001.docx
DOI: 10.3303/CET2189087
Paper Received: 25 May 2021; Revised: 26 August 2021; Accepted: 25 November 2021
Please cite this article as: Zailan R., Lim J.S., Sa'ad S.F., Jamaluddin K., Abdulrazik A., 2021, Optimal Biomass Cogeneration Facilities
Considering Operation and Maintenance, Chemical Engineering Transactions, 89, 517-522 DOI:10.3303/CET2189087
CHEMICAL ENGINEERING TRANSACTIONS
VOL. 89, 2021
A publication of
The Italian Association
of Chemical Engineering
Online at www.cetjournal.it
Guest Editors: Jeng Shiun Lim, Nor Alafiza Yunus, Jiří Jaromír Klemeš
Copyright © 2021, AIDIC Servizi S.r.l.
ISBN 978-88-95608-87-7; ISSN 2283-9216
Optimal Biomass Cogeneration Facilities Considering
Operation and Maintenance
Roziah Zailana,c, Jeng Shiun Lima,*, Siti Fatimah Sa’adb, Khairulnadzmi
Jamaluddinb, Abdulhalim Abdulrazikd
a Process Systems Engineering Centre (PROSPECT), Research Institute for Sustainable Environment (RISE), School of
Chemical and Energy Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 Johor Bahru, Johor,
Malaysia
b School of Chemical and Energy Engineering, Faculty of Engineering, Universiti Teknologi Malaysia, 81310 UTM Johor
Bahru, Johor, Malaysia
c Faculty of Civil Engineering Technology, College of Engineering Technology, Universiti Malaysia Pahang, Gambang
Campus, 26300 Kuantan, Pahang, Malaysia
d Faculty of Chemical Engineering Technology, College of Engineering Technology, Universiti Malaysia Pahang, Gambang
Campus, 26300 Kuantan, Pahang, Malaysia
jslim@utm.my
A development of biomass cogeneration facilities business comes with many challenges such as operational
compliance and budget constraints. Maintenance cost is often oppressing some portion of the annual budget.
Apparent outage costs due to loss of production during planned operation and maintenance are crucial. It
requires a strategic maintenance framework to ensure optimum performance of the biomass cogeneration
business. This paper presents a Mix Integer Linear Programming optimization model for the palm oil mill-based
cogeneration facility considering fuel cost, electricity cost, maintenance duration and maintenance interval. The
model aims to minimize total annual operation costs to ensure tenants can depend on reliable, uninterrupted
heat as they deserved from the heat supplier. The sensitivity analysis results determine the lowest annualised
cost had impacted by lower fuel and outsource electricity costs, six days maintenance duration, and 1.1
maintenance intervals.
1. Introduction
At present, biomass cogeneration facilities (BCF) have been increasingly accepted by industries as primary
energy sources since oil, natural gas, and coal were expected to be depleted in the next 40 to 50 y (Abbas et
al., 2020). The development of BCF mainly in the eco-industrial parks is mutual with an industrial symbiosis
approach (Zailan, 2020). It also promotes the sustainability of resources as the key pillar in the circular economy
concept (Misrol et al., 2020). Taken the Tanjung Langsat Biomass steam plant, Malaysia that relies on palm oil
mills (POM) biomass fuel such as empty fruit bunch (EFB), palm kernel shell (PKS), and mesocarp fibre (MSF)
(Zailan, 2021). The BCF generates and selling steam as its core business. Generally, the BCF installation comes
with challenges given as operational compliance and budget constraint. A reliable, uninterrupted steam supply
is important to satisfy the steam requirement of tenants to meet their daily operation at any cost. The problem
facing by the steam provider is the limitation of plant operators to plan for an optimal operation system including
the efficient operation and maintenance (O&M) schedule. Annual planned outages duration shall be budgeting
appropriately to ensure optimum yearly operation cost.
Several researchers have optimized operation costs concerning O&M costs in energy industries. It includes an
optimization framework for the maintenance program of a power station (Tam et al., 2007). The pellet production
used in the BCF had optimized to reduce the O&M costs (Rentizelas et al., 2014). In the POM biomass field,
Tan et al. (2020) developed an optimal POM complex concentrating on POM effluent. Hence, this work has
been conducted to fill in the research gap in a POM-based BCF development whereby a recent optimization
model has to be developed considering operational optimization and maintenance schedule time constraints.
The objective of this study is to develop an optimization model of POM-based BCF considering fuel cost,
517
mailto:lim@utm.my
electricity cost and maintenance cost using General Algebraic Modelling System (GAMS) software (GAMS,
2021). The optimization model also aims to minimize total annual operation costs and to ensure tenants can
depend on reliable, uninterrupted steam as they deserved from the BCF. Through this model, managers and
engineers can direct the BCF to create lower annual operation costs. Besides, the planned annual maintenance
program has been scheduled to investigate major faults or failures of the BCF system.
2. Methodology
The initial step of the methodology is the data collection for the generation of steam/power for selected BCF.
Secondary data were acquired from previous studies, and some were collected during a field survey at the
selected BCF including fuel, electricity, operating characteristics, tenant heat demand, and O&M. The model
requires a balanced supply and demand relationship between the BCF and the tenants to meet steam demand.
Next, the superstructure has been developed as in Figure 1 and followed with the formulation of the optimization
model. In this study, given a component of multi-POM biomass to operate a BCF. Fuel input for the BCF had
opted POM biomass that is EFB, PKS, and MSF. Steam generation consists of high-pressure (HP) steam that
will distribute to the respective tenants. Some portion of the HP steam will be let down to meet their medium
pressure (MP) and low-pressure (LP) steam usage. In turn, the remaining HP distributes to the steam turbine
for power generation. The generated electricity utilized by the BCF and could reduce the dependence on
electricity from the grid system. A case study of POM-based BCF was modelled followed by a sensitivity
analysis.
Figure 1: The superstructure of palm oil mill biomass cogeneration system
Objective function is to minimize the annual operation cost of the BCF operation through the optimal
maintenance outages frequencies and intervals of the BCF. The operation cost is a summation of variable costs;
fuel, boiler feedwater, outsource electricity generation, and maintenance. The mathematical model is as
expressed in Eq(1).
𝑀𝐼𝑁𝐴𝑁𝑁𝐶𝑂𝑆𝑇 = 𝐴𝑁𝐹𝑈𝐸𝐿𝐶𝑂𝑆𝑇 + 𝐴𝑁𝐵𝐹𝑊𝐶𝑂𝑆𝑇 + 𝐴𝑁𝑂𝑈𝑇𝐸𝐿𝐸𝐶𝑇𝐶𝑂𝑆𝑇 + 𝐴𝑁𝑀𝐴𝐼𝑁𝑇𝐶𝑂𝑆𝑇 (1)
The amount of boiler feedwater consumed in the boiler j, 𝑀𝐵𝐹𝑊𝑗 is determined knowing the amount of HP
steam generated including the percentage of water blown down from the boiler, 𝑏𝑑𝑝𝑒𝑟 according to Eq(2).Then,
it is calculated as a one-year, 𝑡𝑦𝑒𝑎𝑟 boiler feedwater supply, 𝐴𝑁𝑁𝐵𝐹𝑊 to the BCF as shown in Eq(3).
𝑀𝐵𝐹𝑊𝑗 = 𝑀𝐻𝑃𝑗 (1 − (
𝑏𝑑𝑝𝑒𝑟
100
))⁄ ∀𝑗 (2)
518
𝐴𝑁𝑁𝐵𝐹𝑊 = 𝑀𝐵𝐹𝑊 × 𝑡𝑦𝑒𝑎𝑟 (3)
The amount of fuel 𝑀𝐹𝑈𝐸𝐿𝑖,𝑗 entering the boiler j, is sufficient to generate a specified amount of HP steam. It
must be satisfied by the fuel availability, 𝑓𝑢𝑒𝑙𝑎𝑣𝑎𝑖𝑙𝑖 as shown in Eq(4).
∑ 𝑀𝐹𝑈𝐸𝐿𝑖,𝑗𝑗 ≤ 𝑓𝑢𝑒𝑙𝑎𝑣𝑎𝑖𝑙𝑖 ∀𝑖, 𝑗 (4)
An energy balance in the boiler component is as in Eq(5) which the amount of energy into the boiler j, 𝐸𝐼𝑁𝐵𝑂𝐼𝐿𝑖,𝑗
is equal to the fuel entering the boiler j, 𝑀𝐹𝑈𝐸𝐿𝑖,𝑗 and the net heating value of the fuel consumption, 𝑛ℎ𝑣𝑖 . Noted
that the exact energy generated from the fuel consumption relies on the boiler efficiency, 𝑏𝑜𝑖𝑙𝑒𝑓𝑓 .
𝐸𝐼𝑁𝐵𝑂𝐼𝐿𝑖,𝑗 = 𝑀𝐹𝑈𝐸𝐿𝑖,𝑗 × 𝑛ℎ𝑣𝑖 × 𝑏𝑜𝑖𝑙𝑒𝑓𝑓 ∀𝑖, 𝑗 (5)
Energy flows out from the boiler j, 𝐸𝑂𝑈𝑇𝐵𝑂𝐼𝐿𝑖,𝑗 is determined by the amount of HP steam generated and the
enthalpy difference between boiler feed water, ℎ𝐵𝐹𝑊 and HP steam, ℎ𝐻𝑃.
𝐸𝑂𝑈𝑇𝐵𝑂𝐼𝐿𝑖,𝑗 = 𝑀𝐻𝑃𝑗 × (ℎ
𝐻𝑃 − ℎ𝐵𝐹𝑊) ∀𝑖, 𝑗 (6)
At the tenant’s demand side, Eq(7) implied that the supply of steam 𝑀𝑃𝐶𝑝,𝑘 must greater than heat demand
ℎ𝑒𝑎𝑡𝑑𝑒𝑚𝑝,𝑘 of every tenant to ensure stable and uninterrupted steam supply.
𝑀𝑃𝐶𝑝,𝑘 ≥ ℎ𝑒𝑎𝑡𝑑𝑒𝑚𝑝,𝑘 ∀𝑝, 𝑘 (7)
To reduce the operation cost on the utility side, the BCF is practicing self-generation of electricity. The amount
of electricity generated by the steam turbine as given in Eq(8) and Eq(9). The boiler operation for the steam
generation process consuming electricity generated from the in-house power generator. In this model, electricity
requirement, 𝐸𝐿𝐸𝐶𝑇𝐷𝐸𝑀 only considers electricity consume by boilers, boiler control system, and steam turbine.
Eq(10) shows the total electricity demand by the BCF. In case of insufficient electricity, 𝑂𝑈𝑇𝐸𝐿𝐸𝐶𝑇 is outsourcing
from the grid system and calculated as Eq(11). Finally, Eq(12) is used to annualize the total electricity supplied
by the grid system.
𝐸𝐿𝐸𝐶𝑇𝐺𝐸𝑁 𝐻𝑃𝑀𝑃 = 𝑀𝑇𝑈𝑅𝐵𝐻𝑃 × (ℎ𝐻𝑃 − ℎ𝑀𝑃𝑇 ) × 𝑡𝑢𝑟𝑏𝑒𝑓𝑓/3600
(8)
𝐸𝐿𝐸𝐶𝑇𝐺𝐸𝑁 𝑀𝑃𝐿𝑃 = 𝑀𝑇𝑈𝑅𝐵𝑀𝑃 × (ℎ𝑀𝑃𝑇 − ℎ𝐿𝑃𝑇 ) × 𝑡𝑢𝑟𝑏𝑒𝑓𝑓/3600
(9)
𝐸𝐿𝐸𝐶𝑇𝐷𝐸𝑀 = 𝑏𝑜𝑖𝑙𝑒𝑙𝑒𝑐𝑡 + 𝑏𝑜𝑖𝑙𝑐𝑜𝑛𝑡𝑒𝑙𝑒𝑐𝑡 + 𝑠𝑡𝑒𝑙𝑒𝑐𝑡
(10)
𝑂𝑈𝑇𝐸𝐿𝐸𝐶𝑇 ≥ 𝐸𝐿𝐸𝐶𝑇𝐷𝐸𝑀 − (𝐸𝐿𝐸𝐶𝑇𝐺𝐸𝑁 𝐻𝑃𝑀𝑃 + 𝐸𝐿𝐸𝐶𝑇𝐺𝐸𝑁 𝑀𝑃𝐿𝑃)
(11)
𝐴𝑁𝑁𝑂𝑈𝑇𝐸𝐿𝐸𝐶𝑇 = 𝑂𝑈𝑇𝐸𝐿𝐸𝐶𝑇 × 𝑡𝑦𝑒𝑎𝑟
(12)
An O&M cost model deliberately used to assess the effectiveness of O&M at the typical BCF. This model was
developed by Tam et al. (2007) and adopted in this study. The following constraints shall consider where the
outage dimension cost, 𝑂𝐷𝐶 is crucial. It means the cost of loss of operation due to planned maintenance
outages. Eq(13) expressed the 𝑂𝐷𝐶 .
𝑂𝐷𝐶 = 𝑠𝑑𝑡 − 𝑛𝑜𝑡 × 𝑇𝐸𝐴𝑅𝑁 × 1 𝑀𝐼⁄
(13)
where, 𝑛𝑜𝑡 is system non-operating time, referring to the amount of time when the system is not operating due
to equipment failure. In this study, non-operating time is equal to zero. 𝑀𝐼 is maintenance interval (in a year),
𝑠𝑑𝑡 is system downtime for maintenance or planned outages days duration per outages. Meanwhile, 𝑇𝐸𝐴𝑅𝑁 is
the total daily earning without outages as calculate in Eq(14), where amount of steam generated multiply with
steam price, 𝑠𝑡𝑒𝑎𝑚𝑠𝑎𝑙𝑒 .
𝑇𝐸𝐴𝑅𝑁 = ∑ 𝑀𝐻𝑃𝑆𝑇𝐸𝐴𝑀𝑗
𝑗
× 24 × 𝑠𝑡𝑒𝑎𝑚𝑠𝑎𝑙𝑒 (14)
Next, resource dimension cost, R𝐸𝐷𝐶 or the cost needed for performing maintenance action typically consist of
human resource, 𝐻𝑈𝑀𝑅𝐸𝑆 and equipment and tools, 𝐸𝑄𝑈𝐼𝑃𝐶𝑂𝑆𝑇. The equation is express in Eq(15).
𝑅𝐸𝐷𝐶 = 𝐻𝑈𝑀𝑅𝐸𝑆 + 𝐸𝑄𝑈𝐼𝑃𝐶𝑂𝑆𝑇
(15)
519
Finally, the 𝐴𝑁𝑀𝐴𝐼𝑁𝑇𝐶𝑂𝑆𝑇 in Eq(16) represents sum of all two dimensions costs referring to the total annual
maintenance cost for the BCF.
𝐴𝑁𝑀𝐴𝐼𝑁𝑇𝐶𝑂𝑆𝑇 = 𝑂𝐷𝐶 + 𝑅𝐸𝐷𝐶 (16)
3. Case study
As for case study, the core business of the designated POM-based BCF is a steam generation and selling to
the tenants. It also commits electricity generation for the plant's in-house use. Thus, the steam and electricity
generation components are the main perspectives in this cost optimization model. This model is an extension
from the baseline case study in Zailan et al. (2020) where the power generation is newly introduced. The number
of boilers also increased to two units. All parameters and variables employed in this model are set hourly and
subsequently multiplied into the annual cost. Boiler fuel is accessible from the POM biomass supplier that is
EFB, MSF, and PKS. Table 1 presents data for cost, heating value, and fuel availability. Two units of water tube
boilers are operating continuously at the BCF. Each boiler generates 50 t/h of steam. The installed backpressure
steam turbines are HP-MP and MP-LP and able to generate electricity up to 1 MWh. The engagement of the
heat supplier-tenant concept is applied to execute this model. On the demand side, the amount of steam selling
to the tenants presented in Table 2.
Table 1: Palm oil mill biomass fuel data
Fuel Heating value (Mj/t)
(Hamzah et al., 2019)
Cost of fuel (MYR/t)
(Liew et al., 2017)
Availability (t/h)
(Md Jaye et al., 2016)
EFB 18,800 22 300
MSF 19,060 36 180
PKS 20,090 120 120
*1 MYR equals to 0.24 USD
Table 2: Selling steam to tenant
Tenant HP Steam (t/h) MP Steam (t/h) LP steam (t/h)
1 20 20 0
2 0 10 20
3 10 0 10
Energy consumption data for proposed components will determine how much electricity to outsource from the
grid system. The amount of electricity consuming by the boilers is 0.44 MWh while the steam turbine 1.6 MWh.
The baseline data for O&M to determine the optimality of this model is as in Table 3. Assigned equipment and
human resource costs were estimated based on the plant survey at the biomass cogeneration plant. The
planned maintenance is once a year with seven days outage as suggested by Myriad (2021) includes setting
up maintenance work to allow cooling down of a boiler prior to major cleaning process.
Table 3: Operation and maintenance data for annual maintenance program
Detail Description
Maintenance Duration (Baseline) (d) 7
Maintenance Interval (Baseline) (In a year) 1
Equipment Cost (MYR/y) 100,000
Human Resource cost (MYR/y) 30,000
*1 MYR equals to 0.24 USD
4. Results and discussions
The case study data were fitted into the developed MILP model and optimized using the CPLEX solver of GAMS
Studio 1.3.4. It was run on a personal computer using Mac OS X at the solution time 0.01 s. Model statistics
has a total of 67 constraints, 63 single variables, and 160 non-zero elements. The optimal operation of BCF
illustrates in Figure 2. Total fuel consumed by the BCF system is 32.06 t/h for each boiler and generates 90 t/h
of steam. The mass flow rate of EFB, MSF, and PKS that entering the boiler at optimal operation is 10.97 t/h,
10.82 t/h, and 10.27 t/h. Whereby the electricity generated from both turbines is 0.5 MWh. The electricity
generated will be supplied to the BCF components about 2.04 MWh. An insufficient electricity amount of 1.53
MWh is to be outsourcing from the grid system. Given the annual operating hours for the BCF is 7,056 h. Overall,
520
the annual electricity requirement is 14,394.2 MWh/y. Annual electricity generation is 3,556.2 MWh/y and
requires about 10,828 MWh/y outsource electricity. All tenants receive sufficient steam generated from the
boilers and supplementary steam from the turbine operation. Hence, the steam pass through the pressure
reduction valves, HP and MP reduction valves are negligible. Annualized costs attributed to the optimal annual
operation cost of this model are present in Table 4. Comparison with the baseline model without power
generation (90 t/h steam generation), a total of annualized costing is cheaper at MYR 32,113,292 (2.7 %). A
reflected minor difference would be an opportunity to generate power in the BCF to reduce dependency on-grid
system. Detail economic analysis suggested considering an investment of steam turbine and auxiliary
components to produce a quantifiable profit.
Figure 2: The optimal palm oil mill biomass cogeneration system
Table 4: Cost breakdown for optimal palm oil mill biomass cogeneration system
Item Value (MYR/y)
Annual fuel 26,299,382
Annual boiler feedwater 1,905,120
Annual outsource electricity 3,898,080
Annual O&M 886,000
Total 32,988,582
*1 MYR equals to 0.24 USD
Sensitivity analysis aims to examine the effect of the input variable on the performance of the base case
condition. According to the cost breakdown provided in Table 4, the highest cost of annual fuel and outsource
electricity make it significant to assess its impact on annualised cost. The sensitivity bound is varying from the
optimal values about ±10 %. Total annual fuel cost acquired from fuel cost variations (MYR 23,669,444 and
MYR 28,929,321) while the total outsources electricity cost varying with different electricity tariffs (MYR
324/MWh and MYR 396/MWh). Meanwhile, maintenance duration were tested for 6 and 8 d and interval 0.9
and 1.1 to assess impact of maintenance schedule. The sensitivity results portrayed in Figure 3, whereby the
impact of each parameter towards the changes in annualised cost is noticeable. An annualised cost is highly
impacted by the fuel cost (±8 %), electricity cost (±1.2 %), (maintenance duration (-0.8 % & 0.3 %), followed by
the maintenance interval (0.26 % & -0.27 %). Longer maintenance intervals (1.1) poses a significant reduction
of annualised cost. This condition is expecting to increase the effectiveness of an outage and reduce the failure
risk (Tam, 2007). Further studies are appealing to attain the trade-off between the profit and operation cost.
521
Figure 3: Sensitivity analysis of annualised cost of palm oil mill biomass cogeneration system.
5. Conclusion
An optimal operation of BCF ensure tenants received sufficient heat demand and practised self-generation
electricity for in-house use. The sensitivity analysis presented impacts from cost of fuel and electricity,
maintenance duration and maintenance interval to the annualised cost. The optimization model requires a trade-
off between the cost and profit of the BCF in a future BCF business model of selling steam and electricity. Future
optimization also open to explore ash management and the environmental impact of the BCF.
Acknowledgment
The authors thank funding from Universiti Teknologi Malaysia (UTM) via grant Q. J130000.2451.08G48 and R.
J130000.7851.5F388.
References
Abbas T., Issa M., Ilinca A., 2020, Biomass cogeneration technologies: A review, Journal of Sustainable
Bioenergy Systems, 10(1), 1-15.
GAMS, 2021, An introduction to GAMS, General Algebraic Modeling System <https://www.gams.com>
accessed 19.06.2021.
Hamzah N., Tokimatsu K., Yoshikawa K., 2019, Solid fuel from oil palm biomass residues and municipal solid
waste by hydrothermal treatment for electrical power generation in Malaysia: A review, Sustainability, 11(4),
1060.
Jaye I.F., Sadhukhan J., Murphy R.J., 2016, Renewable, local electricity generation from palm oil mills: a case
study from Peninsular Malaysia, International Journal of Smart Grid and Clean Energy, 44, 106-111.
Liew W.L., Muda K., Azraai M., Affam A.C., Kheang S., 2017, Agro-industrial waste sustainable management–
a potential source of economic benefits to palm oil mills in Malaysia, Journal of Urban and Environmental
Engineering, 11(1), 108-118.
Misrol M.A., Alwi S.R.W., Lim J.S., Manan Z.A., 2020, An optimal water-waste nexus for an eco-industrial park,
Chemical Engineering Transactions, 81, 643-648.
Myriad, 2021, Myriad Guide to Servicing Biomass Boiler, Myriad Heat & Power Products
<myriadservice.co.uk/biomass-servicing-requirements> accessed 19.06.2021.
Rentizelas A.A., Tolis A.I., Tatsiopoulos I.P., 2014, Optimisation and investment analysis of two biomass-to-
heat supply chain structures, Biosystems Engineering, 120, 81-91.
Tam A.S., Price J.W., Beveridge A., 2007, A maintenance optimisation framework in application to optimise
power station boiler pressure parts maintenance, Journal of Quality in Maintenance Engineering, 13(4), 364-
384.
Tan Y.D., Lim J.S., Alwi S.R.W., 2020, Optimal design of integrated palm oil complex with palm oil mill effluent
elimination strategy, Chemical Engineering Transactions, 78, 1-6.
Zailan R., Saad S.F., Jamaluddin K., Alwi S.R., Shiun L.J., Tan Y.D., Mohamad S.N., Lawal M., Misrol M.A.,
Kuba N.S., Reba M.N., 2020, Towards Eco-Industrial Park in Malaysia: Promising Opportunities, Challenges
and Regulator Roles, International Journal of Psychosocial Rehabilitation, 24(1), 562-570.
Zailan R., Lim J.S., Manan Z.A., Alwi S.R., Mohammadi-ivatloo B., Jamaluddin K., 2021, Malaysia scenario of
biomass supply chain-cogeneration system and optimization modeling development: A review, Renewable
and Sustainable Energy Reviews, 148, 111289.
-8. 00 -6. 00 -4. 00 -2. 00 0.00 2.00 4.00 6.00 8.00
Mai ntenan ce Interval
Mai ntenan ce Duratio n
Electri city Cost
Fuel Cost
Percentage of Change in Annualised Cost, %
Var iation in par ameter (-10%) Var iation in par ameter (+10%)
522