Energy Conversion and Management 106 (2015) 1071–1081

Contents lists available at ScienceDirect

Energy Conversion and Management

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

Cost-benefit assessment of energy storage for utility and customers:

A case study in Malaysia
Kein Huat Chua ⇑, Yun Seng Lim, Stella Morris
Lee Kong Chian Faculty of Engineering and Science, Universiti Tunku Abdul Rahman, 43000 Kajang, Malaysia

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

Article history: Under the existing commercial framework of electricity in Malaysia, commercial and industrial cus-
Received 15 July 2015 tomers are required to pay for the peak power demand charge every month. Usually, the peak demand
Accepted 14 October 2015 charge can contribute up to 30% to their electricity bills due to the use of open-cycle gas power plants
that deliver expensive electricity to the customers. Therefore, alternative means are sought after in order
to reduce the peak demand for the customers. Distributed small-scaled energy storage can offer a good
Keywords: option to reduce the peak. This paper aims to identify the financial benefits of the energy storage system
Energy storage
for utility companies and customers. An energy dispatch model is developed in HOMER to determine the
Open-cycle gas turbine
Peak demand
cost of electricity. The model considers the heat rates of power plants in calculating the costs of electricity
Battery under different regulatory frameworks of natural gas with various prices of battery components. Apart
HOMER from that, the cost-benefit for the customers under various electric tariff structures is evaluated. Four
Transmission and distribution battery storage technologies, namely lead acid, vanadium redox flow, zinc-bromine, and lithium-ion
are considered. The simulation results show that the storage system with lead acid batteries is more
cost-effective than other battery technologies. The customers can reduce their electricity bills with the
payback period of 2.8 years. The generation cost for the power system with energy storage is lower than
that without energy storage. Besides, the system with energy storage has lower greenhouse gas emissions
than that without energy storage. The deferral of the reinforcement of transmission and distribution
infrastructure can be achieved with the installation of energy storage at distribution network.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction
The inherent variability of the load demand increases the com-
plexities of modern power plant operations. The variable load
demand requires additional equipment to vary the power output
effectively which increases the system complexity and results in
the high cost of electricity production. Peak power plants are used
to cope with high demand during peak periods. They are required
to supply electricity for several hours, say 8 h, per day. To recover
the operating and maintenance costs as well as the capital costs of
the power plants within their lifespans, the electricity of the peak
power plants has to be more expensive than that of any base-load
plants [1]. Furthermore, these power plants have to operate in par-
tial loading that can reduce efficiency of the plants. Additional fuel
is needed for the peak power plants, hence making the price of the
electricity to be further increased. Therefore, utility companies
often charge the commercial and industrial customers at a
⇑ Corresponding author. Tel.: +60 3 90860288; fax: +60 3 90198868.
E-mail address: chuakh@utar.edu.my (K.H. Chua).
premium price based upon their maximum demand during the
billing cycle, in addition to the energy consumption charges.
The maximum demand is calculated by taking the demand over
the successive time period of 30 min and then multiplied by 2
[2]. However, this maximum demand charges increases the operat-
ing costs of the commercial and industrial sectors, hence affecting
their competitiveness in the market.
Malaysia aspires to become a fully developed nation to achieve
a self-sufficient industrialized nation by the year 2020. The vision
encompasses the economic prosperity, social and political stability,
world class education, and psychological balance. However, the
recent price hikes of electricity may alleviate the industrialization
process. In January 2014, the average electricity tariff in Malaysia
hiked 15% from the average rate of RM 0.3354/kW h (USD
0.0906/kW h) to RM 0.3853/kW h (USD 0.104/kW h) [3]. The hike
of electricity prices is part of the government’s strategy to stabilize
the country’s economy by reducing fuel subsidies for the power
sector. Moreover, the increment in electricity price is unavoidable
because the prices of fossil fuels continue to increase. Furthermore,
the continuous increment in the peak demand requires more new
power plants as well as the grid reinforcement.

http://dx.doi.org/10.1016/j.enconman.2015.10.041
0196-8904/Ó 2015 Elsevier Ltd. All rights reserved.
1072 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081

The price-based programs such as time-of-use (TOU), real-time simple installation and commission, low capital cost investment,
pricing (RTP), and demand response (DR) are introduced to reduce minimal space occupation and low maintenance and operating
the peak demand by shifting some of the peak demand to off-peak cost [19,20]. In addition, the distributed energy storage located
periods [4,5]. Relatively high scarcity prices during certain periods downstream from transmission lines can reduce the loading on
of time would provide the customers a motivation for voluntary the transmission system during peak periods, hence deferring the
demand curtailment. From the customer perspective, the price- upgrade of the transmission and distribution (T&D) and extending
based programs can provide a means of controlling their electricity the T&D equipment life [21].
bills because they will be rewarded with relatively low price dur- The authors of [22] carried out an economic assessment to
ing off peak periods [6]. The utilities will be benefited because they understand whether the energy storage system is financially viable
can avoid any significant thermal stress on their networks during or not to provide primary reserves and peak shaving in small iso-
peak hours. However, the lack of the smart metering and commu- lated power systems with the renewable energy sources. However,
nication infrastructure are the main barriers to the implementation the study was carried out with the assumption that the efficiencies
of the price-based programs in Malaysia [7]. The DR program of the gas-fired power plants are constant under the variation of
requires time-sensitive and automated equipment that could loads. It is known that the efficiencies of the power plants are
increase the electricity cost to the customers. Moreover, most of varied with the loads due to the changes in the heat rates of the
the customers may not be willing enough to switch off their air plants. Therefore, the assessment may not be complete without
conditioners in Malaysia during peak hours because such action considering the changes in the heat rates. Also, the authors of
can reduce their productivity, hence reducing their profits. [23] proposed an approach to evaluate the possible contribution
Photovoltaic systems can offer an alternative solution to the of the grid-connected energy storage to the operation cost,
reduction of the peak demand because it can supply power to generation investment, transmission investment, interconnection
the customers during peak hours [8,9]. However, the output of investment, and distribution investment. The findings showed that
the photovoltaic systems is highly intermittent due to its cloudy the energy storage system can bring benefits to the whole power
weather conditions in Malaysia. Hence, additional spinning systems. However, the authors did not specify the types of the bat-
reserves may be required from the utility to regulate network fre- tery technologies being used in the studies. As a result, the studies
quency, hence increasing the operating cost for the utility. may need to be improved by using the updated parameters of the
In the recent years, energy storage has emerged as a promising specific types of the battery technologies. Other research works
technology in cutting down the peak demand. Large-scale energy such as [24] and [25] showed the feasibility of using the energy
storage such as thermal storage, pumped hydro storage, fuel cell storage system for reducing the peak demand by using a dynamic
storage, compressed air storage, batteries, flywheel, ultra capacitor, programming method and a demand tracking management model.
and super conducting magnetic energy offer the similar functional- However, these studies did not consider the savings in the fuel cost
ity of the peaking power plants. Nevertheless, each of these tech- in order to justify the financial viability of the energy storage
nologies still has financial and technical barriers to be resolved system.
[10–15]. Many large-scale battery storage systems have been The authors of [26] developed a sizing algorithm for the storage
installed for the purposes of peak shaving and load shifting world- devices used in the residential buildings based on the storage
wide. Table 1 shows the worldwide large-scale battery storage sys- capacity, power capability, and a grid demand limit. The authors
tems for peak shaving and load shifting applications. of [27] proposed another sizing strategy for the energy storage in
Distributed small-scale energy storage is a good option for the the residential power distribution feeders with the PV systems to
customers to reduce their peak demand. The benefits of distributed reduce peak demand. The appropriate size of the energy storage
small-scaled energy storage are high portability, short setup time, system is important because it enables the customers to achieve
the successful reduction in the peak demands. However, the two
authors didn’t consider the investment cost of the energy storage
Table 1 system in their sizing methods. Therefore, the size of the energy
Worldwide large-scale battery storage systems for peak shaving and load shifting storage system determined by using the two methods may not
applications.
be able to achieve the maximum financial savings for the
Project Operational Location Battery System size customers.
dates type This study is very important to the developing countries like
MW MW h
Crescent electric 1987–2002 Carolina, Flooded 0.5 0.5
Malaysia because the findings can be disseminated to regulatory
membership USA cell, lead bodies, utility companies and customers in order to reduce the
cooperative acid electricity prices for the effective growth of economics in the
BESS a country. This paper presents the research work with the aim at
Sumitomo 2000 to Japan Vanadium 3 0.8
identifying the financial benefits of the energy storage system for
Densetsu present redox flow
office battery utility companies and customers in Malaysia. The savings in the
system a cost of electricity indicate the worth of the energy storage system
New York bus 2008 to New York, Sodium– 1.2 6.5 for the utility companies. The changes in the heat rates or efficien-
terminal present USA sulphur cies of the power plants with respect to the load profiles are con-
energy storage
sidered in the studies. Hence, the savings in the fuel cost or cost
systems a
ZBB energy 2005 to California, Zinc– 2 2 of electricity due to the energy storage systems can be accurately
corporation present USA bromine determined. In addition, the studies include various design options
battery with the appropriate parameters such as the prices of the energy
storage
storage, battery cycle life, and fuel costs with and without subsidy
systems b
Zurich battery 2012 to Dietikon, Li-ion 1 0.5 from the government in order to determine which battery tech-
energy storage present Switzerland nologies and design options allow the utility companies to achieve
system c the maximum financial gains. Apart from that, the energy storage
a
Data collected from [16].
is sized such that it can achieve a balance between the electricity
b
Data collected from [17]. bill savings and the investment cost on the energy storage, hence
c
Data collected from [18]. allowing the customers to achieve the maximum financial gains.
 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081 1073

The deferral cost of reinforcement of the transmission and distri- day. Combined-cycle gas turbines (CCGT) are used as intermediate
bution infrastructure and the reduction of CO2 emissions cost are power plants. The peak power generations are formed by open-
also evaluated. cycle gas turbine (OCGT), co-generation (Co-Gen), hydro, oil steam
The paper begins with a discussion on the power demand and turbine (ST-Oil) gas steam turbine (ST-Gas), and interconnection
generation in Malaysia in Section 2. The methodology for the with Thailand and Singapore. The generation fleet offers the grid
cost-benefit assessment is discussed in Section 3. Section 4 operator the flexibility to achieve economic dispatch of power
presents the assessments of cost-benefit using energy storage for plants while maintaining the stability and reliability of the power
utility and consumers. Finally, the paper is concluded in Section 5. supply.
Fig. 2 shows the electricity generation portfolios of Malaysia in
2013. The OCGT are dedicated as peak power generating units
2. Power demand and generation in Malaysia
because they are capable to be interconnected with the grid within
5–10 min from initiation to synchronization. It can be seen that the
The growth of Gross Domestic Product (GDP) is an important
OCGT contributes 11% to the overall installed capacity [31].
indicator for the power demand growth worldwide [28,29]. In
In Malaysia, the tariffs for the commercial and industrial
Malaysia, GDP is one of the key elements in determining future
consumers are still one of the lowest among the ASEAN countries
demand growth. Table 2 shows the comparison of periodical aver-
because the natural gas supply for the electricity sector is
age growth rates for GDP with electricity sales, generation, and
heavily subsidized by the government. However, the Malaysian
peak demand from 2012 to 2030. The predicted peak demand
government has set out a schedule of subsidy reductions for gas
growth increases the pressure and challenges to the utility to
via Subsidy Rationalization Program (SRP) to increase the gas price
build more power plants as well as reinforce the existing
for power and industrial by RM3/mmbtu every six months from
transmission and distribution (T&D) infrastructure.
The operation of the electric power system is carried out by com-
mitting and dispatching various types of generators in order to meet
the varying power demand. The dispatch of power plants is aided by
using simulation software taking into account many constraints
imposed by different generators, such as ramp rates, minimum
loading, and heat rate. Generally, power plants are classified into
base load power plants, intermediate power plants and peak load
power plants to meet the fluctuating power demand. Fig. 1 shows
the typical mixture of generation in Malaysia on 1 July 2013. It
can be seen that the coal-fired power plants are used as base-load
power plants whose power output are consistent throughout the

Table 2
Comparison of periodical average growth rates for GDP with electricity sales,
generation, and peak demand from 2012 to 2030.

Period Average period growth rates (% p.a.)

GDPa Electricity salesa Generationa Peak demanda
2012–2015 5.7 4.2 3.7 3.6
2016–2020 5.9 3.6 3.4 3.3
2021–2030 6.2 1.9 1.7 1.6
a
Data collected from [30]. Fig. 2. Electricity generation portfolios of Malaysia in 2013.

Fig. 1. Generation profile of Malaysia on 1 July 2013.

1074 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081

2010 to 2014 [32]. However, the actual implementation is behind Table 3

the proposed schedule. As of November 2014, the gas price for Key parameters for the project set up in HOMER.

power sector is RM15.2/mmbtu while the market price was around Discount rate (%) 6a
RM35/mmbtu to RM44/mmbtu. In this study, the sensitivity anal- Inflation rate (%) 2.1b
ysis is carried out for the case with and without gas subsidy [33]. Project lifetime (year) 20c
Optimization option Economic minimization
Dispatch strategy Cycle charging
3. Methodology a
Discount rate is according to ([35].
b
Data is collected from [36].
3.1. Simulation software c
Project lifetime is based on the lifetime of OCGT and CCGT from [35].

There are various software available for hybrid energy system

analysis such as HOMER, SOMES, Hybrid2, ARES, INSEL, and iHOGA
[34]. This study uses the HOMER software developed by the
National Renewable Energy Laboratory (NREL) of the United States.
Two systems are modeled in HOMER, namely the system with ESS
and the system without ESS. HOMER performs two main operating
tasks, namely optimization and sensitivity analysis. In the opti-
mization analysis, HOMER calculates the optimal system configu-
rations with the minimum total net present cost to meet the
constraints defined by users. In the sensitivity analysis, various
possible system configurations and factors that have the greatest
impact on the design and operation of the system can be evaluated.
The optimization and sensitivity analysis provide solutions to the
challenges arising from the large number of design options and
the variability of parameters, such as future fuel prices, load
growth, and increment of operation and maintenance cost.

3.2. Electrical system configurations Fig. 4. Daily electrical load profile.

The power system with ESS and without ESS are modeled in
HOMER as illustrated in Fig. 3. The system without ESS is formed
by a CCGT, an OCGT, and an electrical load. The CCGT acts as the
grid and OCGT acts as the peaking power plant. In the second
model, the same CCGT and load are used but the OCGT is replaced
by the energy storage system in order to study the effect of replac-
ing the OCGT on the cost of electricity and CO2 emission. If the grid
block-set is used instead of the CCGT and OCGT block-sets, then the
grid will supply all the power including the peak power throughout
the time and therefore be impossible to study the effect of replac-
ing the OCGT.
The economic inputs such as discount rate, inflation rate, and
project lifetime can be set in HOMER. Table 3 shows the key
parameters for the projects set up in HOMER.
HOMER allows users to model two dispatch strategies, namely
the load following and cycle charging strategy. In this study, the
cycle charging dispatch strategy is selected for the system with
CCGT and ESS. Under this strategy, the generator operates at full
capacity with surplus power for charging the ESS.
3.3. Electric load profile
The generation profile of the CCGT and OCGT on 1 July 2013 is
chosen as the electrical load profile for the investigation because it
exhibits the generic daily generation profile in Malaysia. Due to the
constraints of HOMER, the actual load profile is scaled down from
the peak of 7959 MW to 41.68 MW. The daily energy consumption
is 803.75 MW h. The daily electrical load profile is shown in Fig. 4.
3.4. System without ESS
The scaled down generation profile of the CCGT and OCGT is
illustrated in Fig. 5. It can be seen that the CCGT is operated at

Fig. 3. Schematic of electrical system in HOMER (a) system without ESS; (b) system with ESS.
 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081 1075

Fig. 5. The scaled down generation profile of the CCGT and OCGT.

the part load during the periods of 0:00–10:00, 17:00–20:00, and
21:00–23:00. The CCGT is operated near to its full load during
the periods of 10:00–17:00 and 20:00–21:00. The OCGT begins to
operate during the periods of 10:00–21:00.
The operating generation at part load requires additional fuel
due to the reduce deficiency of the plants. In addition, non-
steady state operation can increase operation and maintenance
cost [37]. Heat rate is frequently used to indicate the fuel consump-
tion rate for specific levels of power plant output. It is defined as
the amount of energy used by generator to generate 1 kilo-
watt hour (kW h) of electricity. In HOMER, user can input the fuel
consumption rate to generate the electrical efficiency curve.
HOMER assumes that the fuel curve (F) is a linear line as follows:
F ¼ F 0 Y gen þ F 1 P gen
where F0 is the fuel curve intercept coefficient in units/h/kW, F1 is
the fuel curve slope in units/h/kW, Ygen is the rated capacity of
the generator in kW, and Pgen is the electrical output of the genera-
tor in kW. The generator’s electrical efficiency (ggen) is defined as
the electrical energy output divided by the chemical energy of the
fuel input as follows:
3:6P gen
ggen ¼ ð2Þ
mfuel LHV fuel
where mfuel is the mass flow rate of the fuel in kg/h and LHVfuel is the
lower heating value of the fuel in MJ/kg. The factor of 3.6 arises
because 1 kW h is equal to 3.6 MJ. Fig. 6 shows the heat rate and
the electrical efficiency curve of the CCGT and OCGT with respect
to its loading conditions [38].
Operation and maintenance (O&M) expenses consist of fixed
operation and maintenance (FOM) as well as variable operation
and maintenance (VOM) [39]. FOM includes wage, monthly fees
under pertinent operating agreements, plant support equipment,
routine preventive maintenance, general and administrative
ð1Þ expenses, etc. VOM consists of air filter replacements, water treat-
ment expenses, catalyst replacements, major overhaul mainte-
nance, etc. The capital costs are the investments in equipment
and plant to provide physical infrastructure, generation hardware,
and plant auxiliary services. On the other hand, fuel costs are vari-
able costs that include purchase, transportation and storage of the
source of energy. The key parameters of CCGT and OCGT for cost
analysis are summarized in Table 4.

Fig. 6. (a) Heat rate and (b) electrical efficiency curve of CCGT and OCGT with respect to its loading conditions.
1076 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081

Table 4 under US$ 12/tons CO2 [45]. Although Malaysia doesn’t have any
Key parameters of CCGT and OCGT. carbon emissions trading scheme at this moment, it is expected
Technology CCGT OCGT to be implemented in the future. In this study, the carbon price
Plant life time (year) 20 a
20a of US$ 12/tons CO2 is adopted for the assessment.
Efficiency (%) 52–60b 35–42b
Capital cost (RM/kW)c 7752b 3600b 3.7. Reinforcement of the T&D infrastructure
FOM (RM/kW-year)c 117.623b 27.158b
VOM (RM/MW h)c 25.086b 57.165b
Capacity (MW) 36.5 5.18 Utilities are required to enhance the T&D facility when the elec-
Capital (RM million)c 283 19 tricity demand approaches its capacity. The reinforcement of the
O&M (RM/h)c 1406 328 T&D infrastructure depends on the growth of demand and peak
a
Data is collected from ([35]. demand, aging of infrastructure and the upgrade with new tech-
b
Data is collected from [40]. nology. Installation of ESS at distribution level or customer-end
c
The conversion rate of USD to RM is 1:3.7. could results in deferral of T&D reinforcement. During peak
demand periods, the ESS can deliver power to consumers and
hence reduce the power to be transmitted by the transmission
Table 5 lines. This can defer the need for high investment costs in new
Key parameters for LA, VRB, ZnBr and Li-ion. T&D facilities such as substations, transformers, protective equip-
ment, capacitors, cables, and line support towers. In addition, the
Technology DOD Round trip Cycle life Cost of energy
(%) efficiency (%) (RM/kW h) T&D facilities’ life could be extended because these facilities are
operated at a lower temperature [46]. T&D upgrade deferral is esti-
Low High Low High
mated based on the peak demand projection, the assumption of 5%
LA 50a 85b 500c 2800d 185b 1147b
ESS penetration rate, and 5% ESS growth rate from 2015 to 2030.
VRB 100a 85b 12,000c 13342c 647.5b 3700b The growth projection of the peak demand and ESS from 1990 till
ZnBr 100a 75b 1500c 2000c 740b 2220b 2030 is shown in Fig. 7 (Energy Commission, 2013). With the
Li-ion 85 a
90 b
1000 c
10,000 c
2220 b
9250b assumed growth rate for the ESS, it is expected to achieve 7.4% of
a
reduction in overall peak demand growth by 2030.
Data is collected from [41].
b The T&D avoided cost values can be estimated based on either
Data is collected from [16].
c
Data is collected from [42]. the historical annual marginal T&D investment or the future T&D
d
Data is collected from [43]. investment at specific sites. Based on the US experience, the typical
values for annual T&D upgrade deferral fall within the range of
USD55/kW–USD120/kW [47]. In this study, annual T&D upgrade
3.5. System with ESS deferral is assumed to be in the range of RM203/kW–RM444/kW.

The system with ESS consists of a CCGT, electric load profile,
and the ESS that is formed by a converter and batteries. Once the
OCGT is replaced by the energy storage, the heat rate of OCGT
and the fuel cost are not considered in this system. However, the
parameters required for this model are the depth of discharge
(DOD), the efficiency, the cycle lives and the cost of energy of the
energy storage system. Four battery technologies are chosen for
the investigation of system with ESS, namely lead acid (LA), vana-
dium redox flow (VRB), zinc–bromine flow (ZnBr), and lithium-ion
(Li-ion) battery. The key parameters for these batteries are summa-
rized in Table 5.
The PCS is formed by an inverter and a rectifier for the conver-
sion of alternating current (AC) to direct current (DC) or vice versa.
Batteries will be charged and discharged according to the dispatch
strategy in HOMER via the PCS. The price range of the PCS falls
between RM740 and RM5550 with lifetime of 20 years and effi-
ciency of 90% [16].
3.8. Cost-benefit analysis for the customer
ESS installed at end-users’ premises can provide significant ben-
efits such as reduction of the peak demand surcharge, avoidance of
financial losses due to power outage, and improvement of power
quality. In this study, the reduction of the peak demand surcharge
is investigated. In Malaysia, commercial and industrial sectors
were the largest electricity consumers, which consume 78% of total
electricity generated in the year 2012 [48]. These two sectors are
the main contributors to the economy as well as peak demand.
The installation of ESS for peak demand mitigation on these sectors
has significant advantages compared to that of other sectors. The
financial analysis focuses on the four tariff category for commercial
and industrial customers. Table 6 shows the tariff rate for the four
categories, namely C1, C2, E1 and E2.

3.6. Carbon dioxide emissions

Various carbon emissions trading schemes have been intro-

duced throughout the world to combat climate change. For exam-
ple, the European Union Emissions Trading Scheme (EU ETS) has
been established in 1 January 2005 to reduce greenhouse gas emis-
sions cost-effectively. The EU ETS covers more than 11,000 power
stations and industrial plants across 31 countries and introduces
emission allowances which can be traded as needed. Hence, power
producers have to include this cost to their marginal production
cost and increase the electricity price in wholesale power markets
[44]. The carbon dioxide (CO2) emissions price between schemes
differ from each other over a significant range, from US$ 1/tons
CO2 in the Mexican carbon tax up to US$168/tons CO2 in the Swed-
ish carbon tax. The prices in emission trading schemes is clustered Fig. 7. Growth projection of the peak demand and ESS from 1990 till 2030.
 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081
Table 6
Tariff rates for different categories of commercial and industrial customers.
Tariff C1a C2b
Peak [RM (USD)/kW h] – 0.365(0.10)
Off-peak – 0.224(0.06)
[RM (USD)/kW h]
Total energy 0.365(0.10) – 0.337(0.09)
[RM (USD)/kW h]
Maximum demand 30.3(8.19) 45.1(12.19)
[RM (USD)/kW]
a
C1 represents the medium voltage general commercial [3].
b that the batteries are charged and discharged one cycle during
C2 represents the medium voltage peak/off-peak commercial [3].
c
E1 represents the medium voltage general industrial [3].
d
E2 represents the medium voltage peak/off-peak industrial [3].
Maximum demand is computed based on the highest energy
used in kW h within any consecutive period of 30 min in a month,
and this highest energy is multiplied by two to get the maximum
demand. This duration of time is defined as the Demand Evaluation
Period (DEP). The electricity bill (CBill) for commercial and indus-
trial customers is calculated as follows:
Z t end Z tB
C Bill ¼ PðtÞ  C EU  dt þ 2PðtÞ  C MD  dt
t start tA
where tstart and tend are the periods of time during the billing cycle
in a month, P(t) is the power demand during the time interval, tA
and tB are the DEP in the same month, CEU is the charge of the
energy usage, and CMD is the maximum demand charge. Commer-
cial and industrial consumers can use ESS to shave the maximum
demand during peak period to reduce their electricity cost. For
the consumers with tariff rates of C1 and E1, the monthly savings
(Smth) can be calculated as follows:
Smth ¼ ðPshav e  C MD Þ  ðEES  ð1  gbat Þ  C EU Þ
where Pshave is the peak shaved by the distributed energy storage
system, EES is the energy used for peak shaving, and gbat is the round
trip efficiency of the batteries. For the consumers with tariff rate of
C2 and E2, there are two different rates for peak and off-peak periods.
Hence, the monthly savings (Smth) can be calculated as follows:
Smth ¼ ðPshav e  C MD Þ þ ðEES  C P Þ  ðEES  ð2  gbat Þ  C OP Þ
where CP and COP are the charge of energy usage during peak and
off-peak periods, respectively.
The two most important components to be considered in
achieving the cost-benefit of the ESS in peak shaving are the size
of battery bank and the power converter. Improper sizing of the
Fig. 8. Typical daily load profile for one of the buildings at UTAR.
1077
battery bank and power converter may incur additional cost to
the ESS. The optimal size of the ESS can be obtained from the com-
E1c E2d parisons of the cost of ESS and the electricity bills. Firstly, a typical
– 0.365(0.10) daily load profile of a commercial building is analyzed. Fig. 8 illus-
– 0.219(0.06) trates a typical daily load profile for one of the buildings at Univer-
siti Tunku Abdul Rahman (UTAR), Malaysia. It is found that the
– peak demand is 1084 kW and the daily energy consumption is
29.6(8) 37.0(10)
13.5 MW h.
Lead acid battery is chosen for the cost-benefit analysis because
it is the most economic viable storage technology. It is assumed
Monday to Friday, throughout a year. By taking into account the
Malaysian national holidays of 13 days, the ESS will operate
248 days in a year. The key assumptions for the ESS are summa-
rized in Table 7.
Round trip efficiency and cycle life of the ESS are important
parameters to be considered in the cost-benefit analysis because
it affects the overall cost of the system. The ratio of energy dis-
charged from the storage system to the energy charged into the
storage system is known as round trip efficiency. Cycle life of the
batteries is an important cost driver that indicates the number of
discharges before the replacement of the batteries. The battery
DOD has a significant impact on batteries cycle life. The relation
ð3Þ between the cycle life and the DOD appears to be logarithmic in
which the cycle life goes up exponentially with decrement of the
DOD. The cost of batteries (CBat) can be calculated by as follows:
DOD
C Bat ð$Þ ¼ U Bat ð$=kW hÞ  ER ðkW hÞ  ð6Þ
gbat
where UBat is the unit cost of the batteries in kW h, ER is the rated
capacity of the batteries, gBat is the round trip efficiency of the bat-
tery. The cost of power conversion units (Cconv) is as follows:
ð4Þ
C Conv ð$Þ ¼ U Conv ð$=kWÞ  PR ðkWÞ ð7Þ
where UConv is the unit cost of the power conversion units in kW,
and PR is the power rating of the converter. The cost of the ESS
(CES) is as follows:
ð5Þ C ES ¼ C Bat þ C Conv ð8Þ
4. Results and discussions
This section presents the results of the analysis carried out
using HOMER. The cost-benefit of using ESS for utility is evaluated
based on the generation cost, deferral cost of reinforcement of the
T&D infrastructure, and the reduction of CO2 emissions cost. The
analysis also compares the cost of electricity for the case with
gas subsidy and without gas subsidy. For commercial and indus-
trial customers, the economic viability of using ESS to reduce peak
demand is evaluated via the electricity bill savings as well as pay-
back periods.
Table 7
Key assumptions for the lead acid-based ESS.
Cost of converter [RM (USD)/kW] 740(200)
Cost of battery [RM (USD)/kW h] 185(50)
Round-trip Efficiency (%) 90
Efficiency of the converter (%) 85
Depth of discharge 50
Cycle life 2800
Battery life span 11.3
1078 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081

Table 8 expenditure in the year i, Ei is the electricity generation in the year

The operating parameters of the system without ESS. i, n is the expected lifetime of the system, and r is the discount rate.
Items CCGT OCGT The operating parameters of the system without ESS are summa-
Energy production (GW/year) 281 1.85
rized in Table 8.
Percentage of energy (%) 95.8 4.2 Table 9 shows the comparison of cost components for the sys-
Hours of operation 8760 3650 tem without ESS for the cases with and without subsidy. It can
Capacity factor (%) 87.9 27.2 be seen that the levelized COE for the case without subsidy is about
two times more expensive than that of the case with subsidy.

4.1. System without ESS 4.2. System with ESS

HOMER performs the simulation for the system with and with- The sensitivity analysis is carried out for a number of prospec-
out ESS. HOMER simulates the operation of these systems using the tive design configurations, namely the cycle life, PCS costs, battery
energy balance calculations for each year to meet the requirements price, and gas price. Table 10 shows the COE for the four battery
under the designated conditions, and estimates the cost of instal- technologies in the cases of with and without gas subsidy. It can
ling and operating the systems over the lifetime of the project. be seen that the COE of the lead acid batteries with high cycle life,
Based on the inputs such as the technology options, component low PCS and low battery cost in the case of with subsidy is lower
costs, resource availability, and the constraints defined by users, than that of the system without ESS. The VRB batteries with low
HOMER simulates all the possible system configurations and ranks and high cycle life, low PCS and low battery cost in the case of with
the results in ascending order of levelized cost of electricity (COE) subsidy is also lower than that of the system without ESS. The ZnBr
in RM/kW h. Sensitivity analysis is carried out to investigate the batteries with low PCS and low battery cost for low and high cycle
impact of the sensitive variables on the levelized COE such as fuel life has the same COE as the system without ESS.
prices, price of batteries and generators. The levelized COE is used Fig. 9 shows the range of COE for the highest and lowest battery
to measure the overall competitiveness of the two systems. The cost with low PCS and high cycle life. It can be seen that lead acid
levelized COE (LCOE) is as follows: and VRB batteries have COE lower than that of the system without
Pn C i þOi þF i
ESS while ZnBr batteries has the same COE as the system without
i¼1 ð1þrÞi ESS. With the current technology price, these three battery tech-
LCOE ¼ Pn ð9Þ
Ei nologies are the economically viable candidates to replace the
i¼1 ð1þrÞi
OCGT.
where Ci is the investment expenditure in the year i, Oi is the oper- The break-even price of these four battery technologies are
ations and maintenance expenditure in the year i, Fi is the fuel investigated. Fig. 10 shows the break-even price of LA, VRB, ZnBr,
and Li-ion batteries. It can be seen that the break-even price for
LA, VRB, ZnBr, and Li-ion batteries are RM296, RM844, RM740,
Table 9
and RM610, respectively.
Comparison of cost components for the system without ESS for the case of with and
without subsidy.
4.3. Reduction in carbon dioxide emissions
Items With subsidy Without subsidy
Net present cost [RM (USD) million] 947(255.9) 1846(498.9) In this study, the cost of CO2 emissions is investigated for the
Capital [RM (USD) million] 293(76.5) 293(76.5)
O&M cost [RM (USD) million] 180(48.6) 180(48.6)
system with and without ESS. The carbon price of US$ 12/tons
Fuel cost [RM (USD) million] 475(128.4) 1375(371.6) CO2 is adopted. The total power generation of CCGT and OCGT of
Salvage [RM (USD) million] 1.4(0.38) 1.4(0.38) Malaysia in a year is considered for the system without ESS. It is
Levelized COE [RM (USD)/kW h] 0.236(0.064) 0.46(0.124) assumed that the OCGT can be completely replaced by ESS for

Table 10
Cost of electricity for the four battery technologies with and without subsidy.

COE with subsidy [RM (USD)/kW h] COE without subsidy [RM (USD)/kW h]
Low battery cost High battery cost Low battery cost High battery cost
LA Low cycle life PCS Low 0.253(0.068) 0.383(0.104) 0.476(0.129) 0.605(0.164)
High 0.259(0.070) 0.389(0.105) 0.482(0.130) 0.612(0.165)
High cycle life PCS Low 0.233(0.063) 0.262(0.071) 0.456(0.123) 0.484(0.131)
High 0.240(0.065) 0.268(0.072) 0.462(0.125) 0.491(0.133)
VRB Low cycle life PCS Low 0.234(0.063) 0.265(0.072) 0.457(0.124) 0.487(0.132)
High 0.241(0.065) 0.271(0.073) 0.463(0.125) 0.494(0.134)
High cycle life PCS Low 0.234(0.063) 0.265(0.072) 0.457(0.124) 0.487(0.132)
High 0.241(0.065) 0.271(0.073) 0.463(0.125) 0.494(0.134)
ZnBr Low cycle life PCS Low 0.236(0.064) 0.252(0.068) 0.461(0.125) 0.476(0.129)
High 0.243(0.066) 0.258(0.070) 0.467(0.126) 0.483(0.131)
High cycle life PCS Low 0.236(0.064) 0.252(0.068) 0.461(0.125) 0.476(0.129)
High 0.243(0.066) 0.258(0.070) 0.467(0.126) 0.483(0.131)
Li-ion Low cycle life PCS Low 0.346(0.094) 0.722(0.915) 0.568(0.154) 0.944(0.255)
High 0.353(0.095) 0.728(0.197) 0.575(0.155) 0.951(0.257)
High cycle life PCS Low 0.259(0.070) 0.357(0.096) 0.481(0.130) 0.579(0.156)
High 0.265(0.072) 0.363(0.098) 0.488(0.132) 0.586(0.158)
 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081 1079

Fig. 9. Range of COE for the highest and lowest battery cost.

Fig. 10. Break-even price of (a) lead acid; (b) VRB; (c) ZnBr; (d) Li-ion.

the system with ESS. Table 11 shows the summary of the reduction
of CO2 emissions for the system with ESS. It can be seen that there
is a potential to reduce the CO2 emissions by 140,000 tons every
year. The assessment found that the CO2 emission factor for the
system with ESS is 2.5% less than that of the system without ESS.
4.4. Deferral of reinforcement of the T&D infrastructures
The T&D upgrade deferral cost for the power system with ESS is
evaluated based on the low and high cost from the range of RM203/kW
(USD55/kW) to RM444/kW (USD120/kW). Fig. 11 shows the
1080 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081

Table 11
Summary of the reduction of CO2 emissions cost for the system with ESS.

System With ESS

Reduction of CO2 emission (tons/year) 140,000
CO2 emission factor (kg/kW h) 0.407
Saving in the cost of CO2 (RM million) 9.0

Table 12
Monthly electricity bill for the system with and without ESS.

Cost Energy Maximum Monthly Payback Savings

[RM demand [RM bill [RM period [RM
(USD)] (USD)] (USD)] (year) (USD)]
Without ESS C1 147,420 32,845 180,265 – –
(39,843) (8877) (48,720)
C2 132,498 48,888 181,386 – – Fig. 12. Cost of the ESS and electricity bill.
(35,810) (13,213) (49,023)
E1 136,111 32,086 168,197 – –
(36,787) (8672) (45,459)
E2 128,988 40,108 169,096 – – total cost of the ESS and electricity bill exhibits a minimum that
(34,862) (10,840) (45,702) defines the optimal size of the converter. The optimal size of the
With ESS C1 147,939 29,512 177,451 5.9 2814 converter is 110 kW and the corresponding battery size is
(39,983) (7976) (47,960) (761) 395 kW h.
C2 131,480 43,927 175,407 2.8 5,979 Commercial and industrial customers can manage their electric
(35,535) (11,872) (47,407) (1616) energy cost by storing inexpensive energy price and discharging
E1 136,590 28,830 165,421 6.0 2777
the ESS during peak periods to reduce the maximum demand.
(36,916) (7792) (44,708) (750)
E2 128,010 36,038 164,048 3.3 5048 Table 12 shows the monthly electricity bills for the system with
(34,597) (9740) (44,337) (1364) and without ESS for the four categories of tariffs. Tariff C2 shows
the greatest potential savings for the system with ESS, followed
by E2, C1, and E1. The simple payback period for tariff C2 is the
projection of the T&D upgrade deferral based on the low upgrade shortest, namely 2.8 years.
deferral cost and high upgrade deferral cost. It can be seen that
the T&D upgrade deferral in 2030 is in the range of RM385 million
(USD104 million) to RM 840 million (USD227 million). 5. Conclusions

Energy storage is profoundly important in the power systems

4.5. Cost-benefit analysis of peak shaving for commercial and
industrial consumers
The objective of this section is to evaluate the costs and poten-
tial savings of energy storage in peak shaving for commercial and
industrial consumers. Intuitively it can be said that as the incre-
ment of the ESS capacity reduces the electricity bills. However,
the ESS increases the investment cost. An optimum cost of ESS
can be obtained by summing the cost of ESS with the electricity
bill. The electricity bill throughout the life span of ESS is estimated.
Fig. 12 shows the cost of ESS and electricity bill. It is found that the
due to its multifaceted functionality such as increasing deploy-
ment of renewable integration, reducing peak demands, and
enhancing electric grid reliability. In this paper, the cost-benefits
of introducing the energy storage into the grid for utility and cus-
tomers are presented. The evaluations show that the cost of elec-
tricity for the electrical system with lead acid batteries is lower
than that of the system without ESS. For the energy storage system
installed near the customer-ends, the reinforcement of T&D infras-
tructure can be deferred since the energy storage can reduce or
delay the need for additional T&D capacity. The electricity cost sav-
ings using lead acid battery storage system to reduce the maxi-
mum demand is economically viable for commercial and
industrial customers. The payback periods of the ESS ranged from
2.8 years to 6 years. It is expected that the price of batteries may
drop due to the increases of manufacturing scale, lower component
prices, and advancement of manufacturing technology. It is fore-
seen that the price of Li-ion batteries could drop to one-third of
its current price from about USD 600/kW h to USD200/kW h by
2020. This will provide a great opportunity for energy storage to
be adopted in mitigating the peak demand. Although some lead
compounds are very toxic, the lead recycling is a well-
established industry and hence the potential risk to the environ-
ment and human health are low.

Acknowledgments

This work is supported in part by the research Grant from

Malaysian Electricity Supply Industries Trust Account under the
Fig. 11. Projection of T&D upgrade deferral from 2015 to 2030.
Ministry of Energy, Green Technology and Water, Malaysia.
 K.H. Chua et al. / Energy Conversion and Management 106 (2015) 1071–1081 1081

References [23] Pudjianto D, Aunedi M, Djapic P, Strbac G. Whole-systems assessment of the

value of energy storage in low-carbon electricity systems. IEEE Trans Smart
Grid 2014;5(2):1098–109.
[1] Van den Bergh K, Delarue E. Cycling of conventional power plants: technical
[24] Oudalov A, Cherkaoui R, Beguin A. Sizing and optimal operation of battery
limits and actual costs. Energy Convers Manage 2015;97(June):70–7.
energy storage system for peak shaving application. In: Power Tech, 2007 IEEE
[2] Berg SV, Savvides A. The theory of maximum kW demand charges for
Lausanne; 2007. p. 621–5.
electricity. Energy Econ 1983;5(4):258–66.
[25] Purvins A, Papaioannou IT, Debarberis L. Application of battery-based storage
[3] Tenaga Nasional Berhad. Electricity Tariff Schedule; January 2014.
systems in household-demand smoothening in electricity-distribution grids.
[4] Faria P, Vale Z, Baptista J. Constrained consumption shifting management in
Energy Convers Manage 2013;65(January):272–84.
the distributed energy resources scheduling considering demand response.
[26] Leadbetter J, Swan L. Battery storage system for residential electricity peak
Energy Convers Manage Mar. 2015;93:309–20.
demand shaving. Energy Build 2012;55(December):685–92.
[5] He Y, Zhang J. Real-time electricity pricing mechanism in China based on
[27] Venu C, Riffonneau Y, Bacha S, Baghzouz Y. Battery storage system sizing in
system dynamics. Energy Convers Manage 2015;94(April):394–405.
distribution feeders with distributed photovoltaic systems. In: PowerTech,
[6] Allcott H. Rethinking real-time electricity pricing. Resour Energy Econ Nov.
2009 IEEE Bucharest; 2009. p. 1–5.
2011;33(4):820–42.
[28] Borozan D. Exploring the relationship between energy consumption and GDP:
[7] Bradley P, Leach M, Torriti J. A review of the costs and benefits of demand
evidence from Croatia. Energy Pol 2013;59(August):373–81.
response for electricity in the UK. Energy Pol Jan. 2013;52:312–27.
[29] Coers R, Sanders M. The energy–GDP nexus; addressing an old question with
[8] Li N, Hedman KW. Economic assessment of energy storage in systems with
new methods. Energy Econ 2013;36(March):708–15.
high levels of renewable resources. IEEE Trans Sustain Energy Jul. 2015;6
[30] Energy Commission. Peninsular Malaysia electricity supply industry outlook
(3):1103–11.
2013; 2013.
[9] Rüther R, Knob PJ, da Silva Jardim C, Rebechi SH. Potential of building
[31] Suruhanjaya Tenaga Malaysia. Electricity supply industry in Malaysia:
integrated photovoltaic solar energy generators in assisting daytime peaking
performance and statistic information; 2013.
feeders in urban areas in Brazil. Energy Convers Manage 2008;49(5):
[32] IISD. A citizens’ guide to energy subsidies in Malaysia. International Institute
1074–9.
for Sustainable Development; 2013.
[10] Amodeo SJ, Chiacchiarini HG, Solsona JA, Busada CA. High-performance
[33] Malaysian Gas Association. Malaysia: natural gas industry annual review –
sensorless nonlinear power control of a flywheel energy storage system.
2014 edition; 2014.
Energy Convers Manage 2009;50(7):1722–9.
[34] Sinha S, Chandel SS. Review of software tools for hybrid renewable energy
[11] Castillo A, Gayme DF. Grid-scale energy storage applications in renewable
systems. Renew Sustain Energy Rev 2014;32(April):192–205.
energy integration: a survey. Energy Convers Manage 2014;87(November):
[35] Koh SL, Lim YS, Morris S. Cost effective options for greenhouse gas (GHG)
885–94.
emission reduction in the power sector for developing economies — a case
[12] Ngamroo I. An optimization of robust SMES with specified structure H1
study in Sabah, Malaysia. Energies 2011;4(5):780–803.
controller for power system stabilization considering superconducting
[36] Anna B, Lucky L. Lessons learned: Malaysia’s 2013 fuel subsidy reform. Int Inst
magnetic coil size. Energy Convers Manage 2011;52(1):648–51.
Sustain Dev 2013.
[13] Sharma P, Bhatti TS. A review on electrochemical double-layer capacitors.
[37] Kumar N, Besuner PM, Lefton SA, Agan DD, Hilleman DD. Power plant cycling
Energy Convers Manage Dec. 2010;51(12):2901–12.
costs. National Renewable Energy Laboratory; 2012.
[14] Zhang L, Xiang J. The performance of a grid-tied microgrid with hydrogen
[38] Greg B, Debra L, Paul D. Impacts of renewable generation on fossil fuel unit
storage and a hydrogen fuel cell stack. Energy Convers Manage 2014;87
cycling: costs and emissions. NREL/PR-6A20-55828; 2012.
(November):421–7.
[39] Jose FP. The value of reliability in power systems – pricing operating
[15] Zhao P, Wang J, Dai Y. Thermodynamic analysis of an integrated energy system
reserves. Massachusetts Institute of Technology; 1999. July.
based on compressed air energy storage (CAES) system and Kalina cycle.
[40] U.S. Energy information administration. Updated capital cost estimates for
Energy Convers Manage 2015;98(July):161–72.
utility scale electricity generating plants; 2013.
[16] Poullikkas A. A comparative overview of large-scale battery systems for
[41] Abbas AA, Georgianne H, Aileen BC, Benjamin CK, Dan MR. DOE/EPRI electricity
electricity storage. Renew Sustain Energy Rev 2013;27(November):778–88.
storage handbook in collaboration with NRECA; 2015.
[17] Peter L. Demonstration of ZBB energy storage systems; 2012.
[42] Luo X, Wang J, Dooner M, Clarke J. Overview of current development in
[18] Koller M, Borsche T, Ulbig A, Andersson G. Review of grid applications with the
electrical energy storage technologies and the application potential in power
Zurich 1 MW battery energy storage system. Electr Power Syst Res 2015;120
system operation. Appl Energy Jan. 2015;137:511–36.
(March):128–35.
[43] Hoppecke. Installation, commissioning and operating instructions for vented
[19] Masoum AS, Deilami S, Moses PS, Masoum MAS, Abu-Siada A. Smart load
stationary lead–acid batteries; 2014.
management of plug-in electric vehicles in distribution and residential
[44] Jouvet P-A, Solier B. An overview of CO2 cost pass-through to electricity prices
networks with charging stations for peak shaving and loss minimisation
in Europe. Energy Pol 2013;61(October):1370–6.
considering voltage regulation. IET Gener Transm Distrib 2011;5(8):
[45] World Bank. State and trends of carbon pricing 2014: Washington (DC); 2014.
877–88.
[46] Tom B. Development of distributed generation for grid support and
[20] Nair N-KC, Garimella N. Battery energy storage systems: assessment for small-
distribution system infrastructure: a summary analysis of DG benefits and
scale renewable energy integration. Energy Build 2010;42(11):2124–30.
case studies. New York State Energy Research and Development Authority;
[21] Yan X, Zhang X, Chen H, Xu Y, Tan C. Techno-economic and social analysis of
2011. February.
energy storage for commercial buildings. Energy Convers Manage 2014;78
[47] Chris N, Rich S. US experience with efficiency as a transmission and
(February):125–36.
distribution system resource; 2012.
[22] Sigrist L, Lobato E, Rouco L. Energy storage systems providing primary reserve
[48] Suruhanjaya Tenaga Malaysia. Electricity supply industry in Malaysia:
and peak shaving in small isolated power systems: an economic assessment.
performance and statistical information; 2012.
Int J Electr Power Energy Syst 2013;53(December):675–83.