day with average capacity of
175,000 L/day. Most of the plants are using the same Simultaneous
Saccharification and Fermentation (SSF) technology (Siamphakdee,
2011). There are two main kinds of cassava ethanol plants, one is
cassava chip feedstock ethanol plant and the other, fresh cassava
root feedstock ethanol plant. Some plants can use both feedstocks,
but cassava chip is still the main feedstock. Cassava chip feedstock
type just needs one more process of chipping and sundrying from
fresh cassava root type.
In Thailand, cassava chip feedstock ethanol plants dominate
with 70% of all capacity. This research, data was collected by
interviewing with plant manager and engineers from 2 ethanol
plants in the Northeastern region of Thailand. One of the plants
used cassava chip, and another one used fresh cassava as feedstock.
Both of the plants are using SSF technology. Moreover, data from
several previous studies in Thailand were also used (Gheewala
et al., 2013; Moriizumi et al., 2012; Sriroth, 2013). All the data
636 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641
were compiled and used to determine the average and range to be
used in MFCA model. The calculation was based on 100 ton cassava
input for both factories for comparison. The results show that
productivity from different plants is similar (the range is ±6% from
the average), except for energy and water use that have a range
about ±25e30%.
2.3. Scenario analysis
Scenarios of the new improvement measures or cleaner pro-
duction technologies are applied to the MFCA model of both starch
production and ethanol production to analyze capability to reduce
loss and increase processing efficiency and productivity.
Results from this MFCA analysis of the proposed improvement
options to increase productivity and reduce loss are employed to
look at the broader scale of the whole country. This analysis was
employed in order to evaluate the possibility to increase cassava
availability for ethanol production to meet AEDP target against
other options. Framework of the evaluation is shown in Fig. 2.
3. Results and discussion
3.1. Starch production in the business as usual (BAU)
The data of starch production plants collected from 6 starch
plants and from literatures were compiled and calculated to get the
average and uncertainty range as showed in Table 1. The results
show that productivity from different plants is similar (the range is
±7% from average). Energy use has medium uncertainty range of
±30%. Water uses have high uncertainty range at ±50% because
several plants have modified water cycle process in the system. Peel
and sand waste also have high uncertainty range up to ±50%
because it depends on the quality of feedstock from each season
and region.
The average and uncertainty range in Table 1 were used in MFCA
model. The calculation was based on 100 ton cassava input for
comparison. In the starch plants that the data has been collected,
biogas was used to produce hot air directly to dry the starch. This
may not be the most representative system, since many other
Fig. 2. Framework to evaluate the possibility to increase cassava availability for ethanol production.
Table 1
Average data and range of starch production system.
Input/Output Quantity
Input
 Cassava root (ton) 4.1 ± 0.3
 Water (m3) 18 ± 11
 Electricity (kWh) 170 ± 37
 Hot Air (MJ) 1250 ± 350
Output
 Starch (ton) 1
 Peel and sand þ Rhizomes (ton) 0.3 ± 0.15
 Wastewater (m3) 21 ± 9
 Cassavas pulp (ton) 1.5 ± 0.4
factories use either steam or hot oil to transfer heat from the boiler
to the dryer.
Material Flow Cost Accounting of the cassava starch production
system in BAU is shown in Fig. 3. The weight of material shown in
the table is in wet basis. The largest material loss occurred at the
extraction and separation processes accounting for almost 50 ± 10%
of the material loss. Other losses occurred through removal of
rhizome, sand and peel water vapor and small loss along the line of
the processes. Energy loss cost calculated here was done based on
MFCA principle. It was calculated by multiplying cost of energy
input in each sub-process with % energy loss of each sub-process.
Total material cost was 7870 ± 80 USD. Total material loss cost
accounted for 5950 ± 1410 USD. Total energy cost was 705 ± 160
USD with total energy loss cost of 475 ± 245 USD. The largest en-
ergy loss occurred at the extraction and separation processes. By-
product in the form of cassava pulp were sold as feed for about
565 ± 110 USD and wastewater was used to produced biogas that
can substitute fuel oil and saved cost 740 ± 150 USD. Some part of
biogas was used to produce electricity worth 75 ± 15 USD. How-
ever, the value of selling by-product and utilization as energy
(biogas) is small comparing to the loss cost mentioned earlier.
Material cost of cassava root is the largest variable cost in starch
production plant. This cost can be amounted over 80% of the whole
cost. So, the more cassava root changes to starch (reduce starch
loss), the more cost will be saved. Cost reduction can result in
significant profit gaining, especially in fiber and pulp separation
process and starch separation and dewatering process.
 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641 637

Fig. 3. Material flow cost accounting of the cassava starch production system in BAU.

3.2. The scenario of proposed improvement option to increase
processing efficiency and productivity and reduce loss in starch
production
In this scenario, the starch loss reductions were conducted in
the fiber and pulp separation process where the largest starch loss
occurred. Here, lean manufacturing measure was applied and
automated system installed in the coarse fiber and pulp extraction
process resulting in reduction of starch losses 10 kg/1 ton starch
(Roekklang and Pinitwachiraphong, 2010). In the fine extraction
process, fine-screen extractors (centrifugal screen extractors) were
replaced with vertical screen systems (Dutch State Mines Screen
(DSM)). Here, the reduction of starch losses is 2.5 kg/ton starch.
Reduction of water consumption is 2 m3/ton starch and reduction
of electricity consumption is 5 kWh/ton starch (Chavalparit and
Ongwandee, 2009). In this scenario, starch productivity increases
by 0.3 ton starch/100 ton of fresh root feedstock. The material loss
cost was slightly reduced from 5950 ± 1410 to 5920 ± 1406 USD and
energy loss cost was slightly reduced from 475 ± 245 to 460 ± 243
compared to BAU. Material flow cost accounting of the cassava
starch production system in this scenario is shown in Fig. 4. The
weight of material shown in the table is on wet basis. Here,
improvement of efficiency and productivity was small because the
majority of the plants in the industry has already been applying
cleaner technology to improve productivity for over a decade (for
example; good operational control, check and repair the leaks,
training operator to operate efficiently, recycling water from sep-
aration process and hydro cyclone back to rasping process to
recover starch loss, optimizing hot air generator and drying process
and recovering biogas from wastewater to use as a fuel in hot air
generator (NSTDA, 2009)).
3.3. Ethanol production in the business as usual (BAU)
The data of cassava ethanol plants were collected from 2 cassava
ethanol plants and literatures. The collected data has been
compiled and calculated to get the average and uncertainty range
as shown in Table 2. The results show that productivity from

Fig. 4. Material flow cost accounting of the cassava starch production system in the improved scenario.
638 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641
Table 2
Average data and range of cassava ethanol production system.
Input/Output Quantity
Input
 Fresh cassava (t) 5.83 ± 0.38
 Diesel to produce cassava chip (L) 11 ± 3
 Cassava chip (t) 2.4 ± 0.1
 Water (m3) 10 ± 3
 Steam (t) 2.8 ± 0.8
 Electricity (kWh) 354 ± 89
Output
 Ethanol (kL) 1 accounted at 18.5% as well as water input at 99 ton/100 ton of
 Spent wash (Thick slop) 9.5 ± 2.5
 Fusel oil (l) 3±1
different plants is similar (the range is ±6% from average). Energy
use has medium uncertainty range of ±25%. Water and wastewater
uses have high uncertainty range at ±30%.
The average and uncertainty range in Table 2 was used in MFCA
model. The calculation was based on 100 ton cassava input for
comparison. The largest material loss occurred at distillation and
dehydration process accounting for almost the whole of the ma-
terial loss. Material flow cost accounting of the ethanol production
system in BAU is shown in Fig. 5. The weight of material shown in
the table was in wet basis. Total material cost was 9200 ± 230 USD.
Total material loss cost during production was 8920 ± 1980 USD.
Total energy cost was 1370 ± 510 USD with total energy loss cost of
1305 ± 780 USD. The largest energy loss occurred at distillation and
dehydration process. Even though distillery spent wash was used to
produce biogas to substitute coal (fuel in the boiler) and saved cost
344 USD. However, the value of utilization as energy (biogas) is
small comparing to the loss cost mentioned earlier.
3.4. The scenario of proposed improvement options to increase
processing efficiency and productivity and reduce loss in ethanol
production
Here, two new technology modifications including Very High
Gravity (VHG) and Simultaneous Saccharification and Fermentation
(SSF) processes were introduced to increase ethanol production
efficiency. Very High Gravity (VHG) process applies the concept
used in beer production which entails carbohydrate hydrolyzing
enzymes to treat fresh cassava roots. This process produces the
Fig. 5. Material flow cost accounting of the ethanol production system in BAU.
syrups of high sugar contents having very high soluble solid con-
tents (>30%). After mash viscosity reduction and fermentation,
fermented mash with high ethanol concentrations (up to 14.6% w/
w or 18% v/v) can be achieved compared to conventional method
that produces fermented mash with 8% w/w or 10% v/v ethanol
(Sriroth et al., 2010). Simultaneous Saccharification and Fermen-
tation (SSF) technology allows saccharification and fermentation to
be conducted simultaneously. Even though VHG-SSF together has
yet became commercialized. The advantages from VHG-SSF appli-
cation are more plant efficiency and energy reduction input which
cassava roots comparing to the conventional system (Sriroth et al.,
2010; Sriroth, 2013; Piyachomkwan, 2011; Kim et al., 1997).
Material flow cost accounting of the ethanol production system
in the scenario of modifying technology to VHG-SSF is shown in
Fig. 6. The weight of material shown in the table is on wet basis.
Here, the loss in distillation and dehydration process would be
reduced from 94 ± 40% to 88 ± 37%. Total material loss cost during
production can be reduced from 8920 ± 1980 USD to 8620 ± 1850
USD as compared to BAU. Total energy loss cost can be reduced
from 1305 ± 780 USD to 988 ± 640 USD. Total loss cost was
9610 ± 2490 USD. When compared to BAU, loss cost reduction was
612 USD (as per 100 ton material processing). If the scenario is
expanded to the country scale, the total loss cost would be reduced
by about 70 ± 23 million USD.
Cassava pulp is a by-product from starch production system that
has potential for use as raw material for ethanol production. There
are 5 MTPY of cassava pulp being produced by 99 starch factories in
Thailand. Cassava pulp can be sold for animal feed for only 10 USD/
ton. New Energy and Industrial Technology Development Organi-
zation (NEDO) has cooperated with the National Innovation Agency
of Thailand (NIA) to support the development of technology for
ethanol production from cassava pulp. A pilot plant has been con-
structed, operated and tested in Thailand since 2014, and is ex-
pected to extend the project funding to 2016. The application of this
technology is planned to be expanded to starch plants across the
entire country. The pilot plant shows that a batch of 800 L ethanol
requires 9 ton of wet cassava pulp together with 2e3 ton of dry
cassava pulp (NIA, 2013; Li et al., 2011). Therefore, 5 MTPY could be
able to produce about 187 MLPY of ethanol. The estimate produc-
tivity can make adequate contribution to meet the AEDP target.
This scenario would reduce 1.1 MTPY raw material of cassava to be
 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641 639

Fig. 6. Material flow cost accounting of the ethanol production system in the improved scenario.

used to produce ethanol to meet AEDP target as shown in Table 3;
cassava needed to produce ethanol will be reduced from 12.6 ± 0.82
MTPY to only 11.5 ± 0.75 MTPY.
If the improved scenario is expanded to the country scale, the
cassava pulp could be used for raw material in ethanol production
process instead of selling it for feedstock. This option could increase
the value chain of this by-product and would offset about 250 ± 23
million USD. Moreover, by using 5 MTPY cassava pulp for ethanol
production, 1.1 MTPY fresh cassava could be avoided. Therefore,
105 ± 23 million USD of this loss cost could be reduced. However,
100 percent diversion of cassava pulp to use for ethanol production
may affect livestock industry on driving up the cost of feed.
Therefore, more a detailed study needs to be conducted on the
trade-off to find the optimum percent of cassava pulp to be diverted
for ethanol production.
In Fig. 7, the improved scenario showed that the demand for
cassava root for ethanol production is 11.5 ± 0.75 MTPY, if we spare
cassava utilization for feed and for starch production the same as in
2012 with at 11.9 MTPY and 14.2 MTPY respectively without
reduction of exports. Total requirement of cassava would be
37.6 ± 0.75 MTPY. The yield improvements were targeted to in-
crease from 21 ton/ha to 31 ton/ha. This can increase the cassava
production to meet the feedstock requirement. Therefore, the AEDP
target could be met by increasing cassava plantation yield. More-
over, with the proposed scenario, 425 ± 79 million USD/y loss cost
would be offseted/reduced. All in all, AEDP target of 9 MLPD or 2700
MLPY can be achieved by producing 1973 MLPY of ethanol using
cassava with all modified plant using VHG-SSF process, producing
187 MLPY of ethanol using cassava pulp and producing 540 MLPY of
ethanol using molasses as shown in Fig. 7.
3.5. Sensitivity analysis
There are large uncertainties associated with the model when
expanding it to the broader scale. Sensitivity analysis was applied
here to evaluate the effect of parameters used to calculate material
flow and cost accounting in the model. The efficiency of trans-
forming cassava into ethanol (feedstock processing efficiency) is
the key parameter. This research use the average here 5.83 ton fresh
cassava root to produce 1 kL ethanol. This has a large effect on the
estimation of the quantity of ethanol produced. In Fig. 8, the graph
shows the effect of changing feedstock processing efficiency (±25%
from the average) to the total quantity of ethanol produced in the
ethanol production BAU and in the ethanol production improved
scenario that use both cassava root and cassava pulp for ethanol
production.
There are several parameters that affect loss cost value in the
ethanol production improved scenario. The results of sensitivity
analysis in Fig. 9 show that feedstock processing efficiency was the
most sensitive parameter to loss cost. The second most sensitive
parameter is cassava price. However, the energy price has small
effect to the loss cost compared to the previous parameters.

Table 3
Amount of feedstock use and ethanol production in BAU and in the scenario of proposed improvement option to meet the target.

BAU to meet the target

Technology Raw material Raw material (kg)/L ethanol Raw material use (MtPY) Ethanol produced (MLPY)

BAU Cassava chip and fresh cassava 5.83 ± 0.38 (in fresh cassava) 12.6 ± 0.82 2160

Scenario of proposed improvement option to meet the target

Technology Raw material Raw material (kg)/L ethanol Raw material Ethanol produced Loss cost reduced/offset (million USD)
use (MtPY) (MLPY)

Ethanol production with VHG-SSF technology Fresh cassava 5.83 ± 0.38 (in fresh cassava) 11.5 ± 0.75 1973 70 ± 23
Ethanol production using cassava pulp Cassava pulp 11.25 ± 0.45 (wet), 5 ± 0.6 187 250 ± 34 (from offsetting loss cost in
3.125 ± 0.625 (dry) starch production)
105 ± 22 (from cassava avoidance)
Total 2160 425 ± 79
640 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641

Fig. 7. Proposed allocation of raw material in the scenario to meet the AEDP target.

Fig. 8. Sensitivity of feedstock processing efficiency to quantity of ethanol produced in ethanol production BAU and in the ethanol production improved scenario.

Fig. 9. Sensitivity of feedstock processing efficiency, material and energy cost to loss cost in the ethanol production improved scenario.
 N. Jakrawatana et al. / Journal of Cleaner Production 134 (2016) 633e641
4. Conclusions
Under the current system ethanol production incurs more loss
cost than starch production because it is associated with several
conversion processes. Therefore, a number of technologies and
innovations are required to reduce loss and increase productivity in
ethanol production. Although innovation for productivity
improvement could have high investment cost, the loss cost that
can be reduced can make it more cost effective.
Ethanol production from cassava pulp can also play an impor-
tant role in helping meet the AEDP target and utilizing by-product
from starch processing, increasing its value and offsetting loss cost
from the starch production plant. This option can aid the reduction
of cassava feedstock amount required for producing ethanol for
AEDP target. Therefore, the yield improvement option can satisfy
feedstock requirement without the need for plantation land
expansion. MFCA allocates the cost to both negative product cost
(material loss cost) and positive product cost (product that can be
sold) in order to indicate the loss in economic value. Loss cost and
possible loss cost reduction recognized by applying MFCA can be
advantageous information for decision maker in government and
private sectors to justify the cost effectiveness of their subsidies.
Additionally, it allows them to consider and take into account to
invest in implementing the technologies to reduce loss and in-
crease productivity or recover by-products to meet the target. Thus,
more research funding and investment from the government in
these technologies should be encouraged and more detailed tech-
nology assessment should be conducted.
Nevertheless, there are some limitations and cautions to
applying the MFCA method to materials that contain water and are
associated with a large amount of water for processing. Allocation
on wet mass basis may lead to misunderstanding of the results.
Therefore, if there is detailed data available, mass allocation on a
dry basis should be applied. However, if mass allocation on a wet
basis was used, precautionary principle should be kept in mind
when applying improved measures (e.g. improvement in water
issue) in order to avoid other main material loss (e.g. starch loss)
and avoid reduction of the yield and/or quality of the product.
Acknowledgement
This work was financially supported by the National Science and
Technology Development Agency (Grant no. P-12-01003), Thailand
under the Research Network for LCA and Policy on Food, Fuel and
Climate Change.
References
Bank of Thailand, 2014. Ethanol Situation Report in 2013 and Trend in 2014.
NorthEastern Office, Bank of Thailand.
641
Birkeland, J., Schooneveldt, J., 2003. Mapping Regional Metabolism: a Decision
Support Tool for Natural Resource Management. Land and Water Australia,
Canberra.
Blottnitz, H.V., Curran, M.A., 2007. A review of assessments conducted on bio-
ethanol as a transportation fuel from a net energy, greenhouse gas, and envi-
ronmental life cycle perspective. J. Clean. Prod. 15, 607e619.
Chavalparit, O., Ongwandee, M., 2009. Clean technology for the tapioca starch in-
dustry in Thailand. J. Clean. Prod. 17, 105e110.
Chompu-inwai, R., Jaimjit, B., Premsuriyanunt, P., 2015. A combination of material
flow cost accounting and design of experiments techniques in an SME: the case
of a wood products manufacturing company in Northern Thailand. J. Clean.
Prod. 108, 1352e1364.
DEDE., 2014. Curriculum Development Project on Renewable Energy (Ethanol and
Biogas). Department of Alternative Energy Development and Efficiency, Min-
istry of Energy, Bangkok.
FAO, 2010. Bioenergy and Food Security: the BEFS Analysis for Thailand. Food and
Agriculture Organization of the United Nation, Rome.
Gheewala, S.H., Silalertruksa, T., Nilsalab, P., Mungkung, R., Perret, S.R.,
Chaiyawannakarn, N., 2013. Implications of the biofuels policy mandate in
Thailand on water: the case of bioethanol. Bioresour. Technol. 150, 457e465.
Kim, J.S., Kim, B.G., Lee, C.H., Kim, S.W., Jee, H.S., Koh, J.H., Fane, A.G., 1997. Devel-
opment of clean technology in alcohol fermentation industry. J. Clean. Prod. 5
(4), 263e267.
LI, P., CHEN, H.T., ZHU, M.J., 2011. Ethanol Fermentation from Cassava Pulp by a
Novel Sequential Co-culture, pp. 413e417.
Liewa, W.H., Hassim, M.H., Ng, Denny K.S., 2014. Review of evolution, technology
and sustainability assessments of biofuel production. J. Clean. Prod. 71, 11e29.
METI, 2007. Guide for Material Flow Cost Accounting (Ver.1). Industrial Science and
Technology Policy and Environment Bureau.
Moriizumi, Y., Suksri, P., Hondo, H., Wake, Y., 2012. Effect of biogas utilization and
plant co-location on life-cycle greenhouse gas emissions of cassava ethanol
production. J. Clean. Prod. 37, 326e334.
Nguyen, T.L., Gheewala, S.H., Garivait, S., 2007. Energy balance and GHG-abatement
cost of cassava utilization for fuel ethanol in Thailand. Energy Policy 35,
4585e4596.
Nguyen, T.L.T., Gheewala, S.H., 2008. Fossil energy, environmental and cost per-
formance of ethanol in Thailand. J. Clean. Prod. 16, 1814e1821.
NIA., 2013. Factsheet: Demo Plant: Ethanol Production Using Cassava Pulp as
Feedstock. National Innovation Agency, Ministry of Science and Technology,
Bangkok.
NSTDA., 2009. Science, Technology and Innovation that are used for Strengthen the
Competitiveness of Thai Cassava Industry. Ministry of Science and Technology
of Thailand.
OAE, 2013. Background Information of Agricultural Economics. Center for Agricul-
tural Information, Office of Agricultural Economics, Ministry of Agriculture and
Co-operatives.
Piyachomkwan, K., 2011. Cassava Bioethanol Production. Siam City Hotel, Bangkok,
22e24 June 2011.
Roekklang, E., Pinitwachiraphong, T., 2010. The Increasing Efficiency in Tapioca
Production in Applying Lean Manufacturing Coarse Extraction Method of
Northeastern Starch (1987) Co.,Ltd. Thesis. Industrial Engineering Program,
Faculty of Engineering and Architecture Rajamangala University of Technology
Isan.
Siamphakdee, S., 2011. Ethanol Industry, an Open Market for Farmers. Thai Ethanol
Manufacturing Association.
Silalertruksa, T., Gheewala, S.H., 2009. Environmental sustainability assessment of
bio-ethanol production in Thailand. Energy 34, 1933e1946.
Sriroth, K., 2013. Cassava and Starch Technology Research Unit (SCST). In: Confer-
ence 2013. Suranaree University of Technology, Nakhon Ratchasima, 20 June
2013.
Sriroth, K., Piyachomkwan, K., Wanlapatit, S., Nivitchanyong, S., 2010. The promise
of a technology revolution in cassava bioethanol: from Thai practice to the
world practice. Fuel 89, 1333e1338.