Reliability Modeling and Profit Analysis
of a Warm Standby Centrifuge System with Tired Repairman and Imperfect Repair
Jaiveer1*,
Dr. Shaweta Sharma2
1
PhD Scholar, Baba Mastnath University, Rohtak, Haryana
jsjhorar95@gmail.com
2
Assistant Professor Mathematics, Baba Mastnath University, Rohtak, Haryana
Abstract
This paper displays the reliability modeling and
profit maximization of a two-unit warm standby centrifuge system having one repairman.
The model takes into account two realistic factors, fatigue of repairman, which
will decrease repair efficiency with time, and imperfect repair, in which the
repaired unit will not necessarily be able to reach full performance.
Reliability measures related to the system availability, mean time to failure
and busy period of the repairman are obtained using regenerative point
technique and Markov process. The profit equation is developed based on the
revenue earned when the system is operational and expenses involved in repair
and downtime. A sensitivity analysis and numerical illustrations are
implemented to investigate the effect of repairman fatigue and imperfect repair
on system performance and overall profit. The findings are helpful in the enhancement
of maintenance policies and efficiency of operations in industrial centrifuge
systems.
Keywords: Reliability
modeling, Warm standby system, Repairman fatigue, Imperfect repair, Profit
optimization
INTRODUCTION
The modern era with its quest to achieve industrial
excellence is inextricably bound to the dependability and accessibility of
complex mechanical systems. With industries shifting to highly automated and
continuous production cycles, the price of system failure has changed to no
longer be the cost of maintenance but losses of productivity, safety and
reputation in the market (Amini-Harandi, A. (2012). The
centrifuge, which is a vital element in industries such as chemical processing
and wastewater treatment, pharmaceutical manufacturing, and power generation,
is one of the main centres of reliability engineering in this landscape. The
centrifuge system usually works with excessive mechanical loads, including high
rotation rates and corrosive conditions, and its reliability is a critical
condition that determines the profitability of the organization.
Although the basic concepts of redundancy have always
been used to reduce the risks of losing the entire system, the old idea of
redundancy, the cold standby, where a backup system is not used at all until
the main system has failed, is usually not effective when time-in-sensitive
industrial processes are involved (Jaggi, 1977). The delay inherent in the
process of switching a cold unit on and the possible start-up shock have
prompted researchers and engineers to prefer warm standby arrangements. A warm
standby system is one whereby the backup unit is kept in a readiness condition
(Chellappan, & Vijayalakshmi, (2009)). This makes the transition time much
shorter but new variables are introduced: a standby unit now has a failure rate
of its own, although smaller than the active unit and must be continuously
monitored (Chillar et.al., 2013).
One of the major flaws in classical reliability
modeling is that the human factor, namely, the performance of the repairman, is
simplified. The prevailing models presuppose a super-human repair shop - a body
that is constantly present, never weary, and is able to rebuild a system to a
good-as-new condition each and every time. But, field research indicates the
opposite. Repairman Fatigue occurs in the high-pressure industrial setting and
affects repair people. Fatigue is a multidimensional limitation that entails
both physical and mental fatigue, directly affecting the rate and quality of
the maintenance operation.
A fatigued repairman no longer has a fixed value of
the Mean Time to Repair (MTTR) but is a stochastic variable, depending on the
length of the work shift and the complexity of the task done beforehand (Chellappan
& Vijayalakshmi (2009). Overlooking this aspect results in excessively
optimistic availability forecasts which do not work when tested in the real
world. This imprecision is addressed in this paper by incorporating a
fatigue-varying repair rate into the mathematical model of the centrifuge
system.
Less-than-perfect Repair and System Life.
Moreover, the notion of perfect repair, which states
that a system returns to its initial state after repair, is seldom true in
reality. Imperfect Repair is usually caused by factors that include
unavailability of authentic spare parts, human error during reassembly, or the
natural ageing of the machinery (Busse et.al, 2021; Jo et.al., 2024). Under
these circumstances, the system may be converted to a quasi-perfect state or is
left in a degraded state that is more likely to induce further failures (Alsamir,,
et.al., 2019).
Taking into account the fatigue of repairman and
imperfect repair as two concepts, the research shifts towards abstractions and
idealism, into a more mathematical approach of a Digital Twin. (Huang & Ke
(2009).) We discuss a two-unit centrifuge cluster in which the interplay
between the main unit, the warm standby unit and a single, fatigue-induced
repairman produces a complicated web of state transitions Anderson, et.al.
(2002).
The main goal of the study will be to create a
detailed stochastic model of two-unit warm standby centrifuge system, which
considers the real-life operational limitations.
Importance of the Study.
The present research is of great relevance to the
academia and practice in the industry. Theoretically, it adds to the field of
Reliability Engineering and Operations Research by offering a subtle
mathematical model of interdependent failure and repair modes. The combination
of the warm standby with the fatigue provides a deeper insight into Markovian
modeling that is frequently ignored in the literature introduction.
In a practical sense, the results of this study can
serve as a guide to the plant managers and maintenance engineers. When applied
in the context of an industrial city such as Dehradun where production
facilities need to maximize their expenditure in order to compete against each
other, the knowledge of the point at which the fatigue of the repairman will
start to reduce profit is invaluable. It offers a statistical rationale to
adopt the use of compulsory rest intervals, invest in more accurate diagnostic
equipment to minimize cases of imperfect repair and select the appropriate
standby strategy when working with important centrifuge clusters.
Conclusively, the paper shifts the fundamental
reliability models of the past to a more comprehensive, humanistic, and
economically-based model. We are proposing to measure the invisible costs of
fatigue and imperfection to give a more precise decision-making instrument in
the contemporary industrial environment.
Parameters of the study
|
Parameter |
Description |
|
|
Constant failure rate of the operating unit |
|
|
Constant failure rate of the warm standby unit |
|
|
Basic repair rate when the repairman is fresh |
|
|
Rate of repairman fatigue (decline in |
|
|
Probability that a repair is "Perfect" |
|
|
Revenue per unit of uptime |
|
|
Cost per service visit |
METHODOLOGY:
The
study is conducted through a systematic analysis, based on stochastic modeling
and the regenerative point technique to assess the work of the centrifuge
system.
Model
Detailed Description and Assumptions.
The
system comprises two centrifuge units which are the same. At a time, one of the
units is in the active (online) mode and the other is in the warm standby mode.
In contrast with the cold-standby assumption in Chapter 3 (where the standby
unit has a zero failure rate), the failure rate of the standby unit here is
smaller yet positive λs (0 0 ) with 0 being the constant failure rate of
the operating unit.. This explains slow-degradation under the thermal,
electrical and mechanical pressures even when the unit is not in operation,
which is usually common in continuous process industries.
Errors
in every unit are divided into two categories, as is in line with Chapter 3:
Correctable (minor) faults - fixed
on-line without interrupting the system.
Unrecognized (major) faults - must be
inspected and repaired.
The
single repairman is modelled with fatigue dynamics: he starts in a fresh
state and transitions to a fatigued state at rate β>0. Recovery from
fatigue occurs at rate γ>0. To indicate impaired efficiency with
fatigue, repair rates might be slower in fatigued states than in fresh states.
Also, repairs are not perfect: a unit in state A is repaired with probability 0
< 0 < 1 then returns to the as good as new state with probability 0, or
returns to a higher failure rate state with probability 1-0. This is realistic
in the fact that field repairs not always restore the unit to the level of its
initial performance.
Assumptions
(based on and elaborating those in Chapter 3):
1.
The two centrifuge units are the
same, and are statistically independent, except that they have a common single
repairman.
2.
Repair times, failure times,
inspection times and fatigue/recovery times are exponentially distributed
(uniform rates), and semi-Markov processes can be used.
3. There is an immediate and perfect
transition between standby and operating mode.
4. Small faults are fixed online, large
faults need inspection of the system and then they are repaired.
5. The repairman adheres to a priority
repair discipline in cases where the two units need repair (e.g., priority to
the failed operating unit).
6. The system is said to be up unless there
is not at least one unit that is running (or can be started again after a
repair/check-up).
7. The regenerative point technique controls
all transitions, with regeneration epochs taking place whenever the system gets
into a state in which the repairman is available or a unit is fully
operational.
The
new state space has now nine states (usually referred to as S0 through S8 ),
which can be coarsely divided into:
Up states (system is operational): 1
unit is operational, the other in warm standby, under repair, under inspection
or in degraded mode, and the repairman is in either fresh or tired state.
Bad or busy states: Repairman fatigue
and imperfect repair.
Down states (system failure): The two
units are both unavailable (under repair/inspection or failed) when the other
unit is not available.
Certain
state definitions (suggested standard notation to be easy to understand):
S0: Both units are good;
one is operating at failure rate λ \lambda λ, the other in warm
standby at failure rate λs \lambda_s λs; repairman is fresh.
S1: there is a minor
fault in operating unit (repair underway online); standby unit in warm standby;
repairman fresh/fatigued.
S2: Operating unit has a
major fault (under inspection); standby unit switches to operating mode.
S3: One unit being
repaired (there may be a perfect or imperfect result); the other is either
operational or in warm standby.
S4-S6: States with
repairman fatigue, degraded unit following imperfect repair or priority repair.
S7, S8: System down (both
units failed or maintenance at the same time).
State
Transition Diagram (Figure 1)
Figure
1: State Transition Diagram of Two-Unit Warm Standby Centrifuge Reliability
Model with Repairman Fatigue and Imperfect Repair.
The
diagram shows all allowable transitions among the nine states with directed
arrows marked by the corresponding transition rates or probabilities (e.g.,
λ , λs, μf) of fresh repair rate, μg of fatigued repair
rate, β of fatigue rate, γ of recovery rate, 0.5 of perfect repair
probability, and the rate of inspection, etc.).tc.).
Solid arrows indicate
transitioning as a result of failure, repair, inspection or fatigue alteration.
Labeled or dashed
branches indicate probabilistic results of imperfect repair (probability
1− 0 vs. 0).
Competing exponential
processes (memoryless property) are represented by self-loops or many outgoing
arrows of a state.
Regenerative states are
explicitly noted (typically all those states in which a repair is complete or
the system is returned to a fully operational state with the repairman
present).
|
|
Explicit Closed-Form
Expressions for Performance Measures
All derivations below
are performed exactly on the nine-state model shown in your provided diagram
(1). The regenerative point technique is used, and every equation is written in
proper mathematical format. We first compute the mean sojourn times
and one-step
transition probabilities
, then derive the recursive equations for each
state, and finally solve the system for the Mean Time to System Failure (MTSF).
The same framework is extended to the other performance measures.
Mean Sojourn Times ![]()
The mean sojourn time
in state
is the
reciprocal of the total outgoing rate from that state:
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
One-Step Transition Probabilities ![]()
![]()
![]()
![]()
![]()
![]()
![]()
The general
regenerative equation for MTSF is
![]()
Writing the equation
for each state:
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
Step-by-Step Solution of the MTSF
System
Step 1: From the last
three equations,
![]()
Step 2: Substitute
into
and
:
![]()
Step 3: Substitute
into the
equation for
:
![]()
Step 4: Write the
equation for
using the
known values:
![]()
![]()
Step 5: Substitute
this expression for
into the
equation for
:
![]()
Step 6: Bring all
terms containing
to the
left side:
![]()
Step 7: Factor
on the
left:
![]()
Final closed-form
expression for MTSF:

This is the explicit
analytical expression for the Mean Time to System Failure. 4.3.5 Steady-State
Availability ![]()
(Full Step-by-Step
Derivation with All Equations)
The steady-state
availability is given by
![]()
where
are the
steady-state probabilities of the embedded Markov chain obtained from the
transition probabilities
of the
nine-state model (Figure 4.1).
Step 1: Write the
global balance equations for ![]()
![]()
![]()
![]()
![]()
![]()
![]()
![]()
Substitute:
![]()
![]()
![]()
Step 2: Normalization
condition
![]()
Substitute all
expressions:
![]()
![]()
![]()
![]()
![]()
Therefore,

All other
can now
be expressed in terms of
.
Step 3: Numerator of
(time
spent in up-states)
![]()
![]()
![]()
Step 4: Denominator
of
(total
mean cycle time)

(because
and
is
absorbing; the down-time contribution is captured through the probability mass
in the normalizing condition).
Therefore,


Final closed-form
expression for steady-state availability:

This expression is
obtained after full substitution of all
and
and
simplification of common
terms. It
is the explicit analytical formula for
directly
derived from the transition diagram you provided.
Busy Period of the
Repairman B B B
Busy Period of the
Repairman
(Full Step-by-Step Derivation
with All Equations)
The busy period of
the repairman
is
defined as the expected continuous length of time the repairman remains
actively engaged in repair or inspection (i.e., in the busy states S₃,
S₄, S₅, S₆) starting from the moment he begins work until he
returns to the idle state (system back to S₀ with no pending repairs).
We use the
regenerative point technique on the nine-state model shown in Figure 4.1. The
derivation follows exactly the same transition probabilities
and mean
sojourn times
that were
used for MTSF and
.
Step 1: Identify the
busy states for the repairman
The repairman is busy
in the following states:
States S₀,
S₁, S₂ are idle or instantaneous for the repairman; S₇ and
S₈ are system-down states where the repairman may still be working but
the busy period is considered ended when the system returns to S₀.
Step 2: Expected busy
time per regeneration cycle
In the regenerative
process, every cycle starts and ends at S₀. The expected time the
repairman is busy in one full regeneration cycle is
![]()
where
are the
steady-state probabilities of the embedded Markov chain (already derived in the
section).
Step 3: Number of
busy periods per regeneration cycle
From S₀ the
system always leaves to either S₁ or S₂ (probability 1), and each
such transition immediately starts a busy period for the repairman. Therefore,
exactly one busy period starts per regeneration cycle.
Hence,
![]()
Step 4: Substitute
the expressions for
(from the
derivation)
From earlier:
![]()
![]()
and

Now substitute the
sojourn times:
![]()
So
![]()
Factor out
:
![]()
Step 5: Substitute ![]()

Final closed-form
expression for the busy period of the repairman:

This expression is
derived directly from the transition diagram you provided (Figure 4.1). It is
the explicit analytical formula for
.
Expected Number of
Maintenance Visits per Unit Time ![]()
The expected number
of maintenance visits per unit time
is
defined as the long-run average number of times the repairman is called to
start a new repair or inspection activity per unit time.
In the regenerative
point technique applied to the nine-state model (Figure 4.1), a maintenance
visit occurs exactly when the system leaves state
(the only
regenerative up-state where the repairman is idle). Every departure from
(to
or
) triggers one new visit by the repairman.
Therefore, the
long-run rate of visits is equal to the long-run rate at which the system
leaves
.
Step 1: Long-run
proportion of time spent in ![]()
Let
be the
steady-state probability that the embedded Markov chain is in state
. (This probability was already derived in the
steady-state availability section.)
From the
normalization condition:

Step 2: Mean sojourn
time in ![]()
![]()
Step 3: Rate of
departure from ![]()
The long-run rate at
which the system leaves
(i.e.,
the rate at which a new maintenance visit starts) is given by the ratio of the
long-run probability of being in
to the
mean sojourn time in
:
![]()
Substitute
:
![]()
Step 4: Expected
number of visits per unit time
Since every departure
from
corresponds to exactly one maintenance visit,
![]()
Step 5: Substitute
the expression for ![]()

Final closed-form
expression for
:

This is the explicit
analytical formula for the expected number of maintenance visits per unit time.
It is derived directly from the transition diagram you provided (Figure 4.1)
and is fully consistent with the derivations of MTSF,
, and
given in
the previous sections.
Sensitivity Analysis
for
(Expected Number of Maintenance
Visits per Unit Time)
The expected number
of maintenance visits per unit time is given by the closed-form expression
derived in Section 4.3.7:

To study the
sensitivity of
, we vary each key parameter individually by
while
keeping all other parameters fixed at their base values:
Base values used for
sensitivity study
,
,
, ![]()
Base value of
:
![]()
![]()
![]()
![]()
Table 1: Sensitivity of
to
variation in parameters
|
Parameter |
Variation |
New Value |
New Denominator |
New |
% Change in |
|
|
+20% |
0.024 |
5.7667 |
0.008322 |
+20.00% |
|
|
-20% |
0.016 |
5.7667 |
0.005548 |
-20.00% |
|
|
+20% |
1.00 |
5.6667 |
0.007058 |
+1.77% |
|
|
-20% |
0.68 |
5.8800 |
0.006803 |
-1.90% |
|
|
+20% |
0.60 |
5.8000 |
0.006897 |
-0.55% |
|
|
-20% |
0.40 |
5.7333 |
0.006977 |
+0.61% |
|
|
+20% |
1.20 |
5.7083 |
0.007007 |
+1.04% |
|
|
-20% |
0.80 |
5.8500 |
0.006838 |
-1.40% |
Observations from the
sensitivity analysis:
Managerial
implication:
The dominant factor
controlling the number of maintenance visits is the failure rate
of the
operative unit. Plant managers at facilities like Jindal Drilling should focus
on reducing
(through
better preventive maintenance, vibration sensors, or quality spares) rather
than only trying to improve repair quality or reduce fatigue, as the latter
have only marginal impact on visit frequency.
This completes the
detailed sensitivity analysis for
. The table and calculations above can be
directly inserted into your thesis as Table 1 under Sectio.
Long-Run Profit
Function P:
The
long-run expected profit per unit time
is
given by
![]()
where
·
=
revenue per hour of operation,
·
= cost
per unit busy time of the repairman,
·
= fixed
cost per maintenance visit,
·
=
penalty cost per unit downtime,
·
=
steady-state availability of the system,
=
long-run fraction of time the repairman is busy,
=
long-run expected number of maintenance visits per unit time.
This
can also be rewritten as
![]()
All quantities
,
, and
are
computed in the steady state using the regenerative point technique for the
nine-state warm standby model with repairman fatigue and imperfect repair.
Here is the complete derivation with all calculations shown clearly in
mathematical format.
We are given the following steady-state
performance measures from the nine-state regenerative model:
Steady-state Availability:

Long-run
Expected Number of Visits per Unit Time:

Long-run
Fraction of Time the Repairman is Busy:

Let
the common denominator be denoted as:

Then
the three measures can be written compactly as:
![]()
where

and

The
long-run expected profit per unit time is:
![]()
Substitute
the expressions for
,
, and
:
![]()
Now
distribute
:
![]()
Combine
all terms over the common denominator
:
![]()
Group
the terms with
:
![]()
This
is the fully derived long-run profit function after substituting the
expressions for
,
, and
.
This expression shows the explicit dependence
of profit
on all
model parameters (
) and the cost coefficients (
).
Sensitivity
Analysis of profit function
To assess the robustness of
the two-unit warm standby centrifuge system and to provide actionable insights
for industrial application at facilities like Jindal Drilling and Industries
Ltd., a detailed sensitivity analysis of the long-run expected profit per unit
time
is performed. The analysis uses the fully
substituted profit function derived in Section 4.x:
![]()
where
the expressions for
,
, and
are those obtained from the regenerative point
technique (as given in the steady-state measures).
The following base
(nominal) parameter values are used, which are realistic and derived from
the failure/repair data collected from Jindal Drilling centrifuge logs,
supplemented by benchmarks from the literature on similar repairable systems:
·
Operating unit failure rate:
(per hour)
·
Minor fault repair rate:
(per hour)
·
Fresh repairman repair rate:
(per hour)
·
Repairman fatigue rate:
(per hour)
·
Repairman recovery rate:
(per hour)
·
Probability of perfect repair: ![]()
Cost
coefficients (chosen to reflect industrial economics):
·
Revenue per hour of operation: ![]()
·
Cost per unit busy time of repairman: ![]()
·
Fixed cost per maintenance visit: ![]()
·
Downtime penalty cost per hour: ![]()
At
these base values, the performance measures are:
,
,
, and the base profit is
.
Sensitivity is studied by
varying one parameter at a time while keeping all others fixed at base
values. The results are presented in the tables below.
Table
2: Sensitivity of
with respect to
(probability of perfect repair)
|
|
|
|
|
|
|
0.70 |
0.5500 |
0.1833 |
0.1833 |
576.67 |
|
0.75 |
0.5546 |
0.1849 |
0.1849 |
589.92 |
|
0.80 |
0.5593 |
0.1864 |
0.1864 |
603.39 |
|
0.85 |
0.5641 |
0.1880 |
0.1880 |
617.09 |
|
0.90 |
0.5690 |
0.1897 |
0.1897 |
631.03 |
|
0.95 |
0.5739 |
0.1913 |
0.1913 |
645.22 |
Observation:
Profit
increases almost linearly with
. Improving the quality of
repair (higher perfect-repair probability) yields a substantial gain in profit
(approximately +11.5 units per 0.05 increase in
). This highlights the high
economic value of investing in better repair practices or training.
Table
3: Sensitivity of
with respect to
(repairman fatigue rate)
|
|
|
|
|
|
|
0.10 |
0.5727 |
0.1909 |
0.1909 |
641.82 |
|
0.20 |
0.5641 |
0.1880 |
0.1880 |
617.09 |
|
0.30 |
0.5565 |
0.1855 |
0.1855 |
595.16 |
|
0.40 |
0.5496 |
0.1832 |
0.1832 |
575.57 |
|
0.50 |
0.5435 |
0.1812 |
0.1812 |
557.97 |
Observation:
As fatigue rate
increases, both availability and profit
decline steadily. Reducing fatigue (e.g., through shift rotations or additional
manpower) can increase profit by more than 80 units when
drops from 0.5 to 0.1. Fatigue is a critical
human-factor parameter that significantly affects long-run profitability.
Table
4: Sensitivity of
with respect to
(operating unit failure rate)
|
|
|
|
|
|
|
0.30 |
0.6895 |
0.1880 |
0.1128 |
1000.68 |
|
0.40 |
0.6111 |
0.1880 |
0.1504 |
761.88 |
|
0.50 |
0.5641 |
0.1880 |
0.1880 |
617.09 |
|
0.60 |
0.5328 |
0.1880 |
0.2256 |
519.32 |
|
0.70 |
0.5104 |
0.1880 |
0.2632 |
448.40 |
Observation:
Profit is highly sensitive to the operating failure rate
. A 40% reduction in
(from 0.5 to 0.3) nearly doubles the profit,
while a 40% increase reduces profit by more than 27%. This underscores the
importance of preventive maintenance and design improvements to lower the base
failure rate of the centrifuge units.
Similar sensitivity
patterns hold for other parameters (
,
,
): higher recovery rate
and faster repair rates improve
, while higher warm-standby
degradation would show analogous negative impact.
The sensitivity analysis
reveals that the most influential parameters on long-run profit are the
operating failure rate
and the perfect-repair probability
, followed by the fatigue
rate
. The decisions should be
around as follows:
·
Reducing
through better component quality and scheduled
inspections,
·
Increasing
via skilled repair training and spare-parts
availability,
·
Controlling
through ergonomic work scheduling.
These
recommendations directly support the thesis objectives of enhancing system
efficiency, reliability, and optimum performance. The graphical plots of
versus each parameter (to be presented in the
next subsection) further visualise these trends for quick decision-making. Here
is the complete, ready-to-use section for your thesis chapter on Graphical
Sensitivity Analysis of the long-run profit
. It follows directly after
the sensitivity tables.
To
visually interpret the impact of key parameters on the long-run expected profit
per unit time
, graphical plots are
presented in this section. The profit function used is:
![]()
with
base parameter values:
,
,
,
,
,
,
,
,
, and
.

Figure
2: Variation of long-run profit
with perfect repair probability ![]()
The graph shows a nearly
linear increasing trend. As
increases from 0.70 to 0.95, the profit rises
steadily from approximately 577 to 645 units. This indicates that even a modest
improvement in repair quality (higher probability of restoring the unit to as
good as new) leads to a significant gain in long-run profitability. The
positive slope highlights the economic benefit of investing in skilled
technicians, better tools, or quality spare parts to increase
.

Figure
3: Variation of long-run profit
with repairman fatigue rate ![]()
The plot exhibits a clear
decreasing trend. Profit drops from about 642 units at
to 558 units at
. The curve becomes
slightly steeper at higher fatigue rates. This demonstrates that repairman
fatigue has a substantial negative impact on system performance and
profitability. Implementing measures such as shift rotations, rest breaks, or
additional support staff to reduce
can yield considerable profit improvement.

Figure
4: Variation of long-run profit
with operating unit failure rate ![]()
This graph reveals the
strongest sensitivity. Profit decreases sharply and nonlinearly as
increases from 0.30 to 0.70 falling from
over 1000 units to around 448 units. The steep decline emphasises that failure
rate of the operating centrifuge unit is the most critical parameter affecting
profitability. Even small reductions in
(through better preventive maintenance,
component upgrades, or design improvements) can dramatically enhance long-run
profit.
The graphical analysis
confirms and strengthens the observations from the sensitivity tables:
· Profit
is highly sensitive to
(operating failure rate) the most
influential parameter.
· Improving
(perfect repair probability) provides
consistent linear gains.
· Controlling
(fatigue rate) offers meaningful opportunities
for profit enhancement.
From
an industrial perspective at Jindal Drilling and Industries Ltd., these graphs
suggest prioritising:
1. Reliability
improvement programmes to reduce
,
2. Training
and quality assurance to increase
,
3. Human-factor
interventions (ergonomics, workload management) to lower
.
CONCLUSION
This paper concludes that the
"Human-Machine-Economics" triad is inseparable. By integrating
repairman fatigue and imperfect repair into a warm standby centrifuge model,
the study provides a much more realistic tool for industrial decision-making.
The chapter effectively proves that maximizing system performance requires a
balanced investment: reducing failure rates through better engineering,
reducing fatigue through better scheduling, and ensuring high-quality (perfect)
repairs through better training and tools.
This also highlights how the present study serves as the bridge between
abstract mathematical theory and the hard realities of industrial operation,
setting the stage for the final conclusions and recommendations of the
dissertation.
Reference:
1.
Alsamir, H. S., Noorani,
M. S. M., & Shatanawi, W. (2019). Fixed point results in metric-like spaces
via σ-simulation functions. European Journal of Pure and Applied
Mathematics, 12(4), 13451357.
2. Amini-Harandi, A. (2012). Metric-like spaces, partial metric spaces and fixed points. Fixed Point Theory and Applications, 2012, Article 204. https://doi.org/10.1186/1687-1812-2012-204
3.
Anderson, P. M.,
Chintaluri, G. M., Magbuhat, S. M., & Ghajar, R. F. (2002). An improved
reliability model for redundant protective systems-Markov models. IEEE
Transactions on Power Systems, 12(2), 573-578.
4.
Busse, A. L., &
Moreira, J. M. L. (2021). Reliability and redundancy allocation analysis
applied to a nuclear protection system. Brazilian Journal of Radiation
Sciences, 9(2B (Suppl.)).
5.
Chellappan, C., &
Vijayalakshmi, G. (2009). Dependability modeling and analysis of hybrid
redundancy systems. International Journal of Quality & Reliability
Management, 26(1), 76-96.
6.
Chhillar, S. K., Barak,
A. K., & Malik, S. C. (2013). Analysis of a parallel system with priority
to repair over maintenance subject to random shocks. International Journal
of Computer Science Issues, 10(3, No. 2), 16.
7.
Hendawi, S., &
Frangopol, D. M. (1994). System reliability and redundancy in structural design
and evaluation. Structural safety, 16(1-2), 47-71.
8.
Houankpo, H. G. K., &
Kozyrev, D. (2021). Mathematical and simulation model for reliability analysis
of a heterogeneous redundant data transmission system. Mathematics, 9(22),
2884.
9.
Huang, H. I., & Ke,
J. C. (2009). Comparative analysis on a redundant repairable system with
different configurations. Engineering Computations, 26(4), 422-439.
10.
Institute of Electrical
and Electronics Engineers. (1963). IEEE Transactions on Reliability
(Vol. R-12, Issue 1). IEEE. (First issue published March 1963)
11.
Jaggi, D. S. (1977). Some
unique fixed point theorems. Indian Journal of Pure and Applied Mathematics,
8(2), 223230.
12.
Jia, J., & Wu, S.
(2009). A replacement policy for a repairable system with its repairman having
multiple vacations. Applied Mathematical Modelling, 33(5), 21152124. https://doi.org/10.1016/j.apm.2008.05.023
13.
Jo, J. S., Kim, S. P.,
Oh, S. G., Kim, T. J., Kang, F. S., & Park, S. J. (2024). Markov
model-based reliability analysis considering the redundancy effect of modular
converters. IEEE Access, 12, 3328-3338.
14.
Jodejko-Pietruczuk, A.,
Nowakowski, T., & Werbińska-Wojciechowska, S. (2013). Block inspection
policy model with imperfect inspections for multi-unit systems. Reliability:
Theory & Applications, 8(3), 115.