Reliability Engineering and System Safety 91 (2006) 444–451

www.elsevier.com/locate/ress

Maintenance strategy based on a multicriterion

classification of equipments
Félix C. Gómez de León Hijesa,*, José Javier Ruiz Cartagenab,1
a
Department of Mechanical Engineering, University of Murcia, Edificio C, Campus de Espinardo 30071, Murcia, Spain
b
Department of Chemical Engineering, University of Murcia, Faculty of Chemistry, Campus de Espinardo 30071, Murcia, Spain
Received 30 November 2004; received in revised form 28 February 2005; accepted 2 March 2005
Available online 6 May 2005

Abstract
In an industrial plant, the level of maintenance provided to individual equipment is directly related to the availability that is expected from
it. Thus, it is hoped that the most critical equipments will not fail or, at least, that any failure will be rapidly detected and corrected in the
minimum time possible. Since resources are limited, it is necessary to determine how they should be distributed, so that no important
equipment remains neglected while more resources are concentrated on the most critical items. Therefore, it is necessary to classify
equipment in an objective way according to its importance. The method of multicriterion classification of critical equipments (MCCE)2,
which is described in the present work, allows systematic and detailed quantification of the criticality of all equipment, that is to say, it
provides an evaluation of the importance that its correct operation has for the plant. To provide this information, the consequences for a
company of any failure in the equipment concerned are analysed. Finally, a real case example of an urban wastewater treatment plant is
described, in which the MCCE method is applied.
q 2005 Elsevier Ltd. All rights reserved.

Keywords: Maintenance; Reliability; Critical equipment; Criticality; Decision support

1. Introduction and operation cannot always be carried out simultaneously.

Therefore, the development of maintenance strategies must
Any maintenance action in an industrial plant is take account that resources are limited and, therefore,
designed, in the first instance, to ensure the correct operation maintenance will be imperfect [2].
of equipments and to obtain the greatest availability as Several methods have been proposed for planning
possible. Not in vain does J.P. Souris [1] affirm that maintenance activities based on imperfect maintenance.
availability is the result of maintenance. Maintaining the Among these, Whitaker and Samaniego [3] proposed a
optimal functional condition also implies ensuring the method based on an evaluation of reliability. Martorell el al.
safety of the plant. The reasons for this are so sound that it [4,5] suggested that preventive maintenance activities
might be wondered why the maximum level of maintenance should depend on the age of the components and proposed
cannot be ensured for all the equipments in a plant. age-dependent models for evaluating the risks and costs of
However, this is unviable both from a economic point of maintenance; while Chelbi and Ait-Kadi [6] analyzed some
view, because an unlimited increase in the level of inspection strategies for randomly failing equipment.
maintenance would involve unnecessary costs, and With this initial determining factor, the design or
from an operative point of view, because maintenance reorganization of a maintenance plan must begin with an
exhaustive study of all the equipments and facilities, the
* Corresponding author. Tel.: C34 968 367 329; fax: C34 968 364 189.
purpose of which is to obtain all the information necessary
E-mail address: [email protected] (F.C. Gómez de León Hijes). to justify and analyse the viability of each maintenance task;
1
Tel.: C34 968367191; fax: C34 968364189. [email protected]. in other words, to decide on the best maintenance strategy
2
El orden de las iniciales cambia con respecto a las iniciales en for the plant. Such a study must begin with a detailed
castellano (CMEC). inventory of equipments and facilities, including their
0951-8320/$ - see front matter q 2005 Elsevier Ltd. All rights reserved. characteristics and functional interrelationships; records of
doi:10.1016/j.ress.2005.03.001 past failures, if they exist; the cost of acquisition and supply;
 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451 445

the direct and indirect cost of maintenance, if this are termed ‘criticality criteria’. The existence of alternative
information is available; needs and operational factors; the equipment, the cost of the equipment, the risk to operators
type of maintenance to be carried out and any legal or and any repercussions on process continuity are only some
contractual obligations concerning maintenance (such as of the aspects that must be considered in order to determine
periodic inspections subject to regulation, guarantee the consequences of failure.
periods, etc.); the legal or contractual obligations of the Different methods have been formulated for identifying
company, such as those related with legal sanctions or the critical equipments of a process, some centre exclusively
penalties regarding the quantity or quality of the production; on the effect of failure on the service [8] and others are based
the means (tools, auxiliary equipment, etc.) available for on the involved risk, such as the HAZOPs (Hazard and
maintenance; human resources and the qualifications of the Operability Studies) method [9], or safety equipment [10,
personnel available; the maintenance tasks that can or must 11]. In other cases, the aim is to classify the maintenance
be contracted out and any other aspects that are relevant for activities to be carried out rather than to classify the
the case in question. equipment itself, e.g. the FMEA (Failure Mode and Effects
Finally, the projected maintenance plan must determine Analysis) method, where, by means of the Failure Risk
the resources that will be destined to this purpose and the Priority Number [12], maintenance activity is determined
maintenance strategies that will be followed, particularly and prioritised.
the kind of maintenance to be executed on given equipments In the chemical industry, equipment is frequently
and a programming of each of the foreseen activities. Since classified bearing in mind both functional and safety
resources will be limited, it is obvious that, according to the aspects. As regards the first criterion, equipments are
Paretians principles, they must be assigned in such a way distinguished according to whether failure (A) brings the
that the most important equipments should be given priority. installation to a stop, (B) decreases its productive capacity
or (C) reduces or threatens to reduce quality. Therefore,
only three degrees of criticality are established. This is the
2. Identification of critical equipments well know ABC classification [13], based on the ABC
Pareto curve [14]. Other methods of classification used in
The unreliability of equipments is the essential reason for the Petrochemical industry subdivide the three above
maintenance. The actions which need to be carried out on all categories into more intermediate categories, taking into
equipment to improve its extrinsic reliability, which comes consideration the possible existence of alternative or reserve
from its correct use and maintenance, is a decision that must equipment [15].
be taken. Jayabalan and Chaudhuri [7] point out that The above are simple but objective methods for system-
maintenance ‘rejuvenates’ a system, in other words, returns atizing identification of the most critical equipments,
part of the lost reliability to the system. Obviously, the more although other factors of undoubted importance are omitted
important (that is to say critical) an equipment, the greater from the analysis.
its level of maintenance should be.
The starting point is to determine the criticality of all
equipment, that is, its importance in the objectives for which 3. Equipment classification according to its criticality
it is used and, therefore, to the company. This provides a
degree of priority as regards the level of maintenance 3.1. Fundamentals
needed or, in other words, the priority of every equipment
when deciding how the resources destined for maintenance The above mentioned methods catalogue equipments
are to be distributed. into three or more categories but do not solve the problem of
Unfortunately, the criticality of an equipment cannot be classifying equipments in order of importance within each
deduced directly from the information obtained in the of the established categories and, obviously, the lower the
preliminary study, because the purely technical information number of categories, the more serious the problem. It must
and functional interrelationships, although essential, do not not be forgotten that the fundamental objective of criticality
take into account all the aspects that may be of importance analysis is to establish as detailed a classification as possible
to a company. For example, the risk derived from the of equipments, which will serve as a reference for deciding
potential failure of the equipment undoubtedly occupies an the maintenance strategies to be followed and how these
important place. Other criteria, such as the cost of the strategies deal with all equipment.
equipment, the policy of suppliers, the reliability of the If there are a few equipments to be analysed, any
equipment, its maintainability, etc. are also important and technician of the plant will be able to classify them into
must be considered. Estimating all the consequences categories with no need of complicated analytical consider-
resulting from equipment failure constitutes the principal ations. However, things are more complicated when the
focal point of the analysis to be carried out. number of equipments is high. In this case, many functional
All those aspects that, to a greater or lesser degree, are interrelationships will exist and other criteria that will
related with the consequences of equipment failure undoubtedly have some sort of economic, functional or
446 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451

organizational impact on the company must be taken into Let ca, cb. ci. cn be the set of criteria, in a random
account. These criteria will include the mean time between order, that the experts of the plant in question consider must
failures (MTBF), the time that might pass before the be evaluated when determining the relative importance of
detection of the failure; the mean time to repair (MTTR) or an equipment. The procedure aims to obtain a list of criteria,
the cost of the repair, to mention only some of the well ordered from least to most import, out from any disordered
known criteria that influence the criticality of the list of the same. To make the process more intuitive, both
equipment. lists are now placed vertically.
In such a case, an objective and systematic method of The procedure starts by taking the first element of the
classification will not only facilitate the task but also make unarranged list, ca, and placing it as the first and, for the
clear the criticality, sometimes unsuspected, of certain moment, only criterion of the arranged list. The criteria of
equipments. The multicriterion classification of critical the unarranged list are taken one at time and their relative
equipments (MCCE), which is presented in the present importance is compared with that of the criteria already
work, quantifies the criticality of every equipment by giving existing in the arranged list, beginning always from the top
it a value that we call the criticality index (IC). of the list, where the least important criterion is placed, that
The MCCE method is based on the fact that, under is to say, the criterion with the least weight. The criterion
normal circumstances, equipment criticality is not fixed by being analysed ‘sinks’ into the arranged list until it reaches a
simply evaluating one factor, but is the result of the position in which its weight, that is to say, its level of
interrelated action of multiple factors, each of which has its importance, is greater than the previously compared
own importance. When equipment criticality is analysed, all criterion but less or equal to that of the following one. For
the criteria that may be relevant for a particular company example, in studying a certain plant, it may be decided that
should be included, always bearing in mind the conse- the criterion ‘Decrease in production’ is more important
quences of any failure in that equipment. Since, in general, than the criterion ‘Cost of the equipment’, but less important
the level of importance of each of the considered criteria than ‘Penalty for lack of product quality’.
will be different in calculating the criticality index of all If the importance of two criteria is similar, they are
equipment, every criterion is affected by a weighting factor placed at the same level (height) in the list and separated by
representative of its importance, which we shall refer to as a comma, a dash or any other separating character. The
weight. order of criteria at the same level is unimportant, but once a
Deciding the weight of a criterion with respect to criterion has been placed it must not be altered because later
others is not without difficulty. For example, different in the procedure a different subindex will correspond to each
criteria may have the same importance, although there criterion. The obtained list adopts the form of expression
may not always be unanimity in the minds of experts (1), in which n criterion have been considered and arranged
within the company. In fact, if different departments of in a global list that contains m levels (l1, l2,., lm), where,
the company are asked to rank a list of criteria, these obviously, m%n.
will seldom coincide exactly since their points of view
l1 Þ c1 ; c2 ; c3 ; .ci (1)
generally differ.
l 2 Þ cj ; ck ; .
3.2. Criterion arrangement
lm Þ .; cn
A compromise solution may be reached by using some of
the typical procedures of decision theory [16], particularly As can be observed, this process of successive compari-
Saaty’s Analytical Hierarchial Process (AHP) [17,18] and sons provides the relative weight of each criterion according
subsequent developments [19]. Nevertheless, since the to its importance, but not its numerical value. The result is a
number of criteria is not excessively high, the procedure list of criteria arranged vertically in several levels from least
used to establish their relative importance (their weight) is to most importance, in which some levels may be occupied
the direct insertion procedure, which is similar to the by several criteria of equal importance.
procedure that card players use to arrange their cards, Since, in most cases, the set of analyzed criteria ranges in
although with certain differences in order to adapt the number between 10 and 20, and many of them are probably
method to the needs of this case. of similar importance, the final list will frequently have no
In the procedure the importance of each criterion with more than 4 or 5 levels, each of which represents a different
regard to every other is compared, but following a certain level of importance.
order. If a criterion is more important than another, then its
weight will be greater, even though its value is still 3.3. Assignment of weight to each level
unknown. This is one of the differences compared with
the classic procedure of direct insertion [20], in which the Once the initial set of criteria has been classified into
elements of the list have a numerical value that allows its categories of different importance (the same number as
arrangement. there are levels in the arranged list), it is necessary to specify
 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451 447

the weight of each one of them, that is to say, the weighting It must be borne in mind that it is dealing with qualitative
factor that every criterion will have to calculate the criteria referring to different aspects of the equipment and if
criticality index. Evidently, the weight of every criterion comparison is not easy, then it is even more difficult to find a
situated at the same level will be identical, since this was the numerical value that exactly reflects its importance. There-
reason for dividing the list into levels. fore, the weight obtained always refers to a specific plant in
In order to simplify the evaluation of the relative given circumstances, and cannot be extrapolated to other
importance of the different levels, a value equal to unity is cases.
assigned to the criteria of the first level, l1, which are, as When important modifications are carried out in the plant
already explained, those of least weight. That is to say, or in the process (such as changes in the operations or
w(l1)Z1. conditions involved in a process, in the tasks performed by
Next, the relative importance must be valued of each one the operators, or in the level of supervision, etc.), the best
of the levels compared whit the others. Thus, for example, it alternative is not only to calculate the criticality of the
might be estimated that the relative importance of the equipments again but also to carry out a complete review of
second level criteria (l2) is the double (triple, quadruple, the weight initially assigned to each criterion, including
etc.) of those of the first level (l1), in other words: the weight checking whether or not the criterion analysed has been
of a level 2 criterion is approximately equivalent to the sum changed.
of two (three, four, etc.) level 1 criteria.
Supposing, for example, it has been determined that
3.4. Evaluation of equipment criticality according
w(l2)Z2, that is to say, the criterion of level 2 weighs twice
every criterion
as much as a level 1, two references are now available to
value the relative importance, and therefore the weight, of
Now is the moment to analyse all equipment by reference
any level three criterion (l3). The objective now is to find a
to the different criteria. For example, when the ‘The cost of
linear combination of criteria from the previous levels that
equipment ’ is analysed, it may occur that a given equipment
has the importance equivalent to a level 3 criterion.
is very expensive, whereas another is not. Logically, if it is
In this way, following on the previous example, it could
considered only this criterion, the first equipment is more
be estimated that a level 3 criterion is equivalent to the sum
critical than the second and so it should be assigned a higher
of one level 2 criterion and another of level 1, which means
score. Another typical example is the ‘mean time to repair
that w(l3)Z3; or, if this is insufficient, an equivalence might
(MTTR)’ and an equipment with an MTTR of 2 h will be
be established with two level 2 criteria, in which case
less critical than another with an MTTR of 20 h value, but
w(l3)Z4; and so on until the correct equivalence between
only with regard to this criterion.
every level of the list is established.
To follow a common norm, the range of possibilities of
This process of successively comparing the criteria,
every criterion is divided into the same number, d, of
besides fixing the weight of each of them, may uncover a
degrees, or categories, which reflect the different degrees
mistake in the position occupied by a given criterion within
of criticality possible for one equipment with respect to the
the arranged list, that is to say, in the initial appraisal of its
criteria analysed, for example, division into five degrees of
importance. In such an event, the criterion is simply moved
criticality (VERY HIGH, HIGH, NORMAL, LOW or
up to the corresponding level, even if this involves inserting
VERY LOW), with associated values of 4, 3, 2, 1 and 0
or eliminating a level.
points, respectively. Depending on the kind of criterion,
In practice, it may be more convenient and meaningful to
each of these categories can be formulated differently, as is
analyse and place the most important criterion first and
shown in the example presented at the end of this article.
finally analyse the least important one. This is no problem
All equipment will be analyzed from the standpoint of
and merely involves inverting the order of the list.
each of the criteria following the order established in the
Since every level and, consequently, every criterion is
criterion vector. The result is a list of n values, which we call
already associated with a known weight, it is not absolutely
criticality vector, corresponding to the degree of criticality
necessary to maintain the list structure of the levels. Indeed,
reached by the equipment for each one of the n criteria. All
the expression may be simplified into a vector, which we
equipment has its own criticality vector.
shall call the criterion vector.
Criterion vector : ðc1 ; c2 ; c3 ; .; ci ; .; cn Þ (2) 3.5. Determination of criticality index
This vector is accompanied by another vector of identical
length whose elements are the weights of each one of the Although not strictly necessary, it is advisable, as in other
criteria. In this vector, which we call the weighting vector, methods of multicriterion decision, to compare the numeri-
every element, wi, represents the weight of the criterion ci of cal evaluation finally obtained, in this case the criticality
the criterion vector. index, with a characteristic value, that is, to normalize the
results. In the MCCE method, normalization refers to the
Weighting vector : ðw1 ; w2 ; w3 ; .; wi ; .; wn Þ (3) highest possible value. The obtained value will be in
448 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451

percentage terms, which, besides simplifying the 4. A real example

expression, are easily interpretable.
Therefore, the equipment criticality index is a value 4.1. Application of the MCCE method
between zero and one hundred, and takes into consideration
the joint effect of both the criticality degree and the weight
of each criterion. It is calculated by the following 4.1.1. Step 1
expression: Analysis of the inventory showed that 54 equipments
were to be considered. Every equipment sometimes
Pn consisted of several machines, mechanisms, working
iZ1 ðdi !wi Þ devices or control instruments.
IC Z 100 ! Pn (4)
d! iZ1 wi
4.1.2. Step 2
Where: The following criteria were considered relevant to
n is the number of criterion determining the criticality of the above equipments:
d is the number of criticality degrees of the criterion, ca: effect of failure on the service.
di is the criticality degree of a certain equipment cb: where and when a failure might be detected.
according to the criterion ci, cc: state of depuration of the effluent.
wi the weight of the criterion ci. cd: potential risk for plant operators.
As was to be expected, the denominator of expression ce: existence of alternative equipment.
(4), which is the result of multiplying the number of cf: functional regime of the equipment.
criticality degrees by the sum of all the weighting vector cg: other elements of the plant that may be affected.
elements, is the highest score that an equipment may reach ch: labour effects.
before being normalized, whereas the numerator of ci: time necessary to restore the service.
expression (4) is the scalar product of the equipment cj: mean time to repair (MTTR).
criticality vector and the weighting vector. ck: cost of the repair.
In short, to apply the MCCE, the steps to be followed are: cl: mean time between failures (MTBF).

1. Elaborate a list with all the equipments that should be 4.1.3. Step 3
analyzed. Using the direct insertion method, the arranged list of
2. Fix the set of criticality criteria to be analysed. criticality criteria, from least to most importance, was as
3. Order the list of criteria, according to levels of follows:
importance (expression (1)). l1) cf, cg, ch, cj, ck, cl,
4. Find the weight, or level of importance, for every level l 2) c e, c c,
in the list. l3) ca, cb, ci,
5. Invert the list order to obtain the criterion vector l 4) c d
(expression (2)).
6. Form the weighting vector (expression (3)).
4.1.4. Step 4
7. Divide the range of possibilities of every criterion into d
The following weights (Table 1) were assigned to each
degrees of criticality.
level, that is, to each criterion, according to its relative
8. Determine the degree of criticality for all equipment
importance
studied, according to each of the criteria.
9. Calculate the criticality index, IC, of all equipment
(expression (4)). 4.1.5. Step 5
10. Arrange the equipments following the order highest to Since there are 12 criteria, the criterion vector is:
lowest criticality index. ðc1 ; c2 ; c3 ; c4 ; c5 ; c6 ; c7 ; c8 ; c9 ; c10 ; c11 ; c12 Þ
Note that steps 3 and 4, in which the level of importance Where:
of the criteria is arranged and valued, may also be carried
cd / c1 ; ca / c2 ; cb / c3 ; ci / c4 ; cc / c5 ; ce / c6 ;
out using the AHP method instead of the direct insertion
procedure used here. Integration of this method into other cf / c7 ; cg / c8 ; ch / c9 ; cj / c10 ; ck / c11 ; cl / c12 :
methodological structures has been described by several
Table 1
authors [21]. Weight assigned to each criterion
There follows a real case example of an urban
wastewater treatment plant, in which the MCCE method is Level l1 l2 l3 l4
Criterion cf, cg, ch, cj, ck, cl cc, ce ca, cb, ci cd
applied to prioritise the resources and maintenance
Weight 1 2 3 5
strategies.
 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451 449

Table 2 Table 6
Criterion c1: potential risk for plant operators Criterion c5: state of depuration of the effluent

4 Very high 4 The effluent is not depurated

3 High 3 The effluent is hardly depurated
2 Normal 2 The effluent is half depurated
1 Low 1 The effluent is almost depurated
0 Very low 0 The effluent is completely depurated. The failure has
no immediate effect on the process

Table 7
Table 3
Criterion c6: existence of alternative equipment
Criterion c2: effect of failure on the service
4 There is no alternative equipment
4 Plant operation suspended immediately
3 It is possible to provide alternative equipment shortly
3 Plant operation suspended after a short time after the failure
after the failure
2 Correct operation of the plant reduced immediately.
2 There is inferior equipment in reserve
The wastewater is not completely depurated
1 There is similar equipment in reserve
1 Correct operation of the plant reduced after a short time
0 There is redundant equipment available
after the failure
0 The operation is not affected
4.1.8. Step 8
4.1.6. Step 6 To obtain the criticality vector, the degree of criticality
According to the weights indicated in step 4 and the was calculated for all equipment according to the 12 above
criterion vector of step 5, the weighting vector is the criteria, following Tables 2–13. The meaning of every di
following: value of the vector is, therefore, the degree of criticality
selected for the criterion ci.
The results of this step were the 54 criticality vectors
ð5; 3; 3; 3; 2; 2; 1; 1; 1; 1; 1; 1Þ obtained for all the equipments. For example, a given
equipment obtained the following criticality vector:

4.1.7. Step 7 ð1; 0; 2; 1; 1; 2; 2; 1; 2; 0; 1; 1Þ

Five degrees of criticality were established for each
criterion, quantified from 0, which means that the equipment
4.1.9. Step 9
in question is not critical according to the above mentioned
Applying expression (4), the criticality index was
criterion, to 4, which means that the criticality of the
calculated for all equipment.
equipment is maximum according to the above mentioned
Hence, for the criticality vector, that were used as an
criterion.
example in the previous step, it is obtained:
The degrees of criticality, that is to say, the levels of
detail of every criterion, are shown in Tables 2–13. 27
IC Z 100 !
Table 4 96
Criterion c3: where and when a failure might be detected
Table 8
4 Downstream. The effluent water is in bad condition for Criterion c7: functional regime of the equipment
several hours
3 In the efluent water control. The effluent water is in bad 4 Continuous (100%)
condition for a short time 3 Almost continuous (75%)
2 In subsequent steps of the process, but before the effluent 2 Non-continuous (50%)
has been released 1 Non-continuous (25%)
1 In the step of the process where the failure occurs 0 Occasional
0 Detected immediately. Plant operation unaffected

Table 5 Table 9
Criterion c4: time necessary to restore the service Criterion c8: other elements of the plant that may be affected

4 More than 3 days 4 In all cases, other elements must also be stopped and repaired
3 Between 1 and 3 days 3 Sometimes other elements must also be stopped and repaired
2 Between 5 and 24 h 2 In all cases, other equipments must be stopped
1 Between 1 and 5 h 1 Sometimes other equipments must be stopped
0 Less than one h 0 No other equipment or element is affected
450 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451

Table 10 Table 14
Criterion c9: labour effects Maintenance level according to the criticality of the equipment

4 Some operators remain with nothing to do Maintenance criticality index (Ic)

3 More than two operators temporarily perform other tasks
I Predictive on-line ICR50
2 Two operators temporarily perform other tasks
II Predictive off-line 35%IC!50
1 One operator temporarily perform other tasks
III Preventive 20%IC!35
0 Staff unaffected
IV Preventive basic 10%IC!20
V Corrective Ic!10

Table 11 maintenance to the different equipments, which in this case

Criterion c10: mean time to repair (MTTR) were those that are indicated in Table 14.
4 More than a week The purpose here of Table 14 is simply to show that the
3 From 4 to 7 days more critical the equipment, the more intensive was its
2 From 1 to 3 days maintenance, even within the same level. According to the
1 From 2 to 8 h
0 Less than 2 h
cut-off values indicated in that Table (in whose determi-
nation also intervened other factors), the maintenance
procedure would be as follows: five equipments to be
That is to say, the criticality index for the equipment of monitored (10% of the total); off-line predictive mainten-
the example in question is ICZ28%. ance applied to 11 equipments (20%); preventive mainten-
ance to 12 equipments (22%); preventive basic
4.1.10. Step 10 maintenance, that is to say, only periodic greasing, cleaning
In this phase the equipments were arranged from the and replacement of consumable components, applied to 13
highest to lowest criticality index. equipments (24%); for the 13 remaining equipments (24%
In this case the most critical equipment showed a of the total) only corrective maintenance actions were
criticality index of 62%, whereas some equipments had an contemplated, that is to say, once the failure occurs.
index lower than 10%.

4.2. Results 5. Conclusions

The result was a ranking of 54 equipments of the The multicriterion classification of critical equipments
wastewater treatment plant from most to least criticality. (MCCE), described in this work, provides a percentage
The values obtained in this classification were between 9 value for all equipment, called the criticality index, as a
and 62%. representative of its criticality. This index is calculated as a
This classification was used as a necessary reference, pondered function of multiple factors, each of which
together with other economic, technical and strategic factors constitutes a different criterion for analysing the importance
that are not the object of this work, to assign levels of that all equipment has for the company.
The criticality index allows equipment to be classified
Table 12 according to its importance for the company and constitutes,
Criterion c11: cost of the repair
therefore, a basic reference value that, together with other
4 More than 5000V economic, technical and strategic factors, helps to decide
3 Between 2000V and 5000V which level of maintenance should be applied to the
2 Between 500V and 2000V
1 Between 100V and 500V
equipment and how the maintenance resources must be
0 Less than 100V distributed.

References
Table 13
Criterion c12: mean time between failures (MTBF) [1] Souris JP. El mantenimiento, fuente de beneficios. Diaz de Santos,
4 The equipment fails very frequently: MTBF #1 month SA: Edit; 1992.
3 The equipment fails frequently: 1 month!MTBF #3 months [2] You-Tern T, Kuo-Shong W, Lin-Chang T. A study of avalaibility
2 The equipment fails with some frequency: 3 months!MTBF centered preventive maintenance for multi-component systems.
#1 year Reliability Eng Syst Saf 2004;84(3):261–70.
1 The equipment fails occasionally: 1 year!MTBF #2 years [3] Whitaker LR, Samaniego FJ. Estimating the reliability of systems
0 The equipment fails rarely: MTBFO2 years subject to imperfect maintenance. J Am Stat Assoc 1989;84/405:
301–9.
 F.C. Gómez de León Hijes, J.J.R. Cartagena / Reliability Engineering and System Safety 91 (2006) 444–451 451

[4] Martorell S, Munoz A, Serradell V. Age-dependent models for [17] Saaty T, Alexander J. Analytical planning. The organization of
evaluating risks and costs of surveillance and maintenance of systems. systems. Oxford: Pergamos Press; 1985.
IEEE Trans Reliab 1996;45/3:433–41. [18] Saaty TL. The Analytical hierarchy process: planning, priority setting,
[5] Martorell S, Sánchez A, Serradell V. Age-dependent reliability model resource allocation. Pitsburgh: RWS Publications; 1990.
considering effects of maintenance and working conditions. [19] Peniwati K. The analytical hierarchy process: its basics and
Reliability Eng Syst Saf 1999;64(1):19–31. advancements. Jakarta: INSAHP; 2000.
[6] Chelbi A, Ait-Kadi D. An optimal inspection strategy for randomly [20] Wirth N. AlgorithmsCdata structuresZprograms. New Jersey:
failing equipment. Reliability Eng Syst Saf 1999;63(2):127–31. Prentice Hall Inc. Englewood Cliffs; 1976.
[7] Jayabalan V, Chaudhuri D. Cost optimization of maintenance [21] Emblemsvag J, Tonning L. Decision support in selecting maintenance
scheduling for a system with assured reliability. IEEE Trans Reliab organization. J Qual Maintenance Eng 2003;9(1):11–24.
1992;41/1:21–5.
[8] Bahrami-G K, Price J. Mantenimiento basado en el riesgo. Gestión de
activos industriales, año III, número 9, enero/2000. Alción. Madrid:
Edit; 2000. Félix C. Gómez de León Hijes obtained his grade of Industrial Technical
[9] Casal J, Montiel H, Planas E, Vı́lchez J. Análisis del riesgo en Engineer, specialty in Electric Machines, at the Polytechnic University of
instalaciones industriales. Barcelona: Ediciones UPC; 1999. Barcelona in 1979; He is M.S. Industrial Engineer, specialty in Electronic
[10] Hokstad P, Flotten P, Holmstrom S, McKenna F, Onshus T. A and Automatic, at the National University of Education at Distance of
reliability model for optimisation of test schemes for fire and gas Madrid, since 1991, and PH. D. Industrial-Mechanic Engineer, since 1997,
detectors. Reliability Eng Syst Saf 1995;47:15–25. at the University of Murcia. At the moment he is Professor of Mechanical
[11] Cepin M. Optimization of safety equipment outages improves safety. Engineering in the University of Murcia. His research interest is centered in
Reliability Eng Syst Saf 2002;77:71–80. the area of Industrial Maintenance, in which has a good number of
[12] Stamatis DH. Failure mode and effect analysis: fmea from theory to publications and participations in national and international congresses. He
execution. Portland, Or: Publ. American Society for Quality; 2003. is author of the books: Technology of the Industrial Maintenance (1998)
[13] Duffuaa S, Raouf A, Campbell JD. Planning and control of and Basic Manual of Corrosion for Engineers (2004).
maintenance systems: modeling and analysis. New York: Wiley;
2000.
[14] Monchy F. Teorı́a y práctica del mantenimiento industrial. Masson,
Barcelona: Edit; 1990. José Javier Ruiz Cartagena obtained his M.S. Chemical Engineering, at
[15] Gómez de León FC. Tecnologı́a del mantenimiento industrial.: Edit. the University of Murcia, in 2002. He has worked in the industry as
Servicio de publicaciones Universidad de Murcia; 1998. responsible of Maintenance Plans. At the moment, his technical interest is
[16] Romero C. Análisis de las decisiones multicriterio. ISDEFE. Madrid: centered in Hazard Evaluation and Emergency Plans in Chemical
Edit; 1996. Industries.