Pertanika J. Sci. & Technol.

27 (1): 225 - 246 (2019)

SCIENCE & TECHNOLOGY

Journal homepage: http://www.pertanika.upm.edu.my/

Reliability based Redundancy Assessment of a Cogeneration

Plant
Meseret Nasir* , Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof
Mechanical Engineering Section, Universiti Kuala Lumpur, Malaysia France institute, 43650 UniKL, Bandar
Baru Bangi, Selangor, Malaysia

ABSTRACT
Cogeneration systems are extensively used in Malaysia to produce power as a primary
source. However, in the event of cogeneration system failure, the customer or the client are
forced to use a redundancy to avoid power interruptions. There are two methods commonly
used as a backup in the cogeneration systems which are Generator set and public utility. In
order to choose the best redundancy for a particular cogeneration system, it is essential to
evaluate the economic benefit analysis by considering several factors such as Maximum
demand charge, installation cost and Discount interest. In the evaluation of economic
benefit, this study identifies the number of failure and associated downtime using reliability
and availability approach, and then present value method was applied. The result shows
that the usage of public utility as redundancy is beneficial if the cogeneration system
operates within five years period. However, if the cogeneration system operates more than
five years, generator set option would be a better option to minimize the total cost. This
research also addresses the effect of various factors such as installation cost, maximum
demand charge, fuel cost, discount interest rate and production capacity. In general, the
output of the research would be beneficial for the plant operator to select the appropriate
redundancy option based on the economic advantages.
Keywords: Cogeneration, gas turbine, public utility, redundancy, reliability

INTRODUCTION
Cogeneration is a system using a single
ARTICLE INFO
Article history:
source of fuel to generate electricity and
Received: 22 January 2018 waste heat(Chen et al., 2018). This waste
Accepted: 28 August 2018
Published: 24 January 2019 heat is useful to generate chilled water
E-mail addresses: or steam depending on the customer
meseret@unikl.edu.my/meseretreshid@gmail.com (Meseret Nasir)
drwmansor@unikl.edu.my (Wan Mansor Wan Muhamad) need(s) (Reshid et al., 2017a, 2017b). The
azizmaarof@unikl.edu.my (Raja Aziz Raja Maarof)
* Corresponding author performance of cogeneration system is

ISSN: 0128-7680
e-ISSN: 2231-8526 © Universiti Putra Malaysia Press
 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

linked with availability. In the cogeneration system, high availability is the most important
factor to avoid power interruption (Eti et al., 2007). In order to meet high availability, it is
necessary that all equipment/subsystems of cogeneration plant remain in upstate condition
for a longer duration of time. In other words, it is imperative for all subsystems to perform
satisfactorily during their expected life span.
The performance of a cogeneration system relies on the availability and operating
conditions of the equipment (Ramesh & Saravannan, 2011b). In a situation of a 1% reduction
in availability for a macro cogeneration system, this results in unplanned shutdowns which
causes about $500,000 loss of income (Meherwan, 2002). This economic loss has been
estimated to be about 30% of the total cost of electricity generated by the cogeneration
plant (Gräber, 2004; Lemma & Hashim, 2013). Such a proportion of expenditure is
considered higher than what is encountered in other industries. One of the main reasons
for the reduction of availability is failure. Failure is an unavoidable phenomenon which
can occur unexpectedly. When failure occurs, efforts are needed to maintain the system
and avert the associated risk due to it. The common practice to avert the associated risks
due to failure is using redundancy or back up system. Generally cogeneration system uses
two common redundancy options such as Generator set and public utility. Generator set
(Genset) refers to a gas turbine driven generator as a redundancy used in a cogeneration
plant. On the other hand, public utility refers to a cogeneration plant that taps electricity
from the national power grid to avoid power interruptions. When the cogeneration system
fails, the clients who are using the system as primary source of power are forced to use
redundancy. However, the associated cost using redundancy is very high.
There are four major reasons for the need of redundancy (Pham & Wang, 1996). First is
scheduled equipment maintenance. During the scheduled maintenance, the equipment will
cease to function, for preventive maintenance or the Overhaul. Therefore, there is a need
to have redundancy to continue supplying electric power to the user. The second reason is
equipment failure. In such case, the equipment or the system needs to undergo corrective
maintenance. During this time, the redundancy needs to supply the required demand to
the client. Third is demand variations as the cogeneration plant is highly dependent on
environmental conditions. The change in the environmental condition will cause fluctuation
in demand. In order to cope with such circumstance, redundancy should be integrated.
Finally, redundancy is needed due to special operating conditions. This condition refers to
the startup and shutdown of plants which may cause trip or unexpected failure.
The economic analysis of power generating system is more closely linked to system
availability and reliability analysis. This is because production interruption is one of the
major worry for plant owners (Dougan & Reilly, 1993; Lewis & Lewis, 1987; Vega et al.,
1998). The downtime cost in the power plant is very expensive apart from the maintenance
cost of the equipment. During the plant outage, power is purchased from other sources to

226 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

meet the demand of the utility system. Additionally, as Meherwan (2002) highlights, this
can be very costly in terms of operation. Most power purchase agreements have articles
which include maximum demand charge payments. This makes the power plant availability
crucial for the power generation system (Meherwan, 2002). Unplanned outages may
happen during peak generation seasons and usually result in significant losses. Richwine
(2004) estimated that forced outages cost from 3 to 4.5 times as much as planned outages.
Qiu et al. (2011) had established the failure cost model for power generating equipment
in which it estimated the failure cost considering the probabilistic nature of failure. The
probabilistic nature of failure was formulated using Weibull distribution. In their study, the
cost of repair was only estimated, however; the failure cost should include the downtime
cost which was caused by the failure.
Christiansen (2013) estimated unplanned outage events for 388 combined-cycle plants.
The author collected 15-years data over 3000 units of the combined cycle power plant. The
study identified the causes and durations of forced outages and unscheduled maintenance.
Furthermore, reliability and availability were established for each class of plant. The costs
to render the unit serviceable for each main outage were calculated, as were net revenues
lost due to unplanned outages. Furthermore, Grace and Christiansen (2013 ) estimated the
cost of unplanned outage events for combined-cycle systems. The study provided a detailed
listing of events that caused forced and unscheduled maintenance outages in combined-
cycle power plants, costs associated with such events and a quantified assessment of the
economic impact that such outage events could have on overall maintenance costs and
lost revenue.
Although several researches have been done on redundancy of power generating
cogeneration plant, there still is lacking of research regarding the suitability of redundancy
type to cogeneration plant. In fact, the choice of selecting the redundancy is normally
left to the user. Thus, to avoid additional capital expenditure requirements, public utility
is normally chosen to as redundancy, without considering the operation costs. From the
experience of the cogeneration operators, the use of public utility can be expensive due
to high cost of maximum demand charge that comes with, which is neglected in reviewed
papers on redundancy. Thus, this paper analyses various redundancy options by considering
several factors such as maximum demand charge, installation cost interest discount rate
and failure frequency.

METHOD
In order to evaluate the redundancy options of the cogeneration power system, to the main
requirement is to develop appropriate methodology which includes reliability, availability
and economic assessment. The flow chart of the methodology is presented in Figure 1.

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 227

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Figure 1. Methodology flow chart­­­­

Cogeneration System Configuration and Functional Block Diagram

A cogeneration system is a complex repairable system consisting of various subsystems
such as gas turbine, heat recovery steam generator, steam absorption chillers, electric
chillers and thermal energy storage systems which are linked in series, parallel or the
combination of both (Arora & Kumar, 1997). The general network and configuration of
the cogeneration system is depicted in Figure 3 Simulation block for power generation.
The two main systems in the cogeneration system are gas-turbine (GT) and heat recovery
steam generator (HRSG) (Shaaban et al., 2011; Soares et al., 2001). However, the
configuration of cogeneration system differs depending on the consumer requirements
and the site condition. Therefore, it is essential to integrate the cogeneration system with
steam absorption chiller (SAC), auxiliary gas boiler (AGB), thermal energy storage (TES),
and electrical chiller (EC) for the tropical region due to the need for high cooling loads.
This fundamental configuration is useful to mitigate the wasted energy and increases the
utilisation of the cogeneration system.

228 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

As observed in Figure 2, the gas turbine generates electric power and waste heat. The
electric power goes to electric chillers and to the customer for electricity usage. The electric
chiller uses electricity to produce the chilled water to supplement the high cooling load
during the peak hour. This chilled water may also be reserved in the thermal storage. The
waste heat generated from the gas turbine goes to heat recovery steam generator to produce
steam. This steam is used for process heating in the steam absorption chillers. Finally, the
chilled water will be supplied to the customer.

Figure 2. System configuration of cogeneration

Data Acquisition
The required data to develop the models are failure, repair and cost data. The failure
and repair data are commonly used to develop the availability and reliability model.
Unfortunately, the failure and repair data are scarce in many cases due to improper
documentation of maintenance data (Louit et al., 2009). Therefore this study utilizes
production data instead of maintenance data in the evaluation of availability and reliability.
This is because production data is abundantly available. The operational hourly production
data were collected from the plant historical production data and online observation to
develop characterization of the cogeneration system for the period of five year. The collected
data was filtered to exclude the holidays and schedule maintenance which was identified
using calendar and the plant maintenance schedule. The reason is during the holiday or
schedule maintenance, the system will be off or the generation capacity will be deliberately
reduced. The performance data during this period does not reflect the characteristic of

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 229

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

the system. Thus, this substantiates justification to regard the said period as irrelevant to
the analysis. Furthermore, the performance data during start up and shutdown were also
excluded from the analysis; because the system performance is low at these periods but
most importantly, it is not due to the equipment problem.
The system operation and maintenance cost data were also gathered from the plant
and literature to evaluate the redundancy options.

Parameter Estimation
Without reliability and availability assessment, it is difficult to predict the number of failures
and downtimes which is used as an input for consequences assessment. In binary system
performance evaluation, the equipment is characterized into two states such as working
and failed state. These two states of the system can be determined by analysis of the mean
time between failure (MTBF) and the mean time to repair (MTTR). MTBF estimates how
frequently the system will fail. MTBF is also a basic parameters for reliability (Wang &
Sivazlian, 1997). This can be represented by;
Accumulated operating time
MTBF =
Frequency of failure (1)
MTTR gauges how quickly the system is back to service. This can be represented by;

Accumulated down time

MTTR =
Frequency of failure (2)

In this study, exponential distribution can be used to evaluate the system or equipment
reliability and availability for useful period of the bathtub curve (Rausand & Høyland,
2003). The exponential distribution is a good estimation for repairable system as most
of the repairable component or system lies in the useful period of the bathtub curve. The
useful period of the bathtub curve uses a constant failure rate which means that it can be
approximated by the average actual changing rate during the respected time duration.
Equations (3) and (4) are used to define the system or equipment reliability and
availability respectively.

R(t ) = e − λ ⋅t (3)
µ µ
A(t ) =
λ+µ
+
λ+µ
.e − ( )t
λ +µ

(4)

where λ is the failure rate and µ is repair rate of the equipment. λ and µ can be
defined as Equation (5) and (6) respectively.

230 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

1
λ=
MTBF (5)
1
µ=
MTTR (6)

The system reliability of series and parallel system configuration which contain n
equipment can be represented by Equation (7) and (8) respectively (Rausand & Høyland,
2003).

(7)

(8)

Similarly, the availability of the series and parallel system can be defined using Equation
(9) and (10).

(9)

(10)

Using Equations (7)-(10) depending on the configuration of the system, the cumulative
number of failure and down time can be found using Equation (11) and (12) respectively.

N (t ) = λ ⋅ t (11)

D(t ) = [1 − A(t )] ⋅ t (12)

where N (t ) is the cumulative expected number of failure, D(t ) is the cumulative

expected down time and λ is the constant system failure rate.

Estimation of Cost of Redundancy

In this section, the associated cost of public utility and Genset were formulated
mathematically. Each redundancy options depend on various factors.

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 231

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Cost of using Public Utility as Redundancy

When the cogeneration system used public utility as redundancy, the operator need to
consider maximum demand charge due to hookup electricity from the grid, cost of repair
and opportunity loss. Thus, the total expected cost can be estimated by Equation (13)

Total Expected   Expected   Expected cost of electricity supplied   Expected cost of 

cost of failure  = cost of repair  +  by redundancy  + opportunity loss 
       
 Expected 
− 
 fuel save 
(13)

Cost of Repair
When the system failed, corrective maintenance is applied to bring back the system into
functional state. The cost of corrective maintenance ( C cm ) can be defined as Equation (14).
(14)

where N is the frequency of failure per year and Cr is cost of repair per failure.

Maximum Demand Charge

Public utility supplies to a Co-generator in the incident that the Co-generator does not
produce electricity due to plant failure. The Co-generator has an option of firm or non-
firm supply. Non-firm standby means that public utility does not guarantee that supply can
be given when the Co-generator fails. Due to its connection to public utility, maximum
demand charge cost is imposed when the system fails. This cost is highly dependent on the
frequency of the failure. The maximum demand charge cost per year (Cp) can be estimated
using Equation (15)

(15)

where CMax is the Maximum demand charge cost per kw, K is the capacity in kw required
per connection, and Z is the percentage ratio of the system hook-up of electricity from
redundancy. This means only certain failures which get higher restore time will be hooked
up with electricity from redundancy. The minimum waiting time to hook up electricity from
redundancy system is based on the contract agreement between the cogeneration plant and
the user. Thus, Z can be defined by applying Equation (16) based on the historical data of
a cogeneration system.

Number of hook up redundancy due to failure

Z=
Total failure frequency (16)

232 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

Cost of Supplied Power by Public Utility

During system outage, the plant needs to purchase power from the public utility to avoid
the customer damaging cost. This cost can be calculated using Equation (17)

Cs = Dt C R E R
(17)

where Dt is the total amount of time the plant would be out of service per year as a
result of failure, CR is cost of electricity rate per kw from public utility and ER is the amount
of energy supplied by the redundancy per hour.

Loss of Opportunity
Whenever the failure occurs, the system is down for repair action. This unavailability of
the system will cause opportunity loss. This loss can be represented by

(18)

where C is cost per kw charged to clients and L is the possible amount of power
delivered to clients during the service outage.

Fuel Save
When the system uses the public utility as redundancy in the event of failure, it is
not required to supply the fuel for the cogeneration system as the system is down for
maintenance action. This fuel save FS can be estimated using Equation (15)
FS = C f Dt Y
(19)

where Cf is the cost of fuel per GJ and Y is the amount fuel required to operate
cogeneration per hour.
Therefore, the annual expected cost of failure (AECF) can be obtained using equation
(20)

(20)

Cost of using Gen Set

If the plant uses a Gen set as redundancy, three main cost need to be considered, namely
capital which is related to installation cost, cost of repaired which is related to maintenance
and fuel cost which is related to operation cost. All these costs can be represented by
Equation (21)

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 233

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

(21)

where Q is the capacity of redundancy, Ci is the cost of installation per KW and Cf is

the cost of fuel to operate the Gen set.
The annual expected cost of failure can also be represented by the net present value
(PV) using Equation (22) (Sullivan et al., 2000). The present value means the monetary
amount that should be deposited at a certain rate to pay outlay after n years. This means
that all the annual costs are recalculated to the equivalent value of the present time.

(22)

where ( P ,i,m) is the present worth factor, m is number of years and i is the interest
A
rate.

RESULTS AND DISCUSSION

Case Study
Universiti Teknologi PETRONAS (UTP) power generation cogeneration was taken as
a case study. The availability and reliability of the power generation was linked to the
operation of the two parallel gas turbines installed in the plant. The failure of any one of
the two gas turbines would cause reduction of power being generated. If both turbines
failed, no power will be generated to be supplied to the client. To avoid this occurrence,
the system needs to have back up power supply. For the case of this plant, the backup is
obtained from the national grid. This similar case is being practiced by other cogeneration
plants (Haghifam & Manbachi, 2011; Ramesh & Saravannan, 2011a, 2011b; Shaaban et
al., 2011). One of the main disadvantages of using the national grid as back up is the high
cost of maximum demand charge. This, in turn, leads to high cost charged for any hook up
to the national grid when turbine(s) failed. The configuration of power generation system
is indicated in Figure 3. In this research, a cogeneration plant which consists of two gas
turbines is taken as a case study. The turbines are connected using parallel configuration
to produce electricity for the university area as shown in Figure 3. This simulation block
diagram is developed using BlockSim software. When both turbines fail, the system used
public utility as redundancy. To determine MTBF and MTTR, five years of historical
performance data of Gas turbine were used. In order to capture the failure event and
MTBF from gas turbine performance, the minimum acceptable performance of the gas
turbine was determine based five years daily historical and technical data. Thus, 1497KW
is considered as minimum acceptable performance for both turbines. Any performance of

234 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

gas turbine below the minimum level is considered as the gas turbine in failed state. Based
on this assumption, the MTTF, MTBF, downtime, operating hours, and failure event are
estimated as shown Table 1.

Figure 3. Simulation block for power generation

Table 1
Reliability and availability parameters

Parameters Unit Value

Accumulated Operating Hours Hr 14270
Accumulated Downtime Hours Hr 594
Failure Frequency Number of failure 54
MTBF Hr 264.3
MTTR Hr 11

Estimation of Availability and Reliability

The availability and reliability analysis were performed using BlockSim software. Figure
4 illustrates the availability of power generation system. The plot reflects that the use of
redundancy may enhance the performance of the system. The mean availability of the
system with redundancy was about 98% while the mean availability of power generation
without redundancy is about 85%. The increment of performance obviously will enhance
the profitability of the system and create conducive working environment for the utility
plant, even though the cost of redundancy is expensive.
Figure 5 shows the reliability of the system with and without the effect of redundancy.
The reliability of the system working without redundancy is less than the system working
with reliability. If the system was working without redundancy, there is a high probability
that the system may experience a failure compared to the system working with redundancy.

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 235

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Figure 4. Binary system availability of power generation with redundancy (WR) and without redundancy
(WRO)

Figure 5. Binary system reliability of power generation system with redundancy (WR) and without
redundancy (WRO)

236 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

Estimation of Cumulative Number of Failure and Downtimes

Figure 6 and 7 show that cumulative number of failure and downtime of power generation
through time respectively. As indicated in the plots, the model predicted value was compared
with actual failure frequency and downtime. The validation results show that the model
prediction was closer to actual data. This validation results were further confirmed using
t-test by considering five years observation data. Table 2 shows that the summary of
statistical results using cumulative downtime hours. The statistical results indicate that
there is no statistical difference between the predicted and actual downtime and number
of failure. The P value results indicate 0.095 and 0.062 for cumulative failure and down
time. This mean that statistically no significant different between the model and actual
data as the significance value (p) is greater than 0.05 with 95% confidence level.

Figure 7. Cumulative downtime hours

Figure 6. Cumulative number of failures

Table 2
Statstical validation with cumulative number of failures and downtime hours

Statistical Parameters Cumulative Number of failure Cumulative downtime hours

P value 0.095 0.062
t critical value 2.2622 2.30600

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 237

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Validation and Sensitivity Analysis of Redundancy Cost

The system operation and maintenance cost data were also gathered from the plant and
literature to evaluate the cost of redundancy. The sample costs data considered in this study
are shown in Table 3.

Table 3
Input parameters for redundancy cost evaluation

Parameters Unit Values

Cost of Maximum demand RM / KW 48.6
Cost of Electricity RM/ KW 0.22
Cost of repair RM/failure 100000
Cost of fuel RM/GJ 6.066
Fuel flow RM/GJ/Hr 49.74
Investment Cost for Gen set RM/ KW 999
Production cost of Gen set RM/set 0.17
Current demand of the campus KW/day 5000
Maximum demand KW 8400

The expected cost of failure were estimating the cost of failure caused by the actual
downtime and failure frequency, and then compared with the cost calculated with the
predicted failure and down time using Equation (20) and (21). The results are shown in
Table 4 and 5. The estimated present value of failure cost for the actual failure and down
time is -RM 7,722,356 while the present cost of failure for predicted failure and down time
is RM 7,338,172. The annual value for five years using the actual down time and number of
failure is RM 2,545,317 while the annual cost of failure for predicted failure and down time
is RM 2,582,199. Based on the present and annual cost of failure, the deviation between
the actual and predicted value is 1.43% which falls within acceptable margin. Thus, the
developed failure cost model is useful to predict the impact of failure in monetary value.

Table 4
Cumulative failure cost based on actual number of failure and downtimes

Year Expected Cost Expected Expected Cost Expected Fuel cost Total Expected
of production Penalty cost of supplied Cost of repair saving cost of failure
loss ( RM) power
0 0 0 0 0 0 0
1 -53,391 -1,270,353 -69,095 -777,945 18,952 -2,151,833
2 -97,884 -2,328,981 -126,674 -1,426,233 34,746 -3,945,026
3 -134,962 -3,211,171 -174,656 -1,966,473 47,907 -5,439,354
4 -165,859 -3,946,329 -214,641 -2,416,672 58,875 -6,684,628
5 -191,607 -4,558,961 -247,963 -2,791,839 68,015 -7,722,356
Note: The unit of all costs used in this study is Malaysian Ringgit (RM)

238 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

Table 5
Cumulative failure cost based on predicted number of failure and downtime

Year Expected Cost Expected Expected Cost Expected Fuel cost Total
of production Penalty cost of supplied Cost of saving Expected
loss power repair cost of failure
0 0 0 0 0 0 0
1 -73,333 -1,071,630 -56,667 -583,333 -20,115 -1,764,848
2 -128,333 -2,474,955 -99,167 -1,347,222 -35,201 -4,014,476
3 -184,988 -3,325,455 -142,946 -1,810,185 -49,215 -5,414,359
4 -228,095 -4,122,799 -175,845 -2,244,213 -61,039 -6,709,913
5 -265,229 -4,647,799 -204,539 -2,565,715 -71,224 -7,612,057
Note: The unit of all costs used in this study is Malaysian Ringgit (RM)

As can be seen in Figure 8, 58.5% of the failure cost was due to penalty cost (maximum
demand charge) of failure, 35.8% was the cost incurred to restore the system, 3.2 %
was contributed by the electricity used during downtime of the cogeneration systems
and 2.5% was the estimated loss of power due to cogeneration system failure. It can be
observed that Maximum demand charge contributes to the high cost of using public utility
as a redundancy system for cogeneration plant. Public utility is a power supplied by the
national electricity. Basically the use of redundancy is associated with number of failure.
This means that it relates further with reliability and availability. Hence, predicting the
number of failure and down time will support the development of maintenance strategy,
thus reducing the frequency of occurrence for redundancy to be utilized. This also helps
in reducing the cost of maximum demand charge to be borne by the client. Essentially, the
reliability and availability of the power generation system is enhanced. It also minimizes
the cost of failure associated with redundancy.

Figure 8. Contribution of failure cost

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 239

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Spider Plot Analysis

The redundancy selection is affected by several factors such as Maximum demand charge,
installation cost and Discount interest. These factors are not constant from place to place
and through time as well. In order to analyse the effect of each parameter, the spider plot
analysis was used. Figure 9 and 10 showed the effect of various parameters for both
redundancy possibilities such as public utility and Gen set respectively. The intersection
of each curve with the abscissa shows the decision reversal point - the percentage change
from each factor’s most likely value at which the PW is zero. As shown in Figure 9, the slop
of Maximum demand charge and MARR steeper compare to other factors which means
that the PV for public utility is more sensitive to Maximum demand charge and MARR.
Similarly, Figure 10 shows the installation cost is more sensitive to Gen set compared to
other factors. Thus, the cogeneration operator need to look closely on Maximum demand
charge, MARR and installation cost to choose the best redundancy options.

Figure 9. Sensitivity analysis for public utility

Figure 10. Sensitivity analysis for Gen set

240 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

Redundancy Evaluation of Power Generation System

Redundancy is essential for power generation to improve the performance of the system
but it is very expensive to operate as it incurs maximum demand charge penalty. To avoid
the redundancy totally from the system is difficult because the customer damaging cost
of the utility system is substantial. However, one can minimize the effect of redundancy
by selecting the suitable type. Currently, the campus electricity generation system uses
public utility as redundancy, but it can alternately install Gen set as redundancy options.
In order to compare these two redundancy options, Equation (22) were used to estimate
the associated present of each redundancy option for a 20 years’ life span. The results of
present value of each redundancy are presented in Figure 11. This result shows that in the
first 5 years, the present value of public utility is less than Gen set which means that it is a
better option than Genset if the useful period of cogeneration is less than 5 years. However,
when the useful period of cogeneration is greater than five years, Gen set would serve well
as the present value is less than public utility. The present value for public utility and Genset
redundancy at the end of 20 years were RM11, 948,611 and RM8,721,946 respectively.
Thus, by the end of year 20, using Gen set would minimise 24% of the redundancy cost
compared to Public utility.

Figure 11. Comparison of Gen set and public utility based on failure cost evaluation

Effect of Installation Cost, Maximum Demand Charge and Discount Interest On

Redundancy Selection
There are certain factors which can affect the failure cost of Gen set and public utility such
as installation cost, capacity, maximum demand charge, and interest rate. These parameters
may vary from time to time or place to place. Thus, there is a need for sensitivity analysis
in order to identify the breakeven point for decision making. Regarding installation cost,
the information taken from gas turbine hand book (Farmer & De Biasi, 2010) infers that
Gen set installation range from approximately $300 per kW for very large utility-scale

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 241

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

plants to $1,000 per kW for small industrial cogeneration installation. However, the prices
of construction can vary as a result of local labour market conditions and the geographic
conditions of the site. Figure 12 shows the effect of cost of installation on annual cost for
different redundancy options. The variation in installation cost affect the Gen set redundancy
than public utility because the public utility redundancy is already installed and functions
with the existing system. The plot result indicates that if the cost of installation was less than
RM1714.73 per kW, Gen set would be preferable to public utility. However, if the cost of
installation for Gen set is higher than RM1714.73, public utility would be a better option.

Figure 12. Effect of installation cost on redundancy evaluation

The second factor is the maximum demand charge. The maximum demand costs may
vary depending on the plant’s location. In areas where electricity costs are high, for a base-
load cogeneration system, its costs can account for up to 70% of the total plant consequence
costs. The sensitivity was done by varying the maximum demand charge from RM0 to
RM60 per kW. Figure 13 shows that the breakeven value for maximum demand charge
cost is RM28.92 per kw. If the maximum demand charge is less that 28.92 kW per hour,
using public utility as redundancy could be a better option. On the contrary, if the penalty
cost is higher than RM28.92 per kW, Gen set could be a better option.
The installed capacity of the Gen set varies based on the demand of customer and
interest of the owner to make a decision on the redundancy. So, one needs to see the effect
of installed capacity by comparing with the existing installed public utility. Figure 14 shows
the comparison of public utility redundancy with Gen set when the capacity is increasing.
The breakeven capacity is 7.43 MW. If the plant installed the Gen set capacity at less than
7.43 MW, the public utility option should be rejected. If the plant installed more than 7.43
MW, it is better to use public utility as redundancy than Gen set.

242 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

The minimum attractive rate of return (MARR) is one of the factors that may affect the
consequence of failure, which also varies through time. The effect of MARR on redundancy
selection is shown in Figure 15. As it can be seen from the graph, if the MAAR is less than
35%, Gen set can be chosen as redundancy. If discount rate is greater than 35%, public
utility is the better redundancy option. Spacing is different for this section.

Figure 13. Effect of maximum demand charge on redundancy evaluation

Figure 14. Effect of capacity on redundancy evaluation

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 243

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Figure 15. Effect of interest rate on redundancy evaluation

CONCLUSION
This study is very useful as it analyses various redundancy options for cogeneration system
from an economic perceptive. Furthermore, it developed a reliability based cost model
which included reliability and availability concept to analyse the economic benefits on the
selection of redundancy for cogeneration plants. This paper also examines various factors
which can affect the selection of redundancy. The case study findings indicated that gen
set redundancy is better if the cogeneration plant operates for a period of more than five
years. However, in short term cogeneration operations which accounts for less than five
years, the public utility would be a better options. The sensitivity analysis also indicated
Maximum demand charge, MARR and installation cost have significant effect on the
selection of redundancy. In general, this study is very useful for cogeneration operators to
select the best redundancy option which incurs minimum cost.

REFERENCES
Arora, N., & Kumar, D. (1997). Availability analysis of steam and power generation systems in the thermal
power plant. Microelectronics Reliability, 37(5), 795-799.

Chen, H., Xu, J., Xiao, Y., Qi, Z., Xu, G., & Yang, Y. (2018). An Improved Heating System with Waste Pressure
Utilization in a Combined Heat and Power Unit. Energies, 11(6), 1-20.

Christiansen, T. (2013). Risk-Based Assessment of Unplanned Outage Events and Costs for Combined-Cycle
Plants. Journal of Engineering for Gas Turbines and Power, 135, 021801-021801.

Dougan, K. W., & Reilly, M. C. (1993). Quantitative reliability methods improve plant uptime. Hydrocarbon
Processing;(United States), 72(8), 131-141.

244 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)

 Reliability based Redundancy Assessment of a Cogeneration Plant

Eti, M. C., Ogaji, S. O. T., & Probert, S. D. (2007). Integrating reliability, availability, maintainability and
supportability with risk analysis for improved operation of the Afam thermal power-station. Applied
Energy, 84(2), 202-221.

Farmer, R., & De Biasi, B. (2010). Gas turbine world handbook. Gas Turbine World, Fairfield, CT.

Gräber, U. (2004). Advanced maintenance strategies for power plant operators—introducing inter-plant life
cycle management. International Journal of Pressure Vessels and Piping, 81(10), 861-865.

Grace, D., & Christiansen, T. (2012, August 30). Quantifying the cost of unplanned outage events for combined-
cycle plants. Energy-Tech Magazine. Retrieved October 10, 2017, from https://www.energy-tech.com/
ram/article_b362347c-4a9d-542a-b9b4-fdd137136342.html

Haghifam, M. R., & Manbachi, M. (2011). Reliability and availability modelling of combined heat and power
(CHP) systems. International Journal of Electrical Power & Energy Systems, 33(3), 385-393.

Lemma, T. A., & Hashim, F. M. (2013). IFDD: Intelligent fault detection and diagnosis-application to a
cogeneration and cooling plant. Asian Journal of Scientific Research, 6(3), 478-487.

Lewis, E. E., & Lewis, E. E. (1987). Introduction to reliability engineering. New York: Wiley.

Louit, D. M., Pascual, R., & Jardine, A. K. (2009). A practical procedure for the selection of time-to-failure
models based on the assessment of trends in maintenance data. Reliability Engineering and System Safety,
94(10), 1618-1628.

Boyce, M. P. (2011). Gas turbine engineering handbook. Waltham: Elsevier.

Pham, H., & Wang, H. (1996). Imperfect maintenance. European Journal of Operational Research, 94(3),
425-438.

Qiu, G., Xia, C., & Zhang, H. (2011, March). Estimation of Failure Cost in Life Cycle Cost of Power Equipment.
In 2011 Asia-Pacific Power and Energy Engineering Conference (APPEEC) (pp. 1-4). Wuhan, China.

Ramesh, V., & Saravannan, R. (2011a). Reliability assessment of a co-generation power plant in a sugar mill
using fault tree analysis. Energy Sources, Part A: Recovery, Utilization, and Environmental Effects,
33(12), 1168-1183.

Ramesh, V., & Saravannan, R. (2011b). Reliability Assessment of Cogeneration Power Plant in Textile Mill
Using Fault Tree Analysis. Journal of Failure Analysis and Prevention, 11(1), 56-70.

Rausand, M., & Høyland, A. (2003). System reliability theory: models, statistical methods, and applications
(Vol. 396). Hoboken: John Wiley & Sons.

Reshid, M. N., Muhamad, W. M. W., & Maarof, R. A. R. (2017a). Empirical Analysis of Chilled Water
Generation for Off Peak Period of Cogeneration plant Using Neural Network. International Journal of
Applied Engineering Research, 12(24), 14669-14676.

Reshid, M. N., Muhamad, W. M. W., & Majid, M. A. A. (2017b). Transient Simulation of a Waste Heat Recovery
from Gas Turbine Exhaust. Journal of Applied Sciences, 17, 22-31.

Richwine, R. R. (2004). Casom 18: The relationship between scheduled maintenance and forced outages and
its economic impact on selecting availability goals. World Energy Council, Section 6, 1-3.

Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019) 245

 Meseret Nasir­­, Wan Mansor Wan Muhamad and Raja Aziz Raja Maarof

Shaaban, M., Azit, A. H., & Nor, K. M. (2011). Grid integration policies of gas-fired cogeneration in Peninsular
Malaysia: fallacies and counterexamples. Energy Policy, 39(9), 5063-5075.

Soares, J. B., Szklo, A. S., & Tolmasquim, M. T. (2001). Incentive policies for natural gas-fired cogeneration
in Brazil’s industrial sector—case studies: chemical plant and pulp mill. Energy Policy, 29(3), 205-215.

Sullivan, W. G., Wicks, E. M., & Luxhoj, J. T. (2000). Engineering economy (Vol. 12). Upper Saddle River,
NJ: Prentice Hall.

Vega, F., Hill, D. K., & Collins, C. (1998, May). Plant Reliability Analysis in LNG Plants. In Twelfth
International Conference & Exhibition on Liquefied Natural Gas (pp. 1-22). Perth, Austrailia.

Wang, K. H., & Sivazlian, B. D. (1997). Life cycle cost analysis for availability system with parallel components.
Computers & Industrial Engineering, 33(1), 129-132.

246 Pertanika J. Sci. & Technol. 27 (1): 225 - 246 (2019)