Supporting Material is available for this work. For more information, follow the link from
the Table of Contents to "Accessing Supporting Material".
Table of Contents
Go Back
Modelling Modern Maintenance
— A System Dynamics Model Analyzing
the Dynamic Implications of Implementing
Total Productive Maintenance
Jérn-Henrik Thun
Industrieseminar, Mannheim University
D-— 68131 Mannheim, Germany
+49 621 181 15 84/+ 49 621 181 15 79
thun@is.bwl.uni-mannheim.de
Abstract In recent years maintenance has become an important factor for operations
management. Total Productive Maintenance as approach for improving maintenance
has therefore evolved as one of the most popular manufacturing concepts. But ofien the
concept cannot unfold its full potential. In this paper reasons for the failure of Total
Productive Maintenance will be presented with respect to dynamic implications. The
analysis focuses on the changes for the maintenance department and the machine
operators due to the implementation of Total Productive Maintenance. Based on the
ongoing changes a dynamic analysis is performed to identify important implications for
a successful implementation of Total Productive Maintenance.
Keywords Total Productive Maintenance, improvement paradox, dynamic
implications
THE RELEVANCE OF MAINTENANCE FOR COMPETITIVENESS
Total Productive Maintenance has commonly been accepted as main concept for
improving maintenance in the context of operations management. In the last decades the
importance of Total Productive Maintenance has risen because of a dynamic
competitive environment. To succeed in a demanding market arena manufacturing
companies have to fulfil several requirements. One crucial aspect is a reliable
manufacturing process.
In the past, maintenance has become a major issue of operations management, because
of the need for high machine reliability in a demanding environment. A paradigm shift
has evolved concerning the importance of several success factors. Nowadays, cost,
quality, and time must be seen as main competitive success factors, which have to be
considered simultaneously. As a consequence, manufacturing companies must strive for
a superior cost position on the one hand. High process efficiency realized by high
production volume and capacity utilization are crucial aspects for cost reduction. As a
result the importance of maintenance has risen because of its potential to guarantee a
failure-free functioning of the machines in use to enable high process efficiency. On the
other hand, quality aspects like product variety, reliability, and longevity are important
as well. Accordingly, the production on a high quality level is necessary to meet quality
specifications. To guarantee a high process capability maintenance performance must be
enhanced as one consequence. Only machines with a high maintenance standard are
able to produce with less or no failures. Furthermore, besides cost and quality the aspect
of time is crucial. In the context of time based competition [Stalk and Hout, 1990], fast
and on-time deliveries are of great relevance to corporate success. This leads to the
necessity of a cycle time reduction of the manufacturing process with the consequential
need for a high maintenance standard to ensure fast throughput.
Altogether, it can be stated, that manufacturing companies are faced with the need for a
dependable production system. Thereby, the most critical difficulty of the described
situation can be based on the fact, that the requirements are not mutually exclusive but
mainly cumulative [Ferdows, De Meyer, 1990]: Manufacturers have to offer a great
variety of products in the least amount of time on a high quality level for an acceptable
price. As one crucial consequence, machine maintenance has become increasingly
important for manufacturing companies to accomplish these requirements, because a
high-level maintenance standard is the key for supporting challenging quality standards,
achieving high efficiency, and reaching time competence. In the light of the mentioned
requirements for operations management, Total Productive Maintenance has become
one of the most expedient approaches to guarantee high machine dependability.
THE DEVELOPMENT OF MAINTENANCE
In terms of maintenance four development stages can be distinguished. The basis, i.e.
the first stage of maintenance development, is breakdown maintenance that was the
business standard till the 1950ies. The main characteristic of this stage is the way to
cope with machine breakdowns. Actions for maintaining equipment are not undertaken
before a machine breaks down [Nakajima, 1988]. This reactive “fire fighting strategy”
is no longer appropriate in a changing environment.
So the second stage starts with the introduction of Preventive Maintenance since the
beginning of the 1950ies — following the development at General Electric. This
approach is different from the first one in the way that the aim is to strive for a reduction
of down time in advance due to a better planning of maintenance activities. Another
aspect of this stage is corrective maintenance developed in 1957, i.e. measurements
improving the equipment in terms of dependability due to maintenance activities.
In the 1960ies the approach of Maintenance Prevention was introduced. Development
aspects are included in maintenance activities. The effort for necessary maintenance
should be decreased by a better design and planning in the development and purchasing
of machines. The third development stage deals with Productive Maintenance as an
integrative approach for the different maintenance activities that the maintenance
department still has the responsibility for. The fact that the maintenance department was
more and more overloaded with maintenance tasks simple maintenance tasks were
assigned to the machine operators. This was the basis for the fourth stage, the
development towards Total Productive Maintenance [Nakajima, 1988; Nakajima, 1989].
The overall goal of Total Productive Maintenance is to raise the overall equipment
effectiveness [Shirose, 1989]. The overall equipment effectiveness is calculated by
multiplying the availability of the equipment, the performance efficiency of the process
and the rate of quality products [Dal, Tugwell, and Greatbanks, 2000; Ljungberg, 1998].
This measure can be used as an indicator for the dependability of the production system.
Nowadays, five pillars have evolved as a standard of Total Productive Maintenance;
they will be described briefly in the following: The first pillar acts on the “six big
losses” [Shirose and Goto, 1989]. Primary malfunctions are identified and eliminated in
an initial setup project. This is done by project teams consisting of maintenance staff,
machine operators, and engineers. The core of the second pillar is a scheduled
maintenance program. Maintenance activities should be done on a regular basis
following a given time schedule to realize the approach of preventive maintenance
[Ainosuke, 1989]. The third pillar deals with the development of an autonomous
maintenance program [Goto, 1989a]. Autonomous maintenance may be the most
ambitioned part for implementing Total Productive Maintenance because it depends on
shop-floor operators’ commitment. Following the approach of autonomous maintenance
workers perform simple maintenance tasks like cleaning and lubricating. The fourth
pillar of Total Productive Maintenance is training, because maintenance activities
formerly done by the maintenance personnel are assigned to the machine operators.
Therefore, operators need to have a better understanding of the machines and build up
knowledge about maintenance activities [Aso, 1989]. Maintenance prevention, the fifth
pillar of Total Productive Maintenance, strives for making maintenance activities
unnecessary or easier by developing and purchasing “maintenance-free” machines. The
aim is to raise equipment dependability, maintainability, and the ease of operation (Goto,
1989b). The basis of Total Productive Maintenance is the 5S-Programm. The 5S-
program supports the pillars of Total Productive Maintenance, because a tidy and clean
working environment fosters the “Parlor Factory” [Nakajima, 1988].
DYNAMIC IMPLICATIONS OF TOTAL PRODUCTIVE MAINTENANCE
Implications for the Maintenance Department
The discussion of the history of maintenance has shown that a fire-fighting maintenance
strategy in terms of reactive maintenance leads to unexpected machine breakdowns.
Furthermore, the maintenance department is busy most of the time repairing machines.
It does not have the time to do maintenance tasks on a regular basis nor does it have the
time to improve the maintenance system within the production process. This leads to the
fact that preventive maintenance tasks are neglected resulting in more machine
breakdowns. Machine breakdowns eat up the maintenance department’s capacity to
maintain or improve the production system on a regular basis. In the long run, the
vicious circle “Repairs eat up Prevention” results in a situation with many unexpected
machine breakdowns and an overloaded maintenance department. This is the crucial
behaviour of a production system without maintenance free machines and an overloaded
maintenance department.
Time for.
ia |, Repair \
Machine Time for
Breakdowns Maintenance
“Repairs eat up
Maintenance”
Maintenance _
Backlog
Figure 1: The “critical” Feedback Loop of “Breakdown Maintenance”
The implementation of Total Productive Maintenance implies several changes
concerning maintenance. One aspect is a change in responsibilities. As stated before,
due to the implementation of autonomous maintenance simple maintenance tasks are
assigned to machine operators. This leads to the fact that the maintenance department is
relieved, because simple maintenance tasks are done by machine operators. Thereby the
vicious circle will be broken through. A main consequence is that the maintenance
department is not overloaded with fire-fighting activities but can act on preventive
maintenance and necessary improvement. Furthermore, improvement by maintenance
prevention and training of machine operators can be gained.
The New Role of Workers by the Implementation of Total Productive Maintenance
As a consequence of the change of responsibilities machine operators must be trained to
be able to fulfil the new maintenance requirements. Due to the fact that the maintenance
department is relieved, simple maintenance tasks like lubrication must be done by the
machine operator himself. Although the maintenance tasks transferred should be easy,
there will be still a lack of knowledge concerning the know how to fulfil these tasks. So
the machine operators must be trained. The training must be carried out by the
maintenance department to guarantee a sufficient maintenance level of the machine
operators.
The transference of the simple maintenance tasks has two implications regarding the
machine operators. As stated before machine operators must be trained in order to learn
managing the assigned maintenance tasks on the one hand. Thereby, it has to be
considered that learning processes can not be done overnight. They are time consuming,
thus the existing lack of knowledge will be reduced gradually and not immediately. But,
in the long run, the learning process leads to the fact that operators achieve a higher
understanding of the functioning of the machine. Accordingly, they can give insights
about their day to day work for the improvement of maintenance activities, i.e.
contributing to a better design with respect to maintainability.
On the other hand, machine operators might get a feeling of being overstrained. First of
all, the maintenance tasks hinder them from fulfilling their day to day work load,
because the sum of the additional workload and the normal workload are too much.
Furthermore, production pressure initiated by the management is problematic. The
problem works analogously to the “production pressure chokes off PM” cycle described
by Maier [Maier, 2000]. The willingness of the machine operators to perform
maintenance tasks will decrease. Secondly, machine operators might be swamped with
the new maintenance tasks with the consequence that they refuse to do the necessary
maintenance activities. Both aspects lead to the fact that maintenance will not be done
properly. As a consequence the machine operators must be trained to be able to fulfil the
simple maintenance tasks faster and more accurate [Aso, 1989]. Training and patience
must be seen as the remedy against refusal and overstrain.
Implications for Maintainability
The implementation of Total Productive Maintenance goes along with several other
implications. One important point is that the maintenance department has more capacity
to fulfil preventive maintenance tasks, thus the maintenance schedule is done properly.
By assigning simple maintenance tasks to machine operators the vicious circle will be
broken through as stated before. If the maintenance department has the time to do
preventive maintenance activities on a regular basis, machine breakdowns will decline.
This sets free time capacity of the maintenance department. Accordingly, the
maintenance department can work together with engineers elaborating valuable results
with regard to maintainability. In the long run, an improvement of machine
maintainability will decrease the need for simple maintenance tasks. Additionally,
maintenance prevention will make maintenance tasks easier, thus more maintenance
tasks can be assigned to the worker. Furthermore, machine operators will develop a
better understanding of machines fulfilling simple maintenance tasks in the long run.
This understanding can lead to valuable hints for improving machine maintainability as
well. Figure 2 depicts a causal loop diagram of the maintenance system.
total amount of basic
mainienance fask that
a nist be done
maintenance —<
fiee equipment ; _7~. _ basic maintenance tasks
basic maintenance done by maintenance
4 pn tasks done b} department
legree 0} machine operators
autonomous sae
maintenance: oe
” maintenance
lepartments time
lack of for training
Py training
1
|
operators breakdowns
4. necessary repair done by »
—e time for maintenance
training amount of “department
| breakdowns
- Overall Equipment wmsit cance
maintenance Bi (et
prevention BEE eu venese ane: departments
an time
probability of
¥ "breakdowns
Machine preventive +
Availability maintenance done
: following a
maintenance llowing a
backlog schedule
= amount of
planned
maintenance
Figure 2: Causal Loops of Total Productive Maintenance
The described changes due to the implementation of Total Productive Maintenance
result in a counterintuitive behaviour of the underlying system. This dynamic behaviour
can lead to a misunderstanding of the system, thus wrong decisions in terms of
maintenance might be the consequence. Despite the great potential of improvement
programs, e.g. Total Productive Maintenance, for operations management, most
attempts to use them have ended in failure which is described by the “improvement
paradox” [Keating, Oliva, Repenning, 1999; Sterman, Kofman, and Repenning, 1997].
Therefore, a system dynamic model will be useful showing the dynamic behaviour of
maintenance.
ANALYZING THE DYNAMICS OF MAINTENANCE
In the following a system dynamic model will be introduced to analyze the dynamics of
maintenance [Forrester, 1961; Forrester, 1971; Sterman, 2000]. In a first step the initial
model will be described. This model is the basis for the analysis of the consequences of
the implementation of Total Productive Maintenance. A major assumption is that
machines are maintenance free in terms of simple maintenance tasks, i.e. maintenance
tasks are not necessary to keep machines running. Machines just have to be maintained
on a regular basis. The necessary amount of regular maintenance can be done by the
maintenance department. In this situation there are no machine breakdowns because
machine maintenance is done properly. This leads to a model in an equilibrium state.
Figure 3 depicts the basic structure of the maintenance model [Sterman, 2000].
wear and tear by
Operations
one Equipment
ie Defect Defects Defect
Collateral Disvention’ Elimination
Damage through PM through Repairs
overall Breakdown
quipment ate
effectiveness DEGREE OF
ad REACTIVE
MAINTENANCE
breakdown
probability
Backlog
O—— >} Maintenance ee)
maintenance Tasks i nipnane
Time’ tasks per week ntenancé
Neen
Figure 3: The Basic Maintenance Model
This model will be extended with the ongoing changes initiated by the implementation
of Total Productive Maintenance. First, the assumption of maintenance free machines
will be given up. As a consequence simple maintenance tasks must be done. This leads
to the fact that the maintenance department is overloaded. Secondly, autonomous
maintenance is introduced. Simple maintenance tasks are assigned to machine operators.
Accordingly, machine operators have to be trained. Finally, the approach of
maintenance prevention is embodied into the model, thus the amount of necessary
maintenance tasks decreases. Based on the simulation of the maintenance model critical
aspects can be identified. The following figure shows the model including the changes
by the implementation of Total Productive Maintenance.
_wear and tear by
Damuge
Loading Time
limitation
eae
ee | Brrgiciown py
a DEGREE OF \\
N j Pd. TN
\ \ ‘“+— MAINTENANCE \
| \ \
| 1 \
Preventive see \
s me gape
aro Minne SER \ es
=a = en ae a
eae swrrcn 7 FREE DEGREE \ Ivtinensnce Mareenace|
Mbimerance | preventive] reve | MACE, \ Prevention | npinianes | Prevtion
Guess Sens MACHINES pill wan
== brealciown tasks _ ra
KICKOFE | peal — wens
ian tae | "ay y ei REVENTION —_SWITCHMP
RATE or mM
PREVENTIVE Sern
MAINTENANCE | sw +
\ RATIO |
\ Ste vmistrnce 4 swre ;
4 tasls per week | shee sree
maintenance ue
mmc STEPTIVE
AM RATE 1 ng SWITCH TR
Ne vit RATE |
| /
i } cance |. Taining Potential
swan [dou Ee
girndon ones | GORE oases
= 4
e acer
_ TR
Figure 4: Total Productive Maintenance Model
Several simulation runs show the dynamic behaviour of the model. The following figure
depicts the behaviour of the overall equipment effectiveness.
Overall Equipment Effectiveness
100
95
90
_s
| 44
ly
=
es aa
8 [a
80
XQ aA tT
See =
3 gue
75 ——
. +
Lec
|
70
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140
‘Time (Week)
‘overall equipment effectiveness : Traditional
overall equipment effectiveness
overall equipment effe
overall equipment effectiveness
ss : Total Productive Maintenance
Maintenance Prevention
Preventive Maintenance
Basic Run
Figure 5: Simulation runs of Total Productive Maintenance
The simulation shows that the particular approaches of Total Productive Maintenance
have a different impact on the behaviour of the system (see figure 5). In a test
simulation run the system is run with 100% maintenance free machines — there is no
necessity for simple maintenance tasks. The overall equipment effectiveness reaches an
equilibrium after some periods shown by the simulation run “Basic Run”. In a second
run the assumption of “miracle machines” is given up. In a system with realistic
machines maintenance tasks have to be fulfilled. Because of the limited capacity of the
maintenance department maintenance tasks can not be done appropriately. This leads to
the effect that machine breakdowns will occur leading to necessary repairs and therefore
to the “critical” reinforcing feedback loop of “Breakdown Maintenance” depicted in figure 1.
The time used for repairs prevents the maintenance department from doing maintenance
tasks resulting in further unexpected machine breakdowns.
Following the idea of preventive maintenance scheduled maintenance tasks are done by
the maintenance department. Due to the fact that their capacity is limited machine
breakdowns still happen because simple maintenance tasks are not done properly
anymore. The vicious circle “repair eats up maintenance” is still running bringing the
overall equipment effectiveness down. Accordingly, to implement preventive
maintenance effectively simple maintenance tasks have to be assigned to machine
operators in order to disburden the maintenance department. To enable the machine
operators to fulfil the maintenance tasks they must be trained as well. Otherwise a
backlog of simple maintenance tasks will still be accumulated. The resulting simulation
run “Preventive Maintenance” shows that the overall equipment effectiveness will first
decrease for a short period of time and than stabilize on higher level. This situation can
be characterized by a “worse before better”-effect, because before the overall system
can be improved machine operators need to be trained.
In a next step maintenance prevention is done exclusively. Preventive Maintenance,
autonomous maintenance, and training is not considered. The simulation run
“Maintenance Prevention” shows that the overall equipment effectiveness will decrease
in the first periods. In the long run the overall equipment effectiveness will increase. A
reason for that is that less maintenance tasks have to be done improving the
maintenance system in the long run on the one hand, but still repairs must be done
because of missing preventive maintenance on the other hand.
Finally, as well preventive maintenance accompanied by autonomous maintenance and
training as maintenance prevention is incorporated into the model following the idea of
Total Productive Maintenance. The simulation run “Total Productive Maintenance”
shows that the system achieves the highest overall equipment effectiveness in the long
run. Conclusively, for a successful implementation of Total Productive Maintenance it
seems to be necessary to understand the functioning and the interaction of the different
facets of Total Productive Maintenance thus the concept can unfold its whole potential.
CONCLUSION AND FURTHER RESEARCH
In the paper dynamics of an implementation of Total Productive Maintenance are
discussed within the framework of a system dynamics model. Gradually, the pillars of
the concept are built into the model. Simulation runs show that the pillars have a
different impact on the systems behaviour. By the implementation of Total Productive
Maintenance the performance of the overall system might be worse in the beginning but
improve in the long run depending on the management of maintenance.
For further research the interplay of maintenance prevention and preventive
maintenance should be analyzed concerning their impact on the overall equipment
effectiveness. Furthermore, the potential of “maintenance ease”, a facet of maintenance
prevention, should be investigated.
REFERENCES
Ainosuke, M. (1989): “Preventive Maintenance”, in: Nakajima, Seiichi (Ed.): TPM
Development Program — Implementing Total Productive Maintenance, Cambridge/MA,
pp. 85-164.
Aso, M. (1989): “Maintenance Skill Training”, in: Nakajima, Seiichi (Ed.): TPM Development
Program — Implementing Total Productive Maintenance, Cambridge/MA, pp. 85-164.
Dal, B., Tugwell, P., and Greatbanks R., (2000): “Overall equipment effectiveness as a measure
for operational improvement”, in: International Journal of Operations and Productions
Management, Vol. 20, No. 12, pp. 1488-1502.
Ferdows, K. and De Meyer, A., (1990), “Lasting Improvements in Manufacturing Performance:
In Search of a New Theory”, Journal of Operations Management, Vol. 9, pp. 168-184.
Forrester, J. W. (1961), Industrial Dynamics, Cambridge/MA.
Forrester, J. W. (1968), Principles of Systems, Cambridge/MA.
Goto, F. (1989a): “Autonomous Maintenance”, in: Nakajima, Seiichi (Ed.): TPM Development
Program — Implementing Total Productive Maintenance, Cambridge/MA, pp. 165-218.
Goto, F. (1989b): “Maintenance Prevention”, in: Nakajima, Seiichi (Ed.): TPM Development
Program — Implementing Total Productive Maintenance, Cambridge/MA, pp. 85-164.
Keating, E. K., R. Oliva, N. P. Repenning, S. Rockart, and J. D. Sterman (1999): “Overcoming
the Improvement Paradox”, in: European Management Journal, Vol. 17., No. 2, pp. 120—
134.
Ljungberg, O., (1998): “Measurement of overall equipment effectiveness as a basis for TPM
activities”, in: International Journal of Operations and Productions Management, Vol. 18,
No. 5, pp. 495-507.
Maier, F. H. (2000): “Feedback Structures Driving Success and Failure of Preventive
Maintenance Programs”, in: Van Dierdonck, R. und A. Vereecke (Edit.), Operations
Management: Crossing Borders and Boundaries: The Changing Role of Operations — 7"
International Conference of the European Operations Management Association, Vlerick
Leuven Ghent Management School, Belgium, pp. 376-383.
Nakajima, S. (1989): “An Introduction to TPM”, in: Nakajima, Seeiihi (Ed.): TPM
Development Program — Implementing Total Productive Maintenance, Cambridge/MA.
Nakajima, S. (1989): TPM Development Program — Implementing Total Productive
Maintenance, Cambridge/MA.
Nakajima, S., (1988): Introduction to Total Productive Maintenance (TPM), Cambridge/MA.
Shirose, K. (1989): “Equipment Effectiveness, Chronic Losses, and Other TPM Improvement
Concepts”, in: Nakajima, Seiichi (Ed.): TPM Development Program — Implementing
Total Productive Maintenance, Cambridge/MA, pp. 27-84.
10
Shirose, K. and Goto, F. (1989): “Eliminating the Six Big Losses”, in: Nakajima, Seiichi (Ed.):
TPM Development Program Implementing Total Productive Maintenance,
Cambridge/MA, pp. 85-164.
Stalk, G. and T.M. Hout, (1990): Competing Against Time: How Time-based Competition is
Reshaping Global Markets, New York.
Sterman, J. D. (2000), Business Dynamics — System Thinking and Modelling in a Complex
World, Boston et al.
Sterman, J. D., Kofman, F. and Repenning, N. P. (1997): “Unanticipated Side Effects of
Successful Quality Programs: Exploring a Paradox of Organizational Improvement”,
in: Management Science, Vol. 43, No. 4, pp. 503-521.
11
Back to the Top
CC BY-NC-SA 4.0