An Evaluation of Calendar Based Preventive Maintenance System Versus Usage Based Preventive Maintenance System

View metadata, citation and similar papers at core.ac.
uk brought to you by CORE

provided by Lehigh University: Lehigh Preserve
Lehigh University
Lehigh Preserve
Theses and Dissertations
1-1-1977
An evaluation of calendar based preventive

maintenance system versus usage based preventive
maintenance system.
James A. Cichelli
Follow this and additional works at: http://preserve.lehigh.edu/etd

Part of the Industrial Engineering Commons
Recommended Citation
Cichelli, James A., "An evaluation of calendar based preventive maintenance system versus usage based preventive maintenance
system." (1977). Theses and Dissertations. Paper 2098.
This Thesis is brought to you for free and open access by Lehigh Preserve. It has been accepted for inclusion in Theses and Dissertations by an
authorized administrator of Lehigh Preserve. For more information, please contact preserve@lehigh.edu.
AN EVALUATION OF CALENDAR
BASED PREVENTIVE MAINTENANCE
SYSTEM VERSUS USAGE BASED
PREVENTIVE MAINTENANCE SYSTEM
by
James A. Cichelli
A Thesis
Presented to the Graduate Faculty
of Lehigh University
in Candidacy for the Degree of
Master of Science
in Industrial Engineering
Lehigh University
1977
ProQuest Number: EP76371
All rights reserved
INFORMATION TO ALL USERS

The quality of this reproduction is dependent upon the quality of the copy submitted.
In the unlikely event that the author did not send a complete manuscript
and there are missing pages, these will be noted. Also, if material had to be removed,
a note will indicate the deletion.
uest
ProQuest EP76371
Published by ProQuest LLC (2015). Copyright of the Dissertation is held by the Author.
All rights reserved.

This work is protected against unauthorized copying under Title 17, United States Code
Microform Edition © ProQuest LLC.
ProQuest LLC.
789 East Eisenhower Parkway
P.O. Box 1346
Ann Arbor, Ml 48106-1346
CERTIFICATE OF APPROVAL
This thesis is accepted and approved in partial fulfillment of
the requirements for the degree of Master of Science.
7 Datfe
Chairman of department
ii
ACKNOWLEDGEMENTS
The author wishes to express his appreciation to his thesis
advisor, Professor W. J. Richardson for his guidance in the study
presented here, and to Professor M. Groover for serving as a second-
ary advisor.
I also wish to express sincere appreciation to Ms. M. B. Hovik
of the Bethlehem Steel Corporation Research Department for her
assistance and suggestions during the operation of the simulation
model employed in this study.
Most of all I wish to express my deepest thanks to my wife,
Gail, for her patience and support through the many hours work re-
quired for this thesis.
iii
TABLE OF CONTENTS
Page
Acknowledgements iii
Abstract 1
I Background 3
II Problem Statement 13
III Proposed System Description 26
IV System Simulation Model 58
V Results and Conclusions 66
VI Suggestions For Further Investigation 71
Appendix I Simulation Model Major Subroutine 72

Flow Charts
Bibliography 97
Vita 99
xv
LIST OF FIGURES
Figure Page
Number Number
Chart of Management Objectives Relating to 5-7

Non Integrated and Fully Integrated Maintenance
Systems
Block Diagram of Shop Control Scheme With 19

Partially Integrated Feedback Using Direct
Numerical Control
Block Diagram of Plant Control Scheme With Fully 20

Integrated Feedback Using A Distribution Process
Control Computer and Segregated Process Control
Computers
Chart of Management Objectives Relating to 23 - 25

Partially Integrated and Fully Integrated Maint-
enance Systems
Block Diagram of Plant Control Schemes With Fully 30

Integrated Feedback Using Direct Interface Between
the Host Data Processing Computer and Segregated
Process Control Computers
6 Maintenance Weekly PM Schedule Format 42
7 P.M. System Exception Report No. 1 Format 43
8 P.M. System Exception Report No. 2 Format 44
9 P.M. System Work Feedback Card Format 50
10 P.M. System Additional Work Assignment Card 55
11 Simulation Model Data Comparison 67
v
ABSTRACT
Preventive Maintenance Systems as a rule do not lend themselves
to easy evaluation as to effectiveness. No one PM (Preventive Maint-
enance) scheme has been considered the best or least desirable. The
most common and generally accepted system of PM is the periodic
interval system or calendar based. However, this system has a num-
ber of weaknesses.
To overcome some of these shortcomings, the usages based system
has certain advantages. Implementation of such a system involves
more precise data collection requirements and direct interaction with
the process and equipment. It is necessary to make a judgement
whether the advantages of such a system outweigh the possible
difficulties of implementation. Only a complete evaluation of the
operation of the two schemes could develop such an answer. A
practical evaluation of the two schemes from a real-world install-
ation would be time consuming and difficult to complete. A simula-
tion model developed around a real-world situation to perform this
function would provide the data to adequately evaluate both systems.
The simulation model developed in this thesis provided the data to
make valid judgement of the merits of one system over the other.
The model output data showed significant differences in man-
power utilization and equipment performance in favor of the usage
based system. It was found by keeping the manhour parameter constant

and operating the model in both modes of PM, the data showed up to
.6% gain in productivity when the equipment was monitored for the
usage based system. For a large operation this represents a signi-
ficant increase and would justify the establishment of such a pro-
gram.
CHAPTER I
BACKGROUND
Maintenance Management systems from the simplest to the most
complex forms have been applied in all types of industry and service
institutions. No one system has been considered the panacea for all
to follow. The systems that rely on completely manual record keep-
ing, planning, scheduling and accounting are acceptable if the
operation is small. The more advanced systems designed for larger
operations employ computers that include maintenance in the "Manage-
ment Information System". These systems handle not only the record
keeping and accounting but provide data that also assists in deci-
sion making. These advanced systems of Maintenance Management,
however, do not always fully extend to the area of planning and
scheduling.
In a period of escalating material, services, and labor costs,
it is obvious that attention to maintenance costing and cost improve-
ments would appeal to plant managers. The objective is to apply
effective management to gain control of the maintenance situation.
Any number of possible approaches can be employed. One approach
applies the techniques of the management process to develop a
maintenance system that fits the needs of each specific operation.
The system must be designed and operated within an established budget
to verify the overall effectiveness on the operation under control.

The systems discussed throughout this thesis exist within
categories defined as non-integrated, partially integrated, and
fully integrated. The non-integrated system operates with no com-
puter assistance. Under this system record keeping, work initiative
and reporting is done manually. The chart illustrated in Figure 1
has been prepared to make a comparison of a non-integrated system
with a proposed fully integrated system in every aspect of the manage-
ment process. The partially integrated system uses a data processing
computer to supply record keeping, back up maintenance information,
periodic work schedules and performance reports. The fully integrated
system utilizes both a data processing computer system and one or
more process control computer systems. The process control system
is data linked to the data processing system for feedback purposes.
In this system maintenance of equipment controlled by the process
control computer is enhanced by operational data and performance
information supplied to the data processing computer. The fully
integrated system contains all the benefits of the partially inte-
grated system plus the ability to perform more timely maintenance
services to provide more efficient control of manpower and to
collect more accurate equipment operation information. The chart
in Figure 1 points out the differences between maintenance systems
operated at both ends of the spectrum of sophistication. In actual
practice, each new maintenance system must be examined within its
operating environment to determine what level of sophistication is
required. All maintenance systems are not operationally or economi-
cally feasible to operate in a fully integrated mode.

4
CO
0) (
3 O
1 • cr 4J •tH
0) U-l CO • •H u 3
C rl •H 0) 0) B 3 3 0) O 3
O >> tH 60 u 0) J3 O CX iH CO
•H rH 3 CO a 3 rH CJ iH CD 4-1
4J «H cr a> co cfl J3 0) 4J CO 3
U CO 0) •H X C O 4J 3 *4-l •H
3
XI •
XI rl 4J
•H
U 01
4-1
u
ex 0)
X
iH
o BO 4-)
Q O CO XI ^ > a) • 3 CJ • VI • a> «4-l 8
W ri U CD u •H iH B •H u C 4-1 4-1 u CO 3 01
H &.-H rH o 4-13 0) CO •H cd 3 co 0) cO •H s
& B >
2O cd 3
►> D-XJ
H at u tv
CJ XI 4-»
CO CD CO
,3 >> rH
3
O
l-i
0) B
4J ex
x> o
a •
£>
CO (0
CX
rH ■H
3 •
w •H U .3 rH >-, CJ CD • r-l J3 3 i-l A! 3 CO4J O
H CO U • TH i-H CO 4-1 CO U •rl 3 H X)CO i-l 01 C
55
M
xi
•H
cfl co CO
-•
QJ >
N
QJ •H
4J 4-1 4-1
xi 3
•H 0)
sa) >
o
*J 0) CJ 8 4-1 rH
o)
a) a) N a) •H CO CO rH
CO
►4 •H CO QJ •H o • 4J 4-1 4-1 u 3 > o >
•H H CO •H co
^ B o> ,-H >
O CO CO CO CO 3 Vi • •rl O 3 B o 4J X! 4-> B
p ■H M
p CO •H ex 3
&
CX •H "i-> 4-1 3
CD
3 co B 4J 8 T-)
2j O O O 1-1 0) "O
o
rl
01
CO
rl
01
CO
ftp, 01
0) 4-1
> CO 5 CX
CXrH 8 01 •H a O 4J
o o iH B S « •H 0) 19 iH ,3 CX
3 iH
E rH C_> CO PM <J s o o S HJ o W M H pq B NS E <4-l CO O
CO m vo CM co -<r m vo r^ 00
3 XJ
iH cu
cfl cfl
CO r^ U
01 0) 00
3 Pi 01
1 cr 4J
0) • •H CO 3
3 i-l
•H
0)
u
e 3 0) M
O >, 0) 43 > I
•H rH 3 3 r-i o •rl >,
4-» -H cr cO ,£> 0) * 4J r-i
CJ CO QJ 3 o 4-1 CO CJ rH
3 XI rl QJ rl XI 01 3
XI • 4-1 CX 01 u
O CD XI Ai 3 CJ o
rl 1-1 01 rl •H o 3 4J CJ O XI
Q CX -H rH O cd iH cfl 3 QJ 3 •
W cO 3 S B 3 3 QJ rl 4J Cfl CD
H ►> ftfl O QJ 6 3 S
2O rH
Ti
CO
01 01
U X
O •tH
5^
<-\
r-i
r-i
Cfl
rl
X
o
4J
3 •H 4-1
TH 3 CO
CX 01 OJ
B
QJ
XJ
0)
4J
QJ
4J
CD
W xi cO co CO cfl rl cO cr rl 60 CO >,
H •H 0) XI 0) 0) B 0) 3 Cfl I-l OT
a 0) > 01 •rl > 4J CJ 3 00
M N 4J 4-1 4J
• o cfl QJ 0) CJ S 0) 0)
1 •H CD QJ tH CJ 3> >
• •H O
cfl SS 4-1 o
S3 B QJ rH > QJ CD u o 3 3
o •rl CO CX iH •r-i 4-1 01 4-1 B rl rl CX cfl
55 3 CO S 4J O 3 & CO iH 3. fr 01 O I 8
•rj O O CJ r! QJ o o rH B S 0) 8 QJ
S rH O CO P* B (J CJ W M rH *: 4-1 O
rl 55
cfl
-3 O
CN co CM co -* m CJ 4-1
H
QJ (D
O
EQJ 00
QJ
iH H 4J
co 4-1 4J cfl
rl CJ
> O ecl
2P n
3 4J
•H -3 H q co
4-1 c/> v»^ rJ v_^
CJ
QJ
42
O
I
u
cu
a>
c
to •H
•3 60
«
CU
CM oo c
a>
o .3 Xi
w co CO cu
a
o
w
00
CM
•H
00
C
CJ
a) u
xi 3
o v
0J
u
•H
a) oo
•a *a
CO
4-»
CM
•H
H Xi at 3 I-I J3
00 W o o* 0) o CO
M (U •H 0) 4-1 u
I ►> a CM oo cu a)
o •H H w 4J
5 4J
s O Jrf
(U M
&
S op g
a o O (3
fe CJ -H
cu
3
C
•H
4-1
a
o
o
cu v-x
rH
CO T3 rH
4-i d
tJ >. CM "3 CO Cz3
Q ct) ^ J3 •H cu ,C 2
W a) rH a .3 4-) CU jl5
H J3 a 4-1 CO CO o CO o
cfl o CU CO 4-1 •H M
2
O 4-1 4-1
*i
CO
p
*
o
cu 4J
> Pn
w <4-< CM 4J M rH M
H •H •H >> M o 3 CU
55 J3 X .3 o M CM CJ &
M co CO CJ o •H o
1 4-1 CU cu CM a
55 cu >* cu s CU s CM 3
O CO o o •H
55 c§ CO 55 g Q §
X 3
P-l cu •H
u -3 4-1
H CM * •H CU "3 3 CU
55 O U > U 3 cu 00
*H
CU O CU •H CO B 3
r-l CU ? *3 3 ^ a. CO
W 3 rH s—' crs •H Xi
>-J *3 3 CM CU PL, 3 CJ
W CU -3 O 3 ^N u cr cu
J3 CU O -3 CO CU CU rH
to CJ to Si CO •H M M Xi rH 3
CO CO ja o r-l 4J O cu o CM *3 *3
w O CO •H c0 CJ & T-J o 3 CU
o o
o
a;
"3 i-)
CU
O u
cu
o
CO
4J
CU
02
O
ex a
3 -H
>.
4J
CO Xi
Xi CJ
CO
PM 3 •H 3 •3 CM rH CO "M •H o
CO CM (0 3 c0 •• S *H c rH 4-1 60
H
••
> •H > 4-1 •H O 00 CJ 5 •H 3
a •3 O -3 •H •H 3 CU CU o X) >s-H
§
w
00
s
•H
CO CU
4-1
Ou
CO
CO
4-1
CJ
•H
rH
&. M
3 O
M 4-1
•H
rH
3
CJ
3
c0
(X
co 3
CO
CO
rH
•H
4J
•H
rH
4-1
CO
U
o 3 CO CO P. u CO •3 > u (U CO •H CU
< 3 CU CM CU X CO -H CU •a o U > XI Cu
CO PQ O FQ w « x: Xi <j CM Xi < < o
§j rH CJ
CO
* a
a o) CO
rH 0 4-1
O Ou 1 CO
O iH n T3
*-> 3 o
cr «4H T3
•> a) d fs C M
(0 TH 4J CO ft
. 4J 60 • •H O
a • CO rl 3 CD CO > CO U
w a> a) co TH as iH •H 13 ft
H rH 00 CX 4J CD Xi 4-1 U
• ft •
• 3 (3 « 43 CO a
3ow J»5 T3 CO rl
M a) J3 fH
c U (0
M
CD Cfl
4-1
CO
0)
CJ O
3 CJ
•O CD
a)
u
a
4->
CO u
CU o
H
2
> •u
O.U
ft. O CO > U (0
O rl
CJ ft
55 a> o ai *a •H ftS a 3 CD
M cu 4-1 u > CD Pu CD T3 H
1 (V > CD CD CJ M 4J O
r~" at a) •U H C *-> CD O •« a M C
■J rl iH U 41 CO M <4-4 CO •H ft O
0) > co
gfa > 0)
fl) T) t3 "O
« h a a
O Pi pi o co »
B
CD o >>
C
iH
co B
§
•H
4-1
ft
4-1 •H rH CD 4-1 CD
CO 4-1 CO CO CO G
• • • >•. co a CO ►> X
n <f u-» CO B co X> CO CU
T3
4-1 CU
« a 1 3
CO CU •H
rH S «4H
O ft I4H
O tH •H c
4J 3 •a
•
o
a*
•> cU CO CO
CO •H AS
• 4-1 60 CO
• CO rl C • C co T3
CD CU CO -H CO O 4J CU
Q rH 00 ft 4J 4-1 •H 4-1
W • 3 C CO e 4-> o CO
H ^ •o CO rl M CU (0 4J 1-4
g
3O ri CU
P X
£ CJ
£iH
CJ cO
ft o
111
ft 3
o O 4J
E CU •H CD CO
M co CU 4-> U C rH rl 4J
| u •H u CU CO CO u
55 CU > CO "O T3 a o o
o
55 5 CUcu ri a
Pi Pi O cO
go•H 4J
4-1
ft
CU
HH CO 4-> rl
•H rH
• • • • o K 3 o
iH CN on «* 55 W CJ 55
CO
a)
•H CO
4-1 c 4-J 60
•H CU rl CU
4-»
a
4J O •H
•H • ft cu 4J
•
cu cu u
iH c rH ft
rl o
CO c o 3 S ft
a o •H T3 c o cu
o iH 4J cu o o u
cx
co
4J • • CU J= •H c
cO co a rH CJ 4J •H G
cu rl B O ft • CO ft o
u 3 CU iH B co cu O iH
• • T3 rH 4-1 o ►> x: CJ 4J • 4J
>. CU <D X» 3 O CO u X cu ft •
rl e 4-1 U3 O •-{ T-{ •H cu CO rH CU CO
o iH CO O U O JS AS 4-1 4J 4J 4J 4-1 4J 4-1 •• u CO 4J J3 •o
•H rl CJ rl U r4 U U M ^ rH cO rl ft U t> cu
4J CU cO O o O o o O O O Q- CU cu CO fl CJ
3 ft 43 ft ft ft ft ft ft ft U B a u CO o
u 3 T3 CD CU CU cu a)
0) CU 4J O CU X s 4J rl
cu to CU Pi Pi Pi P4 Pi Pi Pi C! o o w P-, CO ft
CU O
Pn O
A wide variety of maintenance costing and accounting systems
befitting different operations are currently available. Therefore,
finding a universally accepted cost system is extremely difficult.
If every hour of maintenance time is assigned by a work order num-
ber then cost allocation is infinitely easier than a system where
maintenance personnel time is charged to a general shop overhead
account. In the latter instance the daily activity and producti-
vity of the maintenance man escapes easy and accurate measurement.
Justification of the installation of new equipment for a com-
puterized "Maintenance Management" system based on benefits of man-
power and parts inventory control could fall short on typical
payback schedules. However, if the major capital items such as the
data processing computer and data collection facilities are already
in service for other accounting and management functions the justi-
fication appears more attractive.
The more sophisticated the requirements for the maintenance
system are, the more stringent the data collection requirements
become. Conventional computerized maintenance systems at most employ
data terminals or card input for updating the host computer. The
timeliness of the input data affects the accuracy of the entire sys-
tem especially if the system is planned to react immediately to
changes. Any significant changes of an operating nature dictate
schedule changes for preventive maintenance, major rebuild activi-
ties or daily shift assignments.

In order to meet the timeliness and accuracy of the data
requirements, a second computer, collecting data on a real time
basis at the shop level, linked to the host computer is required to
handle the problem. If the second computer actively participates in
control of the shop operations it should already contain the data
needed by the host computer to perform the maintenance management
functions. This second computer then becomes a process control
computer.
The process control function with computers has been a rapidly
expanding application of the computer. Early applications involved
very dedicated control of a small portion of a given industrial con-
trol system. Gradually as technique and technology progressed, the
process control computer encompassed an entire control system by
monitoring and controlling many interrelated elaborate devices.
These areas of control include Steelmaking, Power Generation,
Mass Transit, Chemical Production, Warehousing and numerous other
applications. The parameter common to all these applications is
"real time". Each of these systems must deal in "real time" to
successfully achieve the desired output performance.
The great trend toward management information systems readily
lends itself to the process control computer. The process control
computer, as it is being applied in the case under study, contains
information that if properly retained and formatted can be used as
input to an information system without the intermediate step of

manual handling. The most significant advantage Is the timeliness
of the Information. Under current applications In which manufac-
turers of finished products are concerned, very relevant quality
control Information can be transmitted piece by piece or batch by
batch depending on the user requirements.
Embedded subtly in the process information are the data whereby
the maintenance of the process equipment and control equipment can
be monitored and directed. Once this step is achieved the evalua-
tion of maintenance personnel can be ascertained. Needless to say
the process control computer cannot do this on its own. This type
of computer has the prime responsibility to the process and it is
not designed to handle major data processing type functions.
It is proposed that the host data computer linked to the process
computer handles the rough task of data compilation and forms
generation. The entire concept is based on eliminating the lost time
in formulating production feedback to a given information system
through shop terminals. In many processes, the process control sys-
tem already contains much of the data that is also fed back to the
host computer through a separate shop terminal. However, in many
cases that information is lost as the next series of orders is pro-
cessed. Presently, if an information system is tied in to the
process control computer, it does not consider the aspects of the
productivity of the maintenance personnel servicing the system
control equipment.
10
With the continuous effort toward creating real time informa-
tion systems, the capability of linking with process control compu-
ters with data processing computers is rapidly being developed. The
area of study proposed is to investigate systems whereby the process
information is used to schedule, control and manipulate preventive
maintenance procedures and to measure the productivity of those
performing the overall maintenance functions and to compare such
systems with those not employing such information.
Maintenance functions have historically been relegated into
after-the-fact systems. A good universal system has not been de-
fined. The one fact agreed to is that maintenance costs are real.
The maintenance dollar is a very significant cost factor of a produc-
tion facility. Yet, the effectiveness of the maintenance dollar has
been elusive and rarely definable. It is, however, a major contri-
bution to profitability. Some maintenance expense is essential.
This investigation provides a means to determine if scheduling prev-
entive maintenance on an equipment usage basis is a more effective
usage of the preventive maintenance dollar than the calendar based
preventive maintenance scheduling system.
The day to day output of maintenance personnel is difficult to
equate in many kinds of production facilities. However, when a
sophisticated order entry process control and production reporting
system is applied there could be also a system to schedule preven-
tive maintenance and to measure productivity by using the data that
11
are already available. The proposed system gives the added feature
of productivity measurement without extra data collection personnel.
The preventive maintenance function is to be applied to mech-
anical, electrical and electronic equipment including the computer
or computers themselves.
12
CHAPTER II
PROBLEM STATEMENT
Historically, maintenance in the industrial atmosphere has been
accepted as a necessary evil. When times are good, maintenance
costs are eventually ignored as are the activities of the craft
personnel. Machines and controls are serviced on a breakdown basis
only. As operating costs creep upward and orders fluctuate radically,
maintenance expenditures with respect to their influence on profits,
become more apparent. It is very obvious that a better handle on
the maintenance function is required.
The initial steps to improvement included streamlining record
keeping systems and developing preventive maintenance schedules
which were used to inspect and lubricate mechanical equipment, and
to inspect, tune and clean electrical or electronic equipment.
These schedules, repeated on a periodic basis, resulted in a definite
improvement in the service of the equipment. As more and more shops
were included in this maintenance system, paperwork increased tre-
mendously. Both the record keeping system and updating procedures
required dedicated personnel. The next logical step was to stream-
line the records and scheduling system by storing the data in compu-
ters; permitting the schedules to be issued by the computer and
records kept by the computer. The records input data into the
computer would be through cards or other media. This input data
13
contained the completion data about the schedules the computer
issued for the previous inspection period.
So, wherein lies the problem? The fact is we must be able to
develop a sufficient knowledge of maintenance activity to assist us
in forming a well organized and meaningfully controlled maintenance
department. This organization must be developed so that its mem-
bers, maintenance personnel, are made aware of their effect on pro-
duction and profitability. To achieve this end, more than just a
periodic schedule for preventive maintenance is required. The func-
tion must be extended to include historical records of repair on
equipment, cost expenditures on a given machine, standards for maint-
enance job times, distribution scheduling and other maintenance
oriented details. The literature on such advanced systems is
sparse. More development in this area is required.
James K. Hildebrand in his book Maintenance Turns to the Compu-
ter, describes an example of a partially integrated system of
maintenance control using a data processing computer. Mr. Hilde-
brand' s maintenance system involves preventive maintenance and pre-
scheduled work, backlog work and call-in work. Mr. Hildebrand's pre-
ventive maintenance system calls for establishing inspection intervals
in time. These times are based on the risk of downtime on a given
machine. After a period of time, the historical repair record of a
' Cahner Books, Boston Massachusetts, 1972,
14
machine may be reviewed to determine if trends in repairs indicate
reducing or increasing the inspection interval. The computer desig-
nates a work order number to each job assigned. It then waits some
prescribed time then lists a "back-log" for any job work order number
it has not received a completion card as feedback input. Thus an
exception report. Breakdowns are classified as call-in work and
inputs to the computer are coded to be reflected in the equipment
historical record and cost control records.
Mr. Hildebrand's system, and other similar to it, may be
excellent in the environment in which they are applied. Those
applications include machine shops, foundries, job shops, assembly
plants and many other types of production shops. These types of
computerized maintenance systems nevertheless still lose sight of
two factors. The first area of omission is the day to day produc-
tivity of the man responsible for breakdown maintenance. This man
holds the assigned maintenance position. This position relates to
the men who are scheduled to work on operating shifts and repair all
breakdowns as they occur. We are presently able to determine the
output of men whose sole concern is to handle scheduled maintenance.
It is the performance of the assigned maintenance man that has been
elusive to adequate measurement. The second deficiency is in the
method of establishing the interval of inspection listed on preven-
tive maintenance schedules. Most preventive maintenance schemes use
established time intervals in weeks, months, or days as triggers for
15
work schedules. These schedules are totally Ignorant of whether the
machine or whatever being Inspected, tuned or lubricated has even
been in service during any, most, or all of the time since the last
inspection. These schemes keep maintenance personnel busy, but not
necessarily productive. There seems little sense spending time in-
specting, tuning, or lubricating or tightening some piece of equipment
that may have only been in service part of the time after the last
work order appeared on the preventive maintenance schedule.
For example, with today's rising costs in petroleum products,
there is nothing gained from regreasing some piece of equipment or
changing virtually unused hydraulic oil and filters if the machine
only had token service. This type of situation can easily occur in
the larger organization where the maintenance force is so large that
the man doing the job may r\gt always work in the same area or section
where the machine is located. Unless proper information is made
available, he does not know that this machine has not had orders
and was out of service three-fourths of the time since work was
previously performed on it.
It should be noted that the reverse situation can also occur.
This happens when the supervisor trend review shows inactivity on
some machine and the preventive maintenance schedule is then length-
ened. Then, if orders are received and this same machine goes back
into high production, problems can arise because the preventive
maintenance cycle is too long. This situation can develop if
16
nothing is changed on the preventive maintenance schedule until the
next trend review period. The standard scheme cannot adjust to these
situations until the scheduled supervisor trend review makes the
change manually. This is because there is no present mechanism in
the system to record in the computer on a real time basis the actual
production hours of a given piece of equipment.
We are looking for the mechanism that can track the production
status of equipment on a real time basis and at the same time
collect data required to measure the productivity of maintenance
service personnel. The mechanism to achieve the desired results
resides in a process control computer system. However, the process
control computer system cannot act alone. It must be supported by a
large data processing computer. The process control computer
collects the data in a "matter-of-fact" manner because in most cases
some form of the data are used to run the process. The problem is
to be aware that the data are there and to format it properly for
transmission to the data processing computer along with the normally
scheduled data transmissions.
Data collection has always been a difficult problem for over-
seeing the day to day activity of maintenance personnel. This pro-
posed system not only performs the tedious data collection activity,
but also eliminates the step of manually preparing the data collected
for input to some computer system or manual incentive system.
The types of manufacturing or production organizations that lend
17
themselves to this approach are the automated machine shop operating
under direct numerical control as depicted in Figure 2 and manufac-
turing processes that are process computer controlled, such as
metals, paper, glass, chemical or other production complex as shown
in Figure 3. The production complex would include raw material
handling, manufacturing, finishing and secondary processing (if
required), product storage and product shipping.
The amount of investment in equipment in such a complex warrants
efforts to establish a meaningful preventive maintenance program.
Nor must one overlook the additional problems of effective staffing
of maintenance personnel and effective scheduling of maintenance
activities. The problems of staffing and scheduling go hand in
hand with productivity. For example, if a method of effective
measurement were established and work scheduled around the clock
using assigned shift personnel, equipment that would normally be
serviced on a day turn could be completed on off shifts.
The term "assigned maintenance man" previously mentioned is
further defined as groups of service people assigned to maintain a
production unit. The primary function of these people is to repair
electrical, electronic, mechanical, hydraulic and all other service
oriented breakdowns as needed on operating turns. Their work time
is charged against the production shop and normally does not reflect
the work assignment for that shift. The practice in the past for
this type of maintenance operation has been to schedule minimal
18
Work
Schedule w
DATA PROCESSING Production DIRECT NUMERICAL

SYSTEM Data CONTROL PROCESSOR
t i
DNC Commands
Delay And Command i

i
Inputs ' Requests r
PRODUCTION MACHINE PRODUCTION MACHINE

NO. 1 NO. 2
PRODUCTION MACHINE PRODUCTION MACHINE

NO. 3 NO. 4
l 1
1
FIGURE 2
Block Diagram of Shop Control

Scheme With Partially Integrated
Feedback Using Direct Numerical
Control
19
Orfler
Information
COMMUNI-
DATA PROCESSING
I CATION
PROCESSOR
SYSTEM Production
Data
DISTRIBUTION
Process COMPUTER
Signals (May Contain Control
Function)
TRACKING AND RAW i

MATERIAL CONTROL
Process
Data
MANUFACTURING PROCESS CONTROL

CONTROL COMPUTER NO. 1
FINISHING FACILITY PROCESS CONTROL

CONTROL COMPUTER NO. 2
INVENTORY AND PROCESS CONTROL

SHIPPING CONTROL COMPUTER NO. 3
Block Diagram of Plant

Control Scheme With Fully-
Integrated Feedback Using
A Distribution Process PROCESS CONTROL
Control Computer And Segre- COMPUTER NO. "N"
gated Process Control i
i
Computers
- -J
FIGURE 3
20
personnel on non-day shifts. These personnel are primarily concerned
with performing breakdown maintenance and small jobs that can be com-
pleted or left undone if a breakdown should take up most of the
shift. On the day shift, more personnel are scheduled to work.
These people normally handle large jobs and preventive maintenance.
More supervision is also available on day shift to monitor the acti-
vities of those performing assigned jobs. We must be able to judge
whether we are making effective use of the non-day shifts. There is
a good possibility that some equipment used on day shifts may be
available on the other shifts for preventive maintenance work or a
major rebuild that if done on a day shift, could slow down the pro-
duction cycle.
This is leading to evidence of the need for effective use of
maintenance personnel with the production requirements of the manuf-
acturing complex. The key elements for success are effective day to
day personnel activity measurement and day to day update on the usage
of the production equipment.
The proposed solution to be investigated that will satisfy
these needs also can incorporate record keeping and cost control data
on equipment expenditure. These features were included with the sys-
tem described by J. K. Hildebrand. It is not the purpose in this
thesis to include these features, since we know they can be done.
The ojbective here is to investigate the possibility of moving
industry one step closer towards effective maintenance by adding the
21
dimension of real time to the reporting function at minimal costs.
More effective maintenance can be achieved by using the personnel
more productively and performing preventive maintenance of equipment
by actual production hours. Using actual operating hours is a new
approach to preventive maintenance.
The area of investigation then is to take a system that in-
cludes one or more process control computers data linked to a host
data processing computer and to use the process-generated information
in conjunction with the host computer to operate a preventive maint-
enance system and measure the productivity of the personnel involved.
This new fully integrated system can be compared with a partially
integrated maintenance system. Various operating measures will be
used as criteria for evaluation.
The chart in Figure 4 lists the attributes of the two systems
under investigation. Specific concentration will be placed on the
areas that appear to have the most significant overall affect on the
parameters considered for the evaluation in the chart.
22
cu
3 1
e • o4
• O CO CO • CO •H •H
(3 co to CU CO CU B (3 o 3 C
O 4J TJ • 00 CU a CU Xi 1-1 cr o
•H *-> C a) ^-N to 0 «H a iH CJ
•
14-1 cu •H
4J m CD CO CO CU •H 4J
U •H B CO M -H 43 -rl a o 4J CO u rH 3 •
3 43 0J 43 3 •»-• a > CU ri TJ CU rH 43 l-i
TJ (0 H^ O -H •H 4J CX CU l-l ex CO •H a)
O •H _ 43 > CU 4J a CJ • O CO l-l >
M a) 3 6 "H iH O •H CJ e 4J CJ 4J 4J o
cx rH cr 3 t>o 4-i 3 CO •H CO C CU CU 3 CO cx
4J
3
TJ
CU
M
4J J3
co i-l
CJ TJ
cd CU u
§ a c
o CU B
CU l-l 4-1 O
CO Xi
•H
TJ
c
o
<4-l
•H
a>
43
>> 4J si cu
^ en co 4-i o M
iH
rH
u 4-t a
Xi c ■H
CU
4-1
l-l CO
3 .US
§
W 43 CJ • rl U "4H CO M CO CJ •H 3 CO o (3 l-l CU
H co 0) CO O JS CU iH o M cr l-l u O O N
3 CU
V > Pl| (X43
CU •rl O CO
00 &
a >
CU CU
4-1
§ cu 3
CJ
CO •H
4J
> •H
B
O N • 4-1 4J 4J cu •H CU o cd CU CU CJ CU co CU •H
w •H 0) a> •H U Url U 4J U c > > (0 TJ Q CJ (3
p •
H
23
rH
ac >M-H > cu cO cO a>
CO CXiH i-i M 3 3
rH 0 4J O CU 4J i-i
CO
I-I
CU
3
CO
CO
u
CU
&
• •H o o
4-t s l-l u cx > o 4J
CO ■H cx CX CU o <4-l C
•H c
CO
rH
•H
s
iH a) S CJ U CX CJ TJ cx a) 5 o rH s s CU u (3 CU CO O
s TJ o CO cu o id < o S CJ w M M U3 CU •H B 4J
S hJ
o
£
Pn CM CO -* m vo N m>j m vo
• H TJ
CU
00 4J
(3 cO
•r) M
4J 00
CO CU
CO rH *J
CU aj c
• cr
3 pei M
C3
• O CO •H CO ^
c CO CU B c CU r-l
o 4J TJ CJ CU 43 > rH
•rH U a to CO 43 CU • 4J fi<
CJ •H B co 13 O 4-1 CO CJ
3 43 CU X CU l-l TJ CU TJ
T) CO u ^ 4J CX CU l-l i-i C
O •H c CJ • O 43 cfl
M CU 3 S • •H CJ c 4J CJ O
Q cx r-i cr cu ^-v cd •H CO a cu TJ •
W 3 CU 4-1 CO g a a CU u w 4J CU CO
H 4J
l*H
TJ
CU
M CO
>> rt
i-l
iH
o CU S
l-l 4J (X CU
C 4J s
2
o
•H
43
43
CJ
4«S CO >
• M _ M
iH
CO
43 c
•H 4J
CJ •H 3 CO
o
CU
S
OJ
CO CU
rl 4J
00 CO
w CO CO CO O S CU l-l co cr u 00 CU ^
H CU & P-. 4J CU CU fs CU 3 CO 4J C/)
55 CU CU
• 4J
•H C3
4-1 4J CU -rl
> 4J CJ c a
H
>-<
N
•H CO CU
B >M-H
•H CJ 4J
> 0) CO T)
O
l-i •
>
o
CO
C
■H
CU CU
> co
o
CJ
1">. CU
G
a
►J •H cd (X iH i-) l-l CU a? 4J B l-l l-i cx H-l rH cO
.J C rH s 4-1 O CU X > CO •H cx CX CU O rH C3
<! •H CU o CJ M CX -H o O iH S B cu cO CU
M s TJ o cd PNOW •J O W M M W 4-1 -H 4J
H rl 4J C
OS CO l-l •H
<! • • • • • • • • • 43 CO
PH iH CM co <• r-4 cs CO -* m CJ PL, S
BCU cd
B *H
C oo
01 H CJ cu a)
•H H 4J
4-1 4-1 CO
l-l U 00 l-i
CJ O CO C3 4J
OJ 43 4J O CO
C/l N-'
43
O
23
1 u a o* x-\ n
to > TJ 0 •H CU CO co oo
►» 0 CO CO & M ►» ■xa (3
to 4J cu rH CO •H
TJ
r~
„ «W
Q •H
iH
14-1
0
. cu
a
O
u
TJ
r-»
rH U
•H 44 ,*i O
>w CU 43 CO 4J 00 (3 • \-^ 0) . U
u CO 44 CO a O >. fU i* O
CO •H «W •H a) O u CO » iH CO
Q 4-1 CU iH rH cu 4J i-l
W 4H CO 43 ,13 cu CO a (4-t CO cd U
H •H 44 4-1 CO 4* 0 a) •H iH •O O
«* J3 (0 O iH TJ 43 X! a 3
P2 co iH TJ rH CU > 3 O CO 0 TJ O
o i-l 3 CM 43 cu rH CO cu a
w cu CO O
•
TJ U 6 CU Tj 44 a
s
M
3
o
1
§
44
-0 44
cu
0
U
■H
a)
0
cu •
44 CO
3
TJ
a
0
1
cu
44
CO
CO
3
t->
cd
S •
>» • 0) CJ Pu 14-1 (3 cu TJ > • CO TJ 3 Tj
r^ u TJ >.iH 3 •H cu rH V) O *J Tj O CO 00 (3
►J a co CO > O 3 a 0 •H ti (3 CO CU ■H CO
s a) CU
43 S3
M cu CU cr
a. cu
a u
0 cu
5> (3
a) CU s> CU
J3
U
0 3
TJ CD
co 3
0
K £ cd Pn
Q» rH
CO 43 co CO u u O 6 £ CO ft, (0 <j -a
TJ
cu
3
44 00 c
•
•H a
CU rH •H
•0 rQ •H -o 3
CO 43 (3 o
cu O CO 0) CJ
43 4-1 rH a. CO
CO •H CU • •a
CO CO TJ 0 CO 00
w >•» CU CO CU *H •H
H cO 44 4J 4J CO
Tj •H u CO O CO
pi 0 H (3 cO
0 rH CO a 3 cu
w >>-• 44 U a) co TJ u
H CO 44 u 0 cu CU
53 CO iH 3 (0 (3 43 4J
h-l u rH 43 CO CO CO
M-l O > cO
5« •H (3 +J TJ o
►J 43 § O cd • 4-4 cu u
cu
CO 44 5N >, •H >-l
53
p-l cu
CO TJ
>. CU
A CU
O J3
XI
0
43
CO
o >
o
E-i 0
AS
CO
a
o
o
55 a
rH CO
CO 0 cu
u •H 00
CU CU •> >H •4-1 1 3
rH rH cO U ■H a cO cu
3 3 3 44 4J O •H 43 rH
CO TJ TJ rd cO CJ CU 3 CJ 3
w CU CU rH 3 TJ CU a. cr TJ
!=) 43 43 o< O rH M CO CU 44 CU
cr CJ CO O •H 01 ffi cu CO 43
M co 4a CO 4»S 44 44 4-1 3 a
O CO U TJ 0 0 0 1-) CO
S TJ i-> TJ O a co 3 a. M-i TJ
CJ CU CU > C P. A 3 >> cO 00
w O O a 0 CO 44 s 44 3
H 3 •H 3 M-4 <±4 •> rH a 3 PH •H O -H
cO M-l cO O 3 co cO CU rH 4J 4-1
H > iH > •H 00 CJ 00 TJ a TJ iH cd
53 TJ O TJ CO 3 «H a cu a) 3 43 >* P
60 cO CU cO rH CXIH 3 •H 0 u cfl CO 44 CU
s c O. 53 •W 3 > co ^-N r-i 3 iH rH « •H a
M •rH 44 CO 44 PL, cfl 4«S cfl 43 cfl 3 CO 3 CO iH 44 rH O
O 3 CO CO 44 0 M a 44 TJ > W £> cO 3 •H
<! 3 CU >4-l CU >4-( CU CO TJ CU cO CU TJ CU 0 > cu 43 3
cO P3 O PQ O Q PQ w a TJ 43 <tl U 1-1 <: a «* -H
rH CJ
§ co
24
1
•
to U CO CU
•o O >>43
M o cd
O o rH C CO M
O Ctl CU ct] >> O
 CJ CO <4-l
u 00
a • c • a U
o •H CJ O f>s ^-s CM CU
• • 4-1 rH 5 •H H C 4->
CD (0 3 O 4-> O o >. 3 •
a) cu ^ TJ **? O 4-1 •H 43 a co
rH u A! 3 C
• 3 00 CU B cu
4-1
a td 43 • a) -o CU CJ tJ O CO
o ^s 13 cd 43 u CO CU o > c CU CJo
w M CU 43 •o CO 4J H u a 3 4-) CU
H O 43 u cu CO 43 Cu iH >4-l cd 4J u
& o a) S •H u co 3
CO o <*-i CM rH o CU C Vi CU o a
o 0) 4-1 • 4J N iH CU c 43
w (U > ai 6 CU O •H CC) 4J CU 4J
H
25
(0 s
4-1 rH CU iH 4J 4-1 6 4J 3
M •H u 13 4-) 13 CJ •H a CU
60 >. (3
43 CU
0) > cd a CO C 00 cd 3 Tt 6 CO 0
CU a) « t^ CO C CU •H cfl o 4-1 4-1 CU
S PS Pi B3 CO ffl •H PS sd a u U
o
3
CU
00
(0
CU 4J 3
# m # • • • * • CU 3 ed
Pn rH CM ro <r ir> ^o r*» oo CM en ps O B
cu
1 CO 13
13 # cu • • cu
to r< co 43 43 cu B 3
13 O >. 00 H N CU 3
U CJ cd 3 •H 4-1 iH
O CJ rH O U co 4-1
CJ cd CU u • cd >s 3
CU 13 g w o
u oo • 4-1 00
43
c § o
s • fl • co •H 3 s s-x
o •H rj o >. TJ CO & CO PM •
• • 4J rH 5 iH U U B cu 13 •<!■
Q CO co O3 4J O O cu cu o >, CU
W CU CU ^<5 13
13 CJ 4J O 4J .* 4J HP 4J w
H r-4 00 o •MCU 3 Cl CU •H cd PS
• 3 e cd 43
cd • T3 CU )-l *T3 4-1 T3 u C3
J«J 13 cd 43 cu
CJ CO O > rl )-( rH CU CU O
O
W
M CU 43 *r> uCO 4-1 U C s CU O O3 4-1 (3 rH
O 43 CJ cu 43 CO O- iH PM 43 O Cd CU fe
H > CJ cu S •H 4-1 CU •H u 00
53 CO o <4-l PL, rH
o CU c cu o u UH CU
CU 4-1 • 4-1 N iH 4-1 "4H c 4J
CU St CU B CU O •H Cd cu 4-1 rH •H CU o
CO CU 4-1 rH cu rH 4-> 4J B 4J rH u cd T3 00 c
rJ )-l •H CJ *« 4-1 Tj CJ ■H C a. o 3
CU > cd c CO C 00 cd C -H B a. c CU CO CO
H
s CU CU cd !* cd 3 cu
PS PS 33 CO P3 •H PS g^ o cu cd
CJ etJ B
u
cd
4-1
u
o
4J
u
o
, . cu Cu
# # # # • • cu cu
rH CM m <r m vD r^ oo CN PS PS
VJ
o
a
c G o cu
CO co o o 4-1 u
W a iH •H ,*
C3 o 4-J 4-1 CO u cu
o* a, cd co 01 4J O CJ
M CO u co 3 rH u & c
cu 3 S o Cu o cd
§ 1-1 13 CU «H B CO Cu2
O
w >>
rH 4J
cu cu ,M 43 3
O
CJ
>N
cd
CU
u
CM Bo
H u B 4J CO O rH <-\ cu <4-l
o • •H cd cd U O s CU 3 4-1 u
H CO CO 4J 13 4-> CU CO CM 13 o cu CU
Z 3 •H CU •H rH cu
H O > -H ^ 4-1 4-1 4-1 4-1 4-1 4-1 4-1 4J CU
§
w
•H M 4-1 a u U U u u u CO u r-t cu B >.
4-1 CU iH cd o o o O O O B o o cu o cd
^5
<!
3
CJ
CU
3
rH
-H
43
T)
a.
cu
CU cu CU CU CU CU Cu
cu cu cu cu CU 4J CU
M
4-1
CJ
X d
CJ rH
CU
JZ
3
CU CO 43 CU OS PS PS PS PS PS •H PS a W -r) a
X CU o
S w fe C_)
25
CHAPTER III
PROPOSED SYSTEM DESCRIPTION
Keeping in mind the output requirements of the system as des-
cribed in the previous chapter, presented here is a system design
and description to achieve those requirements. Before delving into
the ramifications of the proposed system, a discussion of the design
philosophy is desirable.
Regardless of how they may be described in the trade literature
and books, maintenance organizations peculiar to the older metals
industry have little use for any preventive maintenance system that
takes too much clerical time. If any automated reporting or schedul-
ing system is to be successful, it must be skillfully designed to
contain as little paperwork as possible with the highest effective
information content. Oftentimes, a corporate edict is generated
requiring the installation of some production, maintenance, inventory
control, safety, or other program. Unfortunately the burden of the
program nearly always falls on the line supervisor, who is already
deluged with forms for time, safety, quality, inventory, grievances
and numerous other needed programs. Aside from filling out these
forms he is above all responsible for the activity of his sub-
ordinates, the production or maintenance employee. Therefore, a
primary object of any new system should be to accomplish the job by
avoiding additional work or by reducing existing paperwork. This can
26
be achieved by making the new system a tool that the overburdened
supervisor wants to use in his daily routine. Thus, the objective
is to design a system that is used by the supervisor and not a sys-
tem that uses his time inefficiently.
The output reports of many of these systems are fine, but the -
input requirements put the burden on the first-line supervisor which
may be the wrong level of supervision for this. The position must be
taken that output reports can be anything one may want, but to gen-
erate these reports let the data processing computer manipulate more
data than usual. The raw input from the shop could be in a less
organized form. The approach philosophy must be toward freeing the
line supervisor to pay more attention to production or maintenance.
Most of these line supervisors are promoted from the ranks of the
craft they are supervising. They are trained as craftsmen and not
skilled managers. It is therefore obvious that we must design the
systems as suitable tools for their use without jeopardizing the
time they must spend overseeing activities whereby they can best
assert this influence toward profitability. The system described in
the following paragraphs, attempts to provide such a tool without
interfering with his main functions as a line supervisor.
As indicated, both process control and data processing computers

(
are required for success of this type of application. The data proc-
essing computer does the non real time activity of record keeping,
report generation, schedule generation and data base maintenance.
27
The process control computer collects real time data related to
requirements of productivity measurement and preventive maintenance
scheduling. The productivity measurement depends on the quality and
effectiveness of the preventive maintenance system. Accordingly, the
preventive maintenance system description is what will be discussed
first.
The preventive maintenance system has as its basic and unique
ingredient the actual service production time of a given piece of
equipment. It is designed around groups of area subsystems. The
area subsystems of a manufacturing complex employed as a prototype
for this system are as follows:
1. Raw material handling.
2. Raw material preparation.
3. Product manufacturing equipment.
4. Product line No. 1 handling equipment.
5. Product line No. 1 finishing and inspection equipment.
6. Product line No. 1 storage and shipping equipment.
7. Product line No. 1 secondary process No. 1 equipment.
8. Product line No. 1 secondary process No. 2 equipment.
9. Product line No. 2 handling equipment.
10. Product line No. 2 finishing and storage preparation
line 1.
11. Product line No. 2 finishing and storage preparation
line 2.
28
12. Product line No. 2 storage and shipping equipment.
Figure 5 shows the schematic of the area subsystems and computer
interaction. The need for the breakdown of the complex into these
area subsystems is two-fold. First, the process control computer
control assignments have a minimum amount of overlap with this con-
figuration. Thus, the data collection assignments are easier.
Secondly, when operating various product mixes the equipment within
an area subsystem operates as a unit making production hours report-
ing easier. Using large area subsystems will on the whole reduce
the data processing update time for most products. However, some of
the products manufactured require special coding of the information
about some areas subsystem. This coding is to be done in such a way
that during the data processing update phase certain flags cause a
further breakdown, within these special area subsystems, to take place
internally to the data processing computer.
Let us discuss the sort of data base required to support this
system. Building the data base is the most tedious phase in the
installation of any maintenance system. If this system were applied
right from the installation of the complex all the information would
be readily available and building of the data base would be quickly
completed. On the other hand, as usual, initial budgetary priorities
defer an automated maintenance system until a much later date - if
at all. This usually means when a system is finally installed, build-
ing of a data base could be a burdensome task if the manual system
29
Host Data Process
Processing Raw Control
Computer Material Computer No. 1
Handling
Process
Signals
Raw
Material Process Contro
Process Preparaticln Computer No.2
Control
Computer No,
Production
T
Process ^Control and Manuf- Process
Signals acturing Control Signals
Equipment
Product Flow 1
Product Product
Line No. ll Product Line No. 2
Handling Mfg. Stages Handling
Finishing Finishing Finishing

and Inspec and Storagje and Storagje—'
tion Equip Prep. No.1 Prep. No.2
l_f
Storage
Storage
and and
Inventory Inventory
Secondary
rocess Shipping
1NO. 1
Block Diagram of Plant Control

Scheme With Fully Integrated
Feedback Using Direct Interface
Secondary Between The Host Data Processing
Process Computer And Segregated Process
—*- to. 2 Control Computers
FIGURE 5
30
being used had been well kept up or else had been totally non-
existent.
The requirements of data base construction include means of data
acquisition and a means of changing the known data. To achieve
simplicity the system must have a minimum amount of input formats.
What's needed is one universal input format sheet for use by both
mechanical and electrical personnel, which can be used to add or
change information in the data base. The electrical equipment will
include normal electrical equipment and control and the electronic
equipment and control (including the process control computer). The
mechanical equipment must also be subdivided to consider hydraulic
requirements as well as lubrication requirements. The following
items are required as input on each piece of equipment in the system:
ITEM RECORD
1. Area subsystem No.
2. Device No.
3. Mechanical PM hours.*
4. Lubrication PM hours (mechanical equipment)*
5. Motor lubrication PM hours.*
6. Control PM hours.*
7. Motor PM hours.*
8. Location on grid. \
9. Description of item.
4
10. Function of item.
i
i
31
11. Check list Nos. - electrical and mechanical,
(related to each item 3 through 7)
12. Mechanical drawing Nos.
13. Hydraulic fluid No. (if applicable).
14. Lubricant name and No. (if applicable).
15. Motor data.
16. Motor drawing Nos.
17. Motor control data (voltage phases electrical

control or electronic control location).
18. Control scheme drawing Nos.
19. Control scheme reference manual Nos.
20. Job safety analysis No. (mechanical and electrical).
* Service code on work feedback card points to which
service is to be performed.
These items listed fulfill the informational needs of the fully
integrated system. Each item in the system will be identifiable by
the first two numbers of the Item Record, Area Subsystem No. and
Device No.. Each field will be assigned the same definition for all
items. Therefore, the input format document need only contain the
two identifying fields, the number of the field to be changed and
the changes to be input. This input/output format for the system
data base may in the final analysis require more vigorous programming
effort but this is the sacrifice to be paid for a less cumbersome
system from the user standpoint. The analyst and programming costs
may be higher. This is a one time cost as compared to wasting the
32
user's time on a daily basis.
In treatment of this subject references will be made for use of
formats and program requirements of the supporting data processing
system. These statements are of a general nature and do not consider
all the details or specifics of a finely tuned management type infor-
mation system. This presentation shall be limited in detail to those
areas that are pertinent to the objectives of this thesis. The only
desire here is to investigate the ability to make the proposed compa-
risons.
It is necessary here to elaborate about the system input data so
that its relevance to the overall system can be understood. Although
the primary purpose of this thesis does not require detailing every
aspect of the design of a preventive maintenance system, an outline
knowledge of the proposed usage base system operation is necessary to
establish whether the results of productivity measurement and actual
production hours preventive maintenance intervals are more successful
than present approaches. A description of each input field mentioned
in the Item Record follows:
1. Area Subsystem Number
As previously mentioned the equipment in the complex
that most normally functions as a unit will be grouped
together. The purposes of this grouping is to make record-
ings of data by the process control computer a more unified
approach. This number will require three numeric charac-
ters.
33
2. Device Number
This number is four digits and identifies a specific
piece of equipment in an area subsystem.
3. Mechanical PM Interval Hours
This four digit number signifies the number of oper-
ating hours required between mechanical service intervals.
4. Electrical or Electronic Control PM Interval Hours
This number defines the number of operating hours
between preventive maintenance.
5. Motor PM Interval Hours
This number defines the number of operating hours
between preventive maintenance service calls.
6. Lubrication PM Interval Hours
This number defines the number of operational hours
between mechanical equipment lubrication. This will always
be a multiple of the inspection interval.
7. Motor Lubrication PM Interval Hours
This number defines the number of operational hours
between motor bearing lubrication. This will always be a
multiple of the inspection interval.
8. Grid Interval Location
This four character alpha numeric designation identi-
fies the location on the manufacturing complex plan map
where the piece of equipment is used.
34
9. Description
This is a brief description of the piece of equipment
such as cylinder, transmission, coiler, shear, transfer,
etc.
10. Function
This title is used to briefly describe the use of the
piece of equipment in the system.
11. Check Lists Numbers
This is a four character alpha numeric designator.
One alpha followed by three numerics define the list of
items to be checked during a scheduled preventive mainten-
ance service call on a unit in an area subsystem. The
alpha is an "M" or "E" for Mechanical checklist or Elec-
trical checklist. There is a checklist for each service
required as defined by the service code pointing to items
3 through 7 of the Item Record.
12. Mechanical Drawing Numbers
This field contains space for key engineering draw-
ing numbers. The complete set of engineering drawing
numbers that define the mechanical device are output in
the information document section of the system which need
not be defined here.
13. Hydraulic Fluid Number
This field defines the recommended hydraulic fluid.
35
13. Hydraulic Fluid Number (Continued)
If no fluid is required the field is blank.
14. Lubricant Number
This field contains the lubrication department's
number for the lubricant required on the unit described.
15. Motor Data
This entry contains information on the required
motor or motors for the application described. Motor
Data includes frame number, voltage, frequency (OHZ = DC),
number of phases, full load current, lubricant number
and carbon brush number.
16. Motor Drawing Numbers
This field contains space for pertinent motor engineer-
ing drawing numbers. The complete set of engineering
drawing numbers that define the motor output in the infor-
mation document section of the system which need not be
defined here.
17. Control Data
This data defines the type and location of the starter
or regulator supplying power to the motor defined above.
If no motor is involved this data defines the control device
such as computer, static logic director or servo control.
36
18. Control Scheme Drawing Numbers
This entry lists the number of the engineering drawings
containing schematics and wiring diagrams of the control
described in the previous entry.
19. Control Reference Manuals
This entry lists the control manufacturer's design
numbers associated with the control device defined under
the given unit number.
20. Job Safety Analysis Numbers
This is a four character alpha numeric designator.
One alpha followed by three numerics define the number of
the document which describes the safety procedures to be
followed during a scheduled inspection. The alpha defines
"M" for Mechanical and "E" for Electrical.
As previously mentioned some basic data for the preventive maint-
enance resided in the host computer. This data and information
essentially revolve around the production schedule forecasting.
This forecast is the schedule of products to be manufactured and
product lines to be operated on a forecast basis one week in advance.
The schedule is reforecast on a daily basis as a result of actual
production figures. (In a later description of the preventive
maintenance system operation it will be shown how the daily refore-
casting will not drastically affect the preventive maintenance
schedule for any given week.) Using the production schedule and
37
stored data describing which area subsystems are to be operated for
each product, a schedule of equipment to be used each shift is
easily generated. This ultimately will be used to complete the week-
ly preventive maintenance work schedule.
As mentioned in the early part of this discussion, the bulk of
the data collection job is assigned to the process control computers.
The data from the process that is required to support this system
are confined to a few areas. The data consist of the manufacturing
delays each shift, production, processing and shipping information
each shift. When a process control computer is used, its normal
operation can and does in the prototype used here incorporate a
utilization function. The utilization function measures expected
times between events in the process. When the fixed times are
exceeded a program records the time. While the process continues
it asks for an input delay reason code from the shift supervisor
through a CRT terminal or other input device. These delays are
summarized at the end of a shift and sent to the host data process-
ing computer for report generation or use as data for systems such
as this one. In the prototype system there are three process con-
trol computers controlling equipment in the areas described previous-
ly as subsystem areas. The control algorithms in the computers
record the number of operations performed by the equipment being
controlled. Also included as a side benefit of the process control
function, the computer tracks each manufactured piece through the
38
system, passing pertinent information about a product from one compu-
ter to another as the product passes into the area of control of a
given computer. As a result of this function, inventories and ship-
ping records are accurately updated. This production information
is also summarized and returned to the host computer giving results
of the productivity of each area subsystem on shift basis.
With the host computer now containing delay information, actual
shift production, stored bogey figures, production schedules and
equipment operating schedule, the equipment usage tables can be up-
dated and maintenance productivity information is closer to reality.
With the basic information sources outlined, an explanation of the
fully integrated preventive maintenance system now becomes more
meaningful. This system will be set up on a weekly schedule. The
data processing computer will generate the information for a week
from Sunday to Saturday by the previous Thursday afternoon. The
data processing computer will use the procedures outlined below to
ready the documents for a new week. A set of documents will be gen-
erated for the mechanical supervisor and the electrical supervisor.
1. The supervisor will return the Work Feedback
Cards for the previous week to the data process-
ing center. This will cause an update of the
records. In all cases the computer subtracts
only the hours listed on the feedback cards from
stored data. This allows for accumulation of
39
operating hours added since the Work Feedback
Card was issued.
The system is designed to retain records of which
Work Feedback Cards were issued. After the up-
date is completed a search of the cards returned
is made and the outstanding cards are flagged
for listing in Exception Report No. 1
2. The computer will use forecasted production
schedules to determine what equipment is to be
idle on all shifts in the next week.
3. The host computer will determine which pieces of
equipment from Step 2 have equalled or exceeded
the inspection interval of actual operating hours.
4. The host computer will determine which pieces of
equipment (if any) scheduled for operation have
equalled or exceeded inspection interval. These
items are an exception to the system.
5. The host computer will produce Work Feedback Cards
for the equipment found in Step 3 above.
6. The host computer will produce a printed list of
the equipment for which the cards were issued in
Step 5 with the data and shift projected for the
equipment to be idled.
40
7. The host computer will produce Exception Report
No. 1 consisting of a list of Work Feedback Cards
not returned and the date issued.
8. The host computer will produce Exception Report
No. 2 consisting of the list of items found in
Step 4. These are items that require servicing
but have not been scheduled down. If service is
performed on any item on this list, a Work Feed-
back Card can be filled out and submitted through
normal procedures.
The formats of the documents shown in Steps 6, 7 and 8 above
are shown in Figures 6, 7 and 8.
Every Thursday the maintenance supervisors each receive the
three reports and a deck of Work Feedback Cards for the next week.
They can now plan their work assignments for all shifts for the next
week. Since the projection of the production schedule may not remain
completely accurate job assignments cascade from one shift to another
as required. This system of cascading assignments as the equipment
becomes available highlights one of the advantages of the preventive
maintenance system. By having a minimum quantity of preventive
maintenance personnel on all shifts all pieces "of equipment will be
serviced when the assigned interval expires. This contrasts to
having a high number of preventive maintenance personnel on possibly
a day shift only. However, the equipment that is to be serviced in
these shift operations may not be available on day shifts when
inspection intervals expire.
41
WEEKLY PM WORK SCHEDULE
AREA NO. UNIT NO. DATE AND SHIFT SCHEDULED
FIGURE 6
Maintenance Weekly PM
Schedule Format
42
EXCEPTION REPORT NUMBER 1
OVERDUE WORK FEEDBACK CARDS
REPORT DATE
AREA NO. UNIT NO. DATE CARD ISSUED
FIGURE 7
PM System Exception Report No. 1 Format
43
EXCEPTION REPORT NUMBER 2
UNITS EXCEEDING PM INTERVAL HOURS BUT

NOT SCHEDULED IDLE IN THE COMING WEEK
REPORT DATE
AREA NO. UNIT NO. INTERVAL HOURS ACTUAL SERVICE HOURS
FIGURE 8
PM System Exception Repoift No. 2 Format
44
It is expected that personnel on all shifts participate in
preventive maintenance work and, therefore, as long as there are no
breakdowns or other failures, the assigned maintenance man must help
the preventive maintenance man scheduled and vice versa.
It is obvious that the two reports and the Work Feedback Cards
are not all that is required by the maintenance personnel to perform
the actual PM work. This brings up the Document Section require-
ments of the Data Processing System. It should be emphasized that
the equipment data retained by the data processing system remains in
storage until changed or deleted and the information is not printed
unless requested by the maintenance personnel. Each of the Work
Feedback Cards contain a checklist number, area subsystem numbers
and device numbers. These numbers define documents that were
generated the last time a change was made on the content of a docu-
ment.
The previously defined system input requires the checklist
number be designated on up to two lists for each piece of equipment
in the system. The lists describe in detail the actions to be
taken by either the Mechanical or Electrical repairman performing
the preventive maintenance work. Lists for each item identified
by the area subsystem number and device number are retained by the
maintenance supervisor. When the maintenance supervisor receives
the work list for each week, he provides the proper checklists for
the maintenance personnel. The checklists are retained in the maint-
enance office on file and are not output from data storage unless
45
updated by the maintenance department. The flexibility of this sys-
tem is readily apparent. Each week the computer system generates an
updated report that eliminates looking back to determine what is
outstanding or incomplete. As the routine maintenance is performed
on each piece of equipment in the system, major problems are avoided.
Equipment that operates on an erratic scheduling sequence can still
be pinpointed as requiring attention through around the clock PM
work schedules. If major problems are suspected in the near future
on equipment as a result of a routine PM service call, a shutdown
can be planned and unnecessary long down time avoided. As experience
is gained about each of the items in the system, inspection intervals
can be increased or decreased to suit each individual case.
The Maintenance Reference Manual is the most comprehensive
document supplied by data processing systems. The manual contains
the complete set of data about each device in the system. This
document is indexed by area subsystem numbers and contains all the
information about each device as described previously in the content
of the Item Record. This document is updated by the maintenance
department using a correction card recognized by the data processing
system.
Many references have been made to the Work Feedback Card with-
out detailing its format and content. The object of this card is to
provide a maximum amount of pertinent information, pointers to the
detailed data document. Pointers to the checklists and allowance for
46
feedback input codes to update the maintenance system. The content
of the card is designed to be contained within eighty columns, thus,
making it compatible to any computer system. The card is generated
by the computer at the beginning of the cycle. After the work is
performed the proper codes are marked on the cards and they are
returned daily to the computer center. When the card is received at
the computer center the written code is properly punched on the card.
The cards are then retained until the update sequence is executed
and the next week's cards and reports are generated. The new cards
and reports repeat the cycle.
Once a card is generated, a new card is not generated automatic-
ally until the old one is returned, coded or destroyed. The computer
system continues to add production hours to all the equipment even
after a card has been issued for inspection servicing. When that
card is returned only the hours indicated on the card when it was
issued are subtracted from the accumulated production hours. Thus,
the system retains the hours put on any piece of equipment even in the
interval between the time a Work Feedback Card is issued for service
and the time the card is read back into the system.
Since the work schedule runs from Sunday to Saturday, and the
new week's schedule is generated on the previous Thursday, it is
obvious that some of the feedback cards for the jobs scheduled on
Thursday, Friday and Saturday of the previous week will be listed on
the Exception Report No. 1. This will serve as not only a check of
47
the systems accuracy but also a reminder for completion to the maint-
enance supervisor. However, as previously explained, no hours are
lost since the update procedure subtracts only the accumulated hours
indicated on the Work Feedback Card. There is however a slight
overlap of hours between the time the Work Feedback Card is issued
and when the work is completed. This means for example, Device A
for which a Work Feedback Card was issued in week one with 2,000
hours accumulated time may operate a maximum of 232 additional hours
before the required work is performed. The net results is that when
the update is made 232 of accumulated hours toward the next inspec-
tion is invalid since as presently defined the computer system does
not retain the daily applied hours so that when the system is updated
the hours up to the completion date are deducted from the accumulated
hours. The system could be built to include this daily accounting
feature after the Work Feedback Card is issued but, the extra work
may not be justified. A study can be made after installation to
determine if the frequency of occurrence mandates redesigning this
part of the system.
The previous example was the worse case where Device A ran
twenty-four hours a day for nine days and sixteen hours on the 10th
day. Whereupon the scheduled down time occurred and Device A was
serviced. It is probable this condition will eventually occur, but
the extra complications in the data processing system may not be
justified. This is true since 240 hours only represents ten days of
48
continuous operation. In industry, once a piece of equipment is
debugged it should operate well over ten days without service.
The Work Feedback Card has been discussed quite extensively in
this chapter, and it is appropriate that the card itself be examined
in detail. The Work Feedback Card is shown in Figure 9. The fields
on this card are as follows:
1. Area number.
2. Device number.
3. Service code - electrical, mechanical, hydraulic
lubrication.
4. Date issued.
5. Number of hours since last inspection.
6. Date and shift to be assigned (unit scheduled down).
7. Date and turn complete.
8. Job completion time.
9. Checklist number.
10. Item Repair Code No. 1.
13. Item Repair Code No. A.
Most of the items in this list were previously described in the
Item Record list. Only the service code and repair code remain un-
defined. The Service Code is used because a given piece of equipment
with electrical, mechanical, hydraulic and lubrication components may
49
w
o CO
oo P
1
1
CTl
e
P=i
00 H «*
m W
i Q
m
m
PH
erf a
O d
(4
<r H m
m <! w •
l PH p
rH o o
** £ V S3
Pi
o H CM
m H .
i PH P
r^- O o -
sr 3 CJ S3
Pi
vT> M rH
< w .
I PH p
CO o o
Q
-3- g o S3
o
Pn
< W
CJ CM H
CO , »H
w 1 33 H o to
cj CTi CJ P S3 CJ
<: CO
AS
§ Pn CO ™
CJ
w oo o wH w 1 p cd
w ro M M pj w
fe
m
1
o
• 2 Pi PH
o
S H €
gs
O O W 0)
CO 13 s H U P ai
p^ P*H
o
& A!
co H 1 P
1 w o
oo
CM
H P M
PH
S3
s
o o
H
^ <! CO cj P
1^-
H p w
H
w Pn M w PQ W * ,
H P M PH S3 S3
CM S3 o o o
P ^ CO !2 CO H P
HI H PH
CM # cj
CO
CO CO
1 o Pi S3
PH
M < M S3
13 O S3 P
" <D
rH
rH P < H W rH CO CO P w P
rH
O
rH
CO W Pi > l-H CJ W CJ o P W
ob
r^
1 Q W > M u W S3 P s pq W PH
■*
CO
< Pi W <: S3 P a PQ W PJ
\^
50
carry the same Device Number. When the Work Feedback Card is
generated a service code must be applied so that the proper category
of work is performed.
The Repair Code encompasses a list of normal repair problems
that may be encountered and overcome. Space is provided for up to
four Repair Codes. If less than four describes the work performed,
the remaining spaces contain blanks.
There are three other computer system interactions required to
fulfill the needs of the maintenance supervisor. These are the New
Item Entry, Item Deletion or Change and Document Request. In prac-
tice, the supervisor fills out a coding sheet with all the fields
properly defined on the sheet. The coding sheet is sent to data
processing to be keypunched and used to update the system. These
three additional inputs are seldom used after the system is in opera-
tion, and most of the interaction will occur with the Work Feedback
Card as the input medium. The document request causes the computer
to print out all the stored information on the requested device.
The original premise for the maintenance system was to design an
effective and desirable tool for the maintenance supervisor. To that
end, the system presented here gives the maintenance supervisor the
following advantages:
1. A complete informational document on all pieces
of equipment.
2. A weekly job assignment planning guide.
51
3. An effective preventive maintenance system that
requires a minimum of paper work.
4. A better manpower utilization guide on a week to
week basis.
A system for improving the maintenance of equipment has been
described up to this point. Substantial savings should occur, simply
from scheduling and control. It is important to note that a by-
product of this will be an automatic maintenance labor productivity
control systems appraisal. This by-product provides the solution
to the remaining problem of maintenance personnel productivity.
When a production worker is assigned to some product line, his
output can easily be measured by the units of output achieved in hours
worked. Even if the total output is attributed to a group of workers,
the individual participation of each worker can be defined. On the
other hand, the support groups of maintenance personnel are not so
easily measured. Various systems of incentive payments have been
developed for maintenance personnel. The systems investigated do

I
not monitor the daily activity of each man and relate that activity
to any planned work and unexpected service calls. The system des-
cribed in this thesis simply makes use of data already produced by the
r
product line, collected by the process control computer and trans-
mitted to the host data processing computer. -From this, maintenance
personnel productivity can be measured.
As previously described in this chapter, the product line delays
52
are automatically recorded by the process control computer. The de-
lays are then transmitted to the data processing computer for
summarizing and report generation. This delay data is also applied
to incentive rate calculation for the production workers. The delay
data along with production output is insufficient for measurement
data on maintenance personnel. There must be some way to determine
if the assigned work Was completed on each shift, and if not,
were the delays experienced on a given shift of sufficient magnitude
to offset the incompleted assigned work.
The assigned work is generated through the preventive maint-
enance system. Over a period of time the work assignments of the
preventive maintenance system can be time studied and the total time
for completion assigned to each job. The maintenance supervisor can
then determine the shift assignments using the weekly equipment
maintenance list and the two weekly exception reports. The
standard time for job completion of each job is listed for each
Work Feedback Card issued. This feature is provided so that the
maintenance supervisor can distribute the work load adequately
across all shifts. Space is provided on the Work Feedback Card
/ to enter the actual job completing time.
It may be that there are not enough preventive maintenance
jobs available to fill a given shift. In this case the mainten-
ance supervisor must use the Additional Work Assignment card.
(NOTE: this is only the second piece of paper required for the
53
entire feedback system.) This card contains shift, date, Device
Number, Service Code, Repair Code, Area Number, Estimated Time
for Completion and Maintenance Supervisor Code No. It is returned
with the Work Feedback Cards for that shift. It acts as a catch all
tool for the undefined jobs that always seem to occur. Its format
is shown in Figure 10.
The host computer now has the complete set of data required to
determine the output of the maintenance personnel for each shift in
a given week. As stated earlier, the Work Feedback Cards and now
the Additional Work Assignment Card are only run once a week. The
Additional Work Assignment file is updated for use with the produc-
tivity calculating and reporting procedures. It has no other
function.
With the shift delay, productive output, Work Feedback Cards
and Additional Work Assignment Cards, all the data necessary for
productivity calculation are available. By defining "Perfect
Productivity" to be Unity(l) Work Completed divided by Work Assigned
is the Productivity Achievement Rating. Work Completed is the sum of
the Work Feedback Cards completed plus Additional Work Assignments
completed plus Service Calls completed as a result of production
delays. Work Assigned to the sum of the Work Feedback Cards assigned
and Additional Work Assignment Cards. The data requirements for the
computer are now completed. The data processing computer can now
report the daily shift productivity for the previous week. This is
54
■4-
O W
00 O0
e 1
j
PS
00 H sr .
<!
I Q
m W O
PS S3
d
PS V
st- H CO
to < H .
I P-i Q
o
iH
m S O
CJ a
PS
o M CM
I PH
.
r~. O o
p -3- § U sz Iu
§ PS
o
o .H cd
ST < w • T3
H 1 P-i Q
a 00 W O o cd
sl- PS U 3
ex
§ CM
c
M si-
c
CD
CO 1 pq £J 
•H
oo
g oo O w
H W 1 Q fO en
o CO
<:
m
I . PS
O M o
go pq
O o w
££ O
CO a pn & H U HJ
3a S M
O
o i •
M
H
si-
co w
H
^
Q W 1 Q o
M
o
I
00
H P M
Pn
< Jz PC
s ffl Spj Hw
W pq
o u o o w
cd
C
Q <3 oo
o o
CM ts oo H U HJ oo •H
<
r^ •H
CM
1 XI
pq HJ <d S3 W
CM
CM
1 Pn O PS s <ti H H Q W S3 H H h H W«
rH
T-t
1 oo w
►=> P-I PS >HMO PS u o a w a d
.H
O
i—1
w H PS > M U M U o Q w
00
r^
1
sf
Q W > M CJ W J3 !=> s pq w PS
CO
< w <
\i_ PS 13 DS W W PS
55
achieved by examining the Work Feedback Card Assignment and Comple-
tion dates and turns. The delays for each shift with time started
and time completed and the Additional Work Assignment Card with
dates and turns for issuance and completion, form the data for the
productivity achievement calculation. These data values are then
plugged into the formula, resulting in some number"'from 0 to 1.0 as
the "Productivity Achievement Rating". Productivity ratings greater
than 1.0 show performance better than projected by the maintenance
supervisor. The final report gives the supervisor by date and turn
the "Productivity Achievement Rating" plus a listing of the Work
Assigned and the Work Achieved. The maintenance supervisor can now
analyze what groups of employees work best together, which delay
problems are most frequent, which men best perform certain jobs,
which men minimize delay time, and many other useful points.
The complete utilization of the power of such a system as this
can only be realized when the data processing computer is put to
further use correlating these data with the other available data.
This system also has possibilities in the area of rate calculation.
To achieve this the system would have to be expanded and refined. It
is presently only designed to be a reporting tool for the maintenance
supervisor.
This chapter describes a scheme for a completely integrated
Preventive Maintenance System and a maintenance measurement system.
56
Now a method must be found to investigate and compare the performance
between the more conventional Calendar Bases System and the Usage
Based System. Inasmuch as setting up this scheme in the real world
requires very specialized and expensive equipment, a simulation must
be made to compare whether the new system works better than the pre-
vious or currently used systems. As a matter of fact two simula-
tions are required. A simulation must be developed for the Calendar
Based and the Usage Based System and the results compared.
Chapter IV deals with the experimental procedures associated
with the actual simulation and its development. At this point,
however, an evaluation of the progress toward solution of the pro-
posed problems is justified. This chapter has developed two schemes
which solve the stated problems at the beginning of this thesis.
The Preventive Maintenance System described reacts on a weekly
basis to all equipment which has equalled or exceeded the production
hours of operation determined as an acceptable interval of preventive
maintenance. Reports and documents are generated that supplement
any needs of the maintenance supervisor. Also shown here is that
the data contained in the overall system with minimal supplemental
data manipulation and information can provide output reports describ-
ing the work productivity on a shift by shift basis for the mainten-
ance personnel.
57
CHAPTER IV
SYSTEM SIMULATION MODEL v
To investigate adequately the ability of any preventive maint-
enance system to react according to the parameters outlined in the
chart in Chapter II and the system described in Chapter III, a simu-
lation model must be designed and operated. Such a model incorporat-
ing every detail of the proposed system presents an enormously
difficult task. However, this thesis investigation described some
basic parameters for a preventive maintenance system. These para-
meters encompass the truly necessary simulation model objectives.
As a future area of investigation, the intricacies of the detailed
system could be modeled. Completion of this thesis investigation
requires a simulation model that adequately compares the two maint-
enance philosophies of Calendar Based control and Usage Based control.
Thej data generated by the simulation also must provide a basis for
evaluation of manpower productivity objectives.
For the model it must be assumed that the data collection sys-
tem using process control computers does function. For the simulation
model both the data collection system and the data processing system
perform their respective duties required either internal or external
to the model. The model deals with a system of operating components
called Device (X) where X is 1 to N (N has a maximum of 200) and the
manpower servicing Device (X) for PM and breakdowns. Where equipment
58
operations are scheduled requiring the operations of groups of
devices, all operating sequences are group oriented. This is similar
to equipment in an area subsystem described in Chapter III operating
as a unit, and failing as a unit.
The events that describe Device (X) include Start of Operation,
End of Operation, Start of Breakdown, End of Breakdown, Start of
Preventive Maintenance and End of Preventive Maintenance. Using
the simulation technique of event control, the model is developed
around Gasp II simulation language.
The output requirements of the model must be Device (X) oriented
and system oriented. The output data must be able to substantiate
the possible advantage of one PM system over another. In the model
PM system, Type I is the system using arbitrary time intervals
between PM service calls (Calendar Based) and PM system Type II
uses the actual operating hours (Usage Based) as a trigger for PM
calls.
Output must also consider manpower productivity between the two
PM systems and also during operating and breakdown periods. The
output data expected from the model includes:
1. For Device (X)
a. Total breakdown time.
b. Number of breakdowns.
c. Number of breakdowns avoided because of PM.
d. Number of PM calls.
59
e. Total operating hours.
f. Number of delays.
g. Total delay time.
2. For System
a. Number of times of no PM man available
b. Number of times PM man used for breakdown.
c. Waiting time for repair calls.
d. Statistics on holding queue for repair calls.
e. Summary statistics on Device Variables.
The model is initialized with the type of maintenance system to
be tested, whether the devices are group operated or individually
operated. The number of repair servicemen available and the number
of PM men available are also part of the set-up data.
Using a psuedo event for initialization, the event file is
filled with a single starting event for each device in the system.
The model then continues on its own, scheduling events as the condi-
tions are met.
The model is designed to handle up to 200 devices individually
or group operated. The groups can be any size evenly divided into
200. Each starting event for a given device triggers an ending
event for that device. The Start Operation event schedules an End
Operation event, and the End Operation event schedules the next Start
Operation event for Device (X). This also occurs for the Start PM
event and the End PM event and the Start Breakdown event and End
Breakdown event.
60
The events themselves have an operating priority. The objective
is to have any device ready to operate when a Start Operation event
occurs so that no delay occurs. If for example, a Start PM event
had begun on Device (X), the End PM event must occur before operation
of Device (X) may begin. In this case the End PM event is triggered
as soon as there is a requirement to use Device (X) for operation.
This is probably the simplest situation to handle.
When a Start Breakdown event occurs, the device must be repaired
before the Start Operation event can occur. Thus the End Breakdown
event must occur and the delay time recorded. If a Start Breakdown
event occurs and the device was previously started into operation,
the operation must be suspended and a delay recorded until an End
Breakdown event occurs. After this event the Start Operation event
is resumed. The only event not suspendable is the Start Breakdown
event.
Since there is no data available and every plant probably
experiences a different set of actual scheduling and delay frequen-
cies, the model uses uniform distributions to calculate and trigger
times between events. The model is designed with a plug-in feature
to change the distributions to any set of data relative to an actual
shop performance record. For this reason, the data generated have no
direct correlation to an actual shop. However, for the purposes of
the comparison of this thesis investigation, as long as the two types
61
of maintenance systems use the same distributions the data and
information derived from the data can be used for conclusive eval-
uation.
The model although built to include group operation is not
completely tested in this mode. Time constraints prohibited final
debugging of the group operation mode. Therefore, this will be left
for future completion. There are more than enough data for compari-
son employing the individual operation mode for the Type I - Calendar
Based PM system and the Type II - Usage Based PM system.
Certain utility routines are used to support the model.
"Fileck" is used to search the event file on a periodic basis to be
sure there is a breakdown in the event file for each device. This
function is required since under both modes of operation, breakdowns
are avoided fifty percent of the time if PM is carried out. This
percentage may be higher or lower in reality and can be modified
easily. However, since the model is event triggered, no breakdowns
( would ever occur after an avoided breakdown was removed from the
event file. In fact, no matter how well any maintenance program
works, breakdowns will occur. "Fileck" does the search and when
no event is found for a device a Start Breakdown event is scheduled
at some event time well beyond the normal interval of breakdowns.
The "Initz" Subroutine is used to initialize the event file and
other arrays that are used as reference data in the model. This
r' ■
event only runs once at the very beginning of the program.
62
There are also some basic philosophical differences in the
operation of the two maintenance modes. These are in the usage of
manpower. In the Type I system, PM men are only available on the day
shift and in Type II, PM men are available on all shifts. Therefore,
to insure equal numbers of PM men in both systfems, Type I is initial-
ized with three men and Type II is initialized with one man. This
works out to a total of 21 man shifts per week in each mode. In the
Type I system, PM events are scheduled for day shift only. The Type
II system schedules PM events around the clock.
Productivity performance in the model is not as detailed and
intricate as the system described in Chapter IV. This is simply
because modeling such details exceeds constraints limiting the writer
to the time frame of this thesis. . Nevertheless, a value can be
determined to compare the utilization of the manpower in one system
with the other. This is thus a productivity factor of one system
over another. This factor for the model is the Total Equipment Delay
Hours divided by the Total Scheduled Operating Hours. The difference
in the two systems shows the response time of maintenance personnel
to failure occurrences.
The model uses the basic premise that PM servicemen are availa-
ble for breakdown work when regular service people are busy on other
repairs. This essentially allows for extra manpower on an around-
the-clock basis for Type II operation. In the Type I system there
are no PM service people other than day shift.
63
Productivity can further be compared by evaluating the para-
meters such
sup as maximum number of repair calls in the queue because
all personnel available for repairs are busy. PM productivity
factors between the two systems are also apparent by comparing
statistics regarding the number of PM calls performed and the num-
ber of PM calls scheduled because the device to be serviced was
operating at the time the preventive maintenance was scheduled.
These are all valid factors since the number of PM manhours worked
is constant for both systems.
To collect data for final comparison, eight runs were made for
each mode of the model. The variables that were changed from run
to run were the number of repair service people and the number of
PM service people. The model run time was 10,200 hours in all cases.
The reason for changing the number of service people was to
determine at what point the number of manhours put into the system
had little or no effect on increasing the up-time of the equipment.
This will also show any limits on the PM program with respect to
increased performance hours of the equipment.
The sixteen runs for data collection include eight runs of each
maintenance system. There are eight manhour configurations used
which are run with each system providing eight sets of parallel data
to be compared.
Runs One through Eight for system Type I as listed in Figure 11
included four repair servicemen and three PM men, six repair service-
men and three PM men, eight repair servicemen and three PM men, ten
64
repair servicemen and three PM men, four repair servicemen and six
PM men, six repair servicemen and six PM men, eight repair service-
men and six PM men, and ten repair servicemen and six PM men,
respectively. The corresponding Runs for One through Eight for sys-
tem Type II as listed in Figure 11 included four repair servicemen
and one PM man, six repair servicemen and one PM man, eight repair
servicemen and one PM man, ten repair servicemen and one PM man,
four repair servicemen and two PM men, six repair servicemen and
two PM men, eight repair servicemen and two PM men, and ten repair
servicemen and two PM men, respectively. The data collected from
these runs provided the basis for concluding the results of this
investigation.
Flow charts of the major event subroutines of the simulation
model are included in Appendix I.
65
CHAPTER V
RESULTS AND CONCLUSION
The model was run to simulate 10,200 hours or 425 days using
the eight different manpower configurations for each mode of opera-
tion. The data clearly indicates there are gains in all areas where
scheduling of PM is done on a usage based philosophy (Type II mode).
Figure 11 shows the comparison of the data values collected and
calculated for the two systems. The key to the column numbers
defining the variables compared in Figure II are as follows:
Variable No. 1 - Percent Delay Hours of Scheduled
Operating Hours.
Variable No. 2 - Percent Delay Hours of Scheduled
Manhours.
Variable No. 3 - Number of Breakdowns Avoided
Because of PM.
Variable No. 4 - Number of PM Calls Performed.
Variable No. 5 - Number of PM Calls Rescheduled
Because Equipment Was Operating.
Variable No. 6 - Highest Number of Service Calls
In Repair Queue.
Variable No. 7 - Average Waiting Hours In Repair Queue.
Variable No. 8 - Total Number of Service Calls in
Repair Queue.
66
m
I u rH r^- UO IT. vO -3- 00 vO CO
00 3
* r-t -H M 00 -3- CO CO «X> CO CO CO
OT
Z« cd rH
M n. CJ O*
a) iH 7: CU CU
f^ -tf- CM 00 CM CTi UO U~l
M CTi m CO CM o\ KT •J- CM
3 O CU -H 3 3 m rH u"l rH
Z H 43 P> M O
M vO IO CM rH CM -vf rH vO
C
^ .
CU 00 IH
M
r^
U 60 C )-i
rH 00 rH VO CT\ ~3- CO
rH ■o- CM CM CO CO CM rH
cu to T-I co -H cu
fl K w h Bl 3
1 0) -H 3 ft 0) vO vO <r m VO CM CO CM
3 > fl O 0) 3 M
Z<SS CtiO vO 00 CM u-i r^ rH vO
CM rH rH
4-1
vo 0 a IH CT\ CJ> CN r- r^. r^. rH 00
4-> CU M IH CO CO CO CM CO CO CO rH
>-J CO U O
CU (U CU -H CD (U
,0 42 ,Q > rH 3
S 00 6 U rH CO
3 -H 3 CU cd 3 IH CM CM r^- CO CO u-| Os 00
a a: z w a 0 CO CO CM CM CO CN rH
4-1 CJ
u-i 0 tud
rH 00 •0- -3- -* r^ r^. U"l
M CO rH o> 00 rH r^ 00 CM
IH 00 00 r~. r^. 00 r~^ r^. 00
m m m
Number
Number
PM Cal
Reschei
u-i u-l uo m UO
CO 00 <f u-i CO CO CM vO
vO r^. vO o> <f r^ 0 UO
M vO m vO -3- CO CO CM u-|
0- -a-
00 rH u-| CM 0 00
<f 4-1
O
T3
CO CU M
rH CTi
UO
rH rH v£) -<r UO 0
-d- vO VD 00 00 00 <T\
U 1—1 S IH u-| uo UO -* uo u-| u-i u-i
CU h H fi
rg CU ca 0
B
5
13
£> U 4-1
§ s ^
CM
O
00
r~~
CO
O^
CO
00
a\
u-i
r-^
o>
C3>
r^
<)•
CD>
3
5
Z o-t PM ^ M 00
rH
r-~
rH
r^
rH
r^
rH
r^
rH
r^
rH
r^
rH
r~~
rH
8
CO 4-1 Pn
CO 4-1 3 0 1-^ rH 00 00 u-| CM CM -J-
O 3 M 00 m CM r-^ rH O vO CM
U O *3 CU M 0 rH rH rH CO CO CM CO
CU l-l T3 CU CO CM CM CM CM CM CN CM CM
^ QJ 4d T3 3
B 43 n) vH cd
3 6 QJ 0 O CO rH O-i <fr vO O CO UO
Z 3 U > QJ 2 IH vO CO r-^ uo <fr -<1- ul -J-
2 M <J pq PM vO vO vO VO vO vO
co
CM ir-l CN 00 rH 00 u-l rH
3 1 1 M
U O T3 3 M 00 CM CTi r-~ u-i 0 00 r^
QJ -U PC CU cd1 rH rH rH rH
_Q cl <-• 5T
6 cu >-, 0 co
3 tJ Jd wo u rH 00 0 CTi VO VO CO <!-
Z rl rH QJ 3 H
QJ QJ 4-1 rH 0 CM <r rH 00 r-^ CM CTi 00
PM Q O 3 ,3 CM rH rH rH rH
- e2M l~- 0 00 O CO 00 o>

ON o> l-~ r~ o> r^» r-- r^
3 3 00 M
>-< O .5 3 M CM CM CM CM CM CM CM CM
cu 4-j nc ^ -H
c
-9
B .c w
cu >-, *o cd co r~ vo CM u-i CO O CTi O
3 0 cd ui u u UO CO CM CM ~3" co O CO
Z r-l rH QJ 3 M
QJ QJ 4-1 ft O CO CO CO CO CO CO CO CO
PM Q O O EC
rH CM CO -3- U~l VO r-~ 00.

QJ |
rH O,
43 -H B O O 0 0 O O O 0
Cd U CD CU Z Z z z Z Z Z Z
•HOC 4J QJ
t-i CO O CD Cu c 3 c 3 3 c 3 c
cd QJ -H
> Q 4-1 ££ 3
Pd
3
PtS
3
Pd
3 3 3 3
pci 3
67
The most significant difference between the two systems is
found in the preventive maintenance aspects of the systems. The
number of breakdowns avoided in system Type II were in all cases
at least three times greater than system Type I. This is easily
explained wit,h the support of Variables Four and Five.
Variable Four depicts the total number of PM calls performed.
This value for Type II is almost 2.5 times greater than Type I.
There are several reasons that account for this. In the Type I
system PM occurs on day shift only and if the device was operating
when scheduled for PM, the work could not be performed. The work
had to be rescheduled until the unit was operating and it was also
day shift. Since the Type II system has PM men in all shifts, the
PM work is scheduled for as soon as the device ended operation.
Thus the chance of missing a PM check was diminished.
Variable Five shows the number-of PM calls that had to be
rescheduled because the device was operating at the time the PM was
due. The most significant factor about this variable is for system
Type I, where the total number of PM calls rescheduled is significant-
ly larger than the actual number of PM calls performed. This means
that some devices were rescheduled so often that they overlapped to
the next PM cycle before they were serviced. The Type II system
shows that although many devices had PM rescheduled, they always
were serviced soon after the End Operation event occurred. The
next Start PM event could not occur until the accumulation of
68
operating hours triggered the event.
Variables Six, Seven and Eight deal with the repair queue.
The only outstanding fact evidenced here shows the total number in
the queue over the entire run was significantly less in the Type II
system when the number of servicemen was less than eight. The
reason for this is that the PM man was available on all shifts to
assist the servicemen on repair work if the need arose. This
essentially provided additional help when needed to keep the queue
low.
Variable One indicates there were more productive operating
hours available because the failures were handled faster. This is
reflected as the percentage of delay hours of scheduled operating
hours. This number is derived by dividing the Total Operating Hours
by the summation of Total Operating Hours and Delay Hours in the eight
comparison runs of the model. The differences ranged from .31 to .6
percent gain in productive hours. These gains are derived from the
hypothetical scheduling data used in this model. The actual gains
may be greater if actual plant equipment failure and scheduling
data could be used in the model. The fact that identical scheduling
input data used in the comparison runs produced gains indicates
some degree of success.
Variable two expresses these gains in production time as a per-
centage of the Delay Hours of Maintenance Manhours expended. In all
cases the Type II runs showed better percentages. The combined
69
information gained from Variables One and Two indicates that as a
direct result of the utilization of manpower in the Type II system
production gains are achieved. This supports a basic premise of
this investigation. The Type II system of PM by actual operating
hours and around the clock scheduling of PM personnel can provide
better PM performance and also help reduce delays resulting from
equipment failures.
The around the clock scheduling of PM personnel also provides
additional emergency support for other servicemen when conditions
occur that require additional repair support. This minor reallo-
cation of manpower from the Type I system results in production
gains for the entire plant or shop. This should result in profita-
bility gains for the plant or shop.
70
CHAPTER VI
SUGGESTIONS FOR FURTHER INVESTIGATION
It seems clear that there is merit in Type II operation of
Preventive Maintenance Systems. However, the problem will be to
get a real test situation started. Only actual data can truly
substantiate what has been discovered statistically. It may be
that some additional work should be done on the model to investi-
gate some of the other areas to determine their effects on the
overall system. These areas include not only the completion group
operation philosophy, but also the cost accounting aspects of the
model. Both of these areas could provide valuable information for
s tudy.
The model as it is presently configured included as comprehen-
sive a set of parameters governing maintenance activities as is
typically experienced. The ground work prepared in this model pro-
vides an excellent spring board for the future study of the areas
mentioned.
The task at hand is to work toward introduction of this system
into an operating shop. Typically such an extreme change from the
norm will not be easy but well worth the effort.
71
APPENDIX I
SIMULATION MODEL
MAJOR SUBROUTINE FLOWCHARTS
72
SUBROUTINE
START OPERATION
Store Time To Cal-

culate Time For
START Accumulated Hours
For PM Type II
Collect Service
Is Device (X) Statistics On
Used In Group Oper Device (X)
ation Set Flag For Device
(X) Busy For Oper.
No
Is Device (X) Is End Simulation

Out of Service Flag = 1
No Yes No
Is Device PM Flag
Is Device (X) Set From End PM
Undergoing PM Event For Type II
System
No
Schedule Next End

Operation Event
Store Time For Total RETURN
Operating Hours
73
SUBROUTINE
START OPERATION
Is Group Ready Is Device (X)

For Operation Busy For PM
Yes Yes
Call End PM Event

For Device (X)
Is This The
Device In The Group i jqQ
That Is Out Of
Service
Yes
Store Time To Store Time To

Calculate Break- Calculate Delay
down Time For The Group
Set Start Delay Set Start Delay

Flag For Device Flag For Group
(X) (N)
0
74
SUBROUTINE
START OPERATION
Is Any Device (X) Yes

In Group Busy
For PM ZL
Call End PM
For Devices Busy
No On PM
Set Group Busy

Flag
Call Start
Operation Events
For All Devices
In The Group
Call End PM
Event For Device
(X)
0
75
SUBROUTINE
START OPERATION
Has Device (X)

Reached Number Of YES
Operating Hours To
Schedule PM In Type
II- System
NO
Schedule Next
Start PM Event
For Device (X)
Collect Statistics
On Time Between
PM Event
Set.PM Operating
Hours Equal To
Zero For Device (X)
Reset Flags To
Trigger Next PM
Event For Device
(X)
0
76
SUBROUTINE
END OPERATION
Reset Device (X)

Busy For Operation
Flag
START
Is This A Group RETURN

Operated Device
NO
Is The End
Simulation Flag Call End Operation
Set Event For All
Other Devices
In The Group
NO
Schedule Next Reset Group

Start Operation Busy For Opera-
Event For tion Flag
Device (X)
Use Current Time

To Calculate
Total Time Unit
Operated
77
SUBROUTINE
END OPERATION
Calculate Group Calculate Delay

Breakdown Time Then Time And Add To
Subtract Time From Total Delay Time
Accumulated Operat- For Device (X)
ing Hours
<D
Reset Start Delay
RETURN Flag
Subtract Delay Time

From Accumulated
Operating Hours For
Device (X)
78
SUBROUTINE
START BREAKDOWN
Collect Statistics
On Number Of
Failures While In
Operation For Device
(X)
Is Device (X) Store Time To

In Operation Now Calculate Total
Time Of Failure
NO
Set Flag
Is Maintenance IBWOF (X) = 1
Man Available
YES
Set Maintenance
Man Busy Flag
Is From Queue
Equal to 1
79
SUBROUTINE
START BREAKDOWN
Schedule Next
Is Number In End Breakdown
Queue Equal To For Device (X)
Zero
Is Device (X)
Group Operated
Set Device (X)
Out Of Service
YES
Store Time To
© Set Group Out
Of Service Flag
Calculate Delay Via Device (X)
In End Breakdown
Store Time To
Collect Statistics Calculate Group
On Number Of Delay Time
Breakdowns Device
(X)
JL
Collect Time
Related Statistics RETURN
On Service
80
SUBROUTINE
START BREAKDOWN
© Collect Statistics
On Number In
Is Flag For Break- \ YES Queue
down While Operating\ /" N
Positive For 1—( 5 J
Device (X) / ^—'
NO Reduce Number In
Queue By One
Store Time And
File Breakdown
In Queue
Set Device (X)

Just Stored In
Queue Out Of
Service
Remove Item In
Queue With Longest
Waiting Time
Collect Statistics
On Waiting Time
81
SUBROUTINE
START BREAKDOWN
Call End Breakdown

Set Emergency
Service Flag Subroutine
Positive
Set Device'(X) Reset Queue

Out Of Service Bypass Flag
IBYQUE - 0
Find First End

Breakdown Event
Call Subroutine
End Breakdown
Set Up Attributes
For Original
Start Breakdown
Event
Set Flags To
Bypass Queue In
End Breakdown
IBYQUE = 1
82
SUBROUTINE
START BREAKDOWN
Q
Is This Type I
TYPE I Is From Queue
Or Type II
System Flag Equal To One
TYPE II
Is PM Man
Available
YES
Is The Time Of The
Set PM Man Event Equal To Day
Busy Flag Shift Hours
NO
Set PM Man Is Breakdown While

Assigned To Operating Flag
Breakdown Work Positive
Flag (IBWOF(X))
Collect Statistics
On Number Of Is From Queue
Times PM Man Used Flag Equal To 1
For Breakdown
©
83
SUBROUTINE
START BREAKDOWN
©
Collect Statistics
On The Number Of
Store Time And
Times PM Man Used
Set Device (X)
For Breakdown
Out Of Service
Add 1 To Number
© Set PM Man Busy
For Breakdown Work
In Queue And Store
Event In Queue
Is From Queue Flag \ YES

Equal to 1
NO
Is PM Man Available
©
84
SUBROUTINE
START BREAKDOWN
£
Collect Statistics
On Number Of Times
1
Is Breakdown While
Operating Failure
No PM Man Available Flag Positive
For Breakdown Work
NO
Is PMBK Put Event In

Flag Equal to Zero Queue File
NO
Find First End PM If From Queue YES

Event To Free PM Flag Equal to 1
Man
Call End PM Subrou-
tine
NO
Store Time Event
Is Filed In Queue
And Set Device (X)
Out Of Service
Add One To
Number In Queue
85
SUBROUTINE
END BREAKDOWN
Is Emergency Service
START Flag Set To One [_[ ^Q
NO
Is Emergency Calculate Delay

Service Flag Time
Set To One NO
YES
Deduct Delay Hours

Set Bypass Flag
To One and Set From Total Operat-
Skip Flag to One ing Hours And
Accumulated PM Hours
Reset Flag For

Is Breakdown While Breakdown While In'
Operating Flag Set Operation To Zero
To One For Device For Device (X)
(X)
NO
Is Bypass Flag \ N0
Set To One
YES
©
86
SUBROUTINE
END BREAKDOWN
Reschedule Current
Is End Simulation
YES Event For Attrib(l)
Flag Equal To One
+ 2 To 10 Time
Units
NO
Reset Bypass Flag

Schedule Next Start To Zero
Breakdown Event
For Device (X)
Reset Emergency
Service Flag To
Zero
Is This Type I Or
Type II System
TYPE II RETURN
-«
Is Flag Set For PM

Man Busy On Break-
down Work
NO
©
87
SUBROUTINE
END BREAKDOWN
Collect Statistics
On PM Man Status
On Breakdown Work
Collect Statistics
On Repair Man Status
And Add One To Avail'
able Serviceman Pool
If SUP flat Is Zero Reset Flag For PM
Man Busy To Zero
Is Start Delay Flag

Set Reset Flag For PM
Man Busy On Break-
down Work to Zero
NO
Calculate Breakdown
Time For Device (X)
Y
,ur\
Did Breakdown Occur 3L
On Day Shift
Reset Flag For
Device (X) Out Of
Service To Zero
88
SUBROUTINE
END BREAKDOWN
Calculate Delay
Time For Group
Is This Device A
Group Operated
Device
YES
Reset Start Group
Delay Flag
Is Flag Set
Indicating This
Device Caused The
Group To Be Out Of
Service
NO
Reset Group Failure

Flag For Device (X)
Is Start Delay
Flag Set For This
Group
NO
Reset Group Out

Of Service Flag
6
89
SUBROUTINE
END BREAKDOWN
Set From Que Flag

Equal To One
Is By Que Flag
Set To One
YES NO Remove First

Event In Que
Is The Number Of NO
Repairs In The Que
Equal To Zero Call Subroutine
Start Breakdown
YES
Reset From Que

Flag To Zero
&
90
SUBROUTINE
START PM
Is PM Man
Available
START
Is Device (X)
Busy For Operation
YES
Collect Statistics
On Number Of Times
Device (X) Was Busy
Is This Type I Reschedule PM For

Or Type II System T Now Plus 24
/TYPE I Time Units
TYPE II
Reschedule PM For Collect Statistics

Next End Operation On Number of Times
Event For Device PM Rescheduled For
(X) Plus One Time Type I System
Unit
91
SUBROUTINE
START PM
Collect Statistics Store Time To

On Number Of Times Calculate Total PM
PM Rescheduled For Time
System II Set Device (X) Busy
For PM
, (
\
Collect Statistics
On PM Man Busy
Set PM Man Busy
RETURN Flag
Cancel Next
Scheduled Breakdown
For Device (X)
50% Of The Time
Collect Statistics
On Number Of Times
No PM Man Available Set Flag If Break-
For Service down Avoided
Reschedule PM For
TNOW Plus One Hour
Is Flag Set For
Breakdown Avoided
YES
92
SUBROUTINE
START PM
Collect Statistics
On Number Of Break-
downs Avoided For
Device (X)
Schedule Next End

Reset Breakdown
PM Event For Device
Avoided Flag
(X)
<D
Collect Statistics
On Number Of PM
Calls For Device (X)
0
93
SUBROUTINE
END PM
Reset Device (X)

Busy For PM Flag
START
<D
Reset PM Man Busy
Flag Only If Flag
For PM Man On Break-
down Work Is Not
Is This Type I \TYPE I
Or Type II System
-Sfi£
TYPE II
RETURN
Set Schedule PM
Flag For Device
(X)
Is End Simulation YES

Collect Statistics Flag Equal To One
On Number Of
PM Hours
NO
Schedule Next
Collect Statistics Start PM Event
On PM Man Status During Day Shift
Only
94
SUBROUTINE
END PM
Collect Statistics
On PM Man Busy
Reset Device (X)

Busy For PM Flag
95
SUBROUTINE
END SIMULATION
Call All End

START Operation Events
Set Ends Call All Start

Flag Equal To One PM Events
Call All Start

Breakdown Events Call All End PM
Events
Collect Final
Call All End Statistics
Breakdown Events MSTOP - 1
NORPT = 0
Reset Ends
Call All Start Flag Equal To Zero
Operation Events
RETURN
96
BIBLIOGRAPHY
Books
Glans, Grad, Holstein, Meyers, Schmidt, Management Systems, Holt,

Rlnehart, Winston, New York, New York, 1958.
Gutterman, Milton M. ED., Computer Applications - 1962, Spartan Book,

Baltimore, 1964.
Hildebrand, James K., Maintenance Turns To The Computer, Cahner Books,

Division of Cahners Publishing Co., Inc., Boston, Massachusetts,
1972.
Maynard H. B., ED., Industrial Engineering Hand Book, Second Edition,

McGraw-Hill Book Company, New York, 1966.
Morrow, L. C, ED., Maintenance Engineering Hand Book, Second Edition,

McGraw-Hill Book Company, New York, 1966.
Mundel, Motion and Time Study Principles and Practices, Prentice Hall,
Incorporated, Englewood Cliffs, New Jersey, 1970.
Murdick and Ross, Information Systems For Modern Management, Prentice

Hall, Incorporated, Englewood Cliffs, New Jersey, 1971.
Pritsker and Kiviat, Simulation With Gasp II, Prentice Hall, Incorpor-
ated, Englewood Cliffs, New Jersey, 1969.
Savas, Emanuel S., ET AL, Computer Control of Industrial Processes,

McGraw-Hill, New York, 1965.
Schmidt and Taylor, Simulation and Analysis of Industrial Systems,

Richard D. Irwin, Inc., Homewood, Illinois, 1970.
Schoderbek, Management Systems, John Wiley and Sons, Incorporated,

New York, New York, 1971.
Wolf, Arthur E., Computerize Plant Information Systems, Prentice Hall,

Incorporated, Englewood Cliffs, New Jersey, 1974.
97
BIBLIOGRAPHY
Articles
Highway Research Record No. 347, Maintenance Planning and Supervi-

sion, Highway Research Board, Division of Engineering, National
Research Council National Academy of Sciences - National Academy of
Engineering, Washington, D.C., 1971.
Highway Research Record No. 61, Maintenance Practices - 1963, Method

and Materials, Highway Research Board, Division of Engineering,
National Research Council National Academy of Sciences - National
Academy of Engineering, Washington, D.C., 1964.
Highway Research Record No. 359, Maintenance Operations and Applied

Systems Engineering, Highway Research Board, Division of Engineering,
National Research Council National Academy of Sciences - National
Academy of Engineering, Washington, D.C., 1971.
98
VITA
PERSONAL HISTORY
Name: James Andrew Cichelli
Birthplace: Baltimore, Maryland
Birthdate: April 11, 1944
Parents: Andrew E. Cichelli

C. Doris Cichelli
Wife: Gail Patricia
Children: Jami Ann, Todd Jason
EDUCATIONAL BACKGROUND
Villanova University
B.E.E. College of Engineering - 1966 1962 - 1966
Lehigh University
M.S. in Industrial Engineering - 1977 1971 - 1977
PROFESSIONAL EXPERIENCE
Bethlehem Steel Corporation
Looper 1966 - 1967
Technical Assistant 1967 - 1969
Assistant Electrical Engineer 1969 - 1972
Electronics Foreman 1972 - 1973
Electrical Engineer 1973 - 1976
Electrical General Foreman 1976 -
99

An Evaluation of Calendar Based Preventive Maintenance System Versus Usage Based Preventive Maintenance System

Uploaded by

Copyright:

Available Formats

You might also like

An Evaluation of Calendar Based Preventive Maintenance System Versus Usage Based Preventive Maintenance System

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

An Evaluation of Calendar Based Preventive Maintenance System Versus Usage Based Preventive Maintenance System

Uploaded by

Copyright:

Available Formats

View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

An evaluation of calendar based preventive

Follow this and additional works at: http://preserve.lehigh.edu/etd

BASED PREVENTIVE MAINTENANCE

SYSTEM VERSUS USAGE BASED

PREVENTIVE MAINTENANCE SYSTEM

Presented to the Graduate Faculty

in Candidacy for the Degree of

All rights reserved

INFORMATION TO ALL USERS

All rights reserved.

This thesis is accepted and approved in partial fulfillment of

the requirements for the degree of Master of Science.

The author wishes to express his appreciation to his thesis

advisor, Professor W. J. Richardson for his guidance in the study

presented here, and to Professor M. Groover for serving as a second-

I also wish to express sincere appreciation to Ms. M. B. Hovik

of the Bethlehem Steel Corporation Research Department for her

assistance and suggestions during the operation of the simulation

model employed in this study.

Most of all I wish to express my deepest thanks to my wife,

quired for this thesis.

III Proposed System Description 26

IV System Simulation Model 58

V Results and Conclusions 66

VI Suggestions For Further Investigation 71

Appendix I Simulation Model Major Subroutine 72

Chart of Management Objectives Relating to 5-7

Block Diagram of Shop Control Scheme With 19

Block Diagram of Plant Control Scheme With Fully 20

Chart of Management Objectives Relating to 23 - 25

Block Diagram of Plant Control Schemes With Fully 30

6 Maintenance Weekly PM Schedule Format 42

7 P.M. System Exception Report No. 1 Format 43

8 P.M. System Exception Report No. 2 Format 44

9 P.M. System Work Feedback Card Format 50

10 P.M. System Additional Work Assignment Card 55

11 Simulation Model Data Comparison 67

Preventive Maintenance Systems as a rule do not lend themselves

to easy evaluation as to effectiveness. No one PM (Preventive Maint-

most common and generally accepted system of PM is the periodic

interval system or calendar based. However, this system has a num-

To overcome some of these shortcomings, the usages based system

has certain advantages. Implementation of such a system involves

more precise data collection requirements and direct interaction with

the process and equipment. It is necessary to make a judgement

whether the advantages of such a system outweigh the possible

difficulties of implementation. Only a complete evaluation of the

operation of the two schemes could develop such an answer. A

practical evaluation of the two schemes from a real-world install-

ation would be time consuming and difficult to complete. A simula-

tion model developed around a real-world situation to perform this

function would provide the data to adequately evaluate both systems.

The simulation model developed in this thesis provided the data to

The model output data showed significant differences in man-

power utilization and equipment performance in favor of the usage

based system. It was found by keeping the manhour parameter constant