Document created: 29 September 03
Air University Review, May-June 1974
Lieutenant Colonel Monroe T. Smith
On 1 December 1971, Major George McKee, then Director of Maintenance Engineering, Hq USAF, pushed a button starting a test program of enormous impact upon Air Force base-level maintenance management. This service test, known as Maintenance Management Information and Control System (MMICS), was conducted at K. I. Sawyer AFB, Michigan, from 1 December 1971 to 28 February 1973.1 This article deals with that service test and the ramifications such a system will have on future base-level maintenance management. I say “will have” because that system is currently being implemented in small packages by the Air Force Data Design Center, Gunter AFS, Alabama.
Before describing the service test, let’s focus on the subject of Air Force maintenance. Air Force maintenance is big business. Over one-third of the Air Force’s people are involved in maintenance. They receive an annual salary of over $1.5 billion, spending or influencing the expenditure of one-third of the Air Force budget.2 I could continue with the “gee whiz” figures, but the point is that the number of dollars and people involved in Air Force maintenance offers untold challenges in its proper management.
Maintenance management offers its greatest challenge at base level. This is where the “rubber meets the road.” All the plans, programs, schemes, etc., up and down the chain of command come into focus at base level.
How do we do this job today—how do we manage our base-level maintenance
organizations in this nuclear age of third-generation computers? Manually, that’s
how! We manage with pencil, paper, and grease boards. With all our
sophisticated weapon systems and electronic marvels of the 20th century, we
manage these fabulous systems with “little or no aid from mechanical or
electronic devices.”3
Typical of this manual operation is the requirement from the maintenance “bible,”
Air Force Manual 66-1, Maintenance Management: “Visual aids will be neat
in appearance, covered with a transparent material to permit posting with
grease pencil. . . .”4
MMICS was a test to see if a real-time, large-capacity, on-line computer could be used at base level to improve maintenance management.5 For the sake of clarity, “real time” in this system means that the system obtains and presents data and produces outputs fast enough to affect the minute-by-minute performance of maintenance. Similarly, “on-line” means that the system receives input data when and where they originate through input and output devices.6 That is as far as I will go into the computer terminology. Rather than talk about MMICS in computer terms, it is more meaningful to talk about the system as viewed from jobs within the system.
general viewpoint
Perhaps the single most onerous task on the flight line or in the shop is the “mountain of paperwork” required in the present system. To correct a defect requiring specialist help involves several people recording that defect on several different forms (AFTO Forms 781, 349, 350, etc.) and grease boards at various times. Each form has its own set of technical orders and manuals detailing form accomplishment. No one who has spent time on any flight line would be surprised at what the RAND researchers had to say about the current management system:
Operating under the current AFM
66-1, managers in the maintenance complex (planners, controllers, crew chiefs,
shop chiefs) spend much of their time processing data—posting to status boards
and filling out or checking a myriad of forms.7
Under MMICS, a discrepancy is recorded once by placing it in the
computer. From that point on, anyone dealing with that item—whether he is a
planner scheduling corrective action, a supervisor inquiring about the aircraft
in question, or the mechanic involved in the actual fix—all will receive the
same information from the output devices of the computer. As each individual
within the management system adds his action to the repair process, he needs
only to input his addition, without reworking or re-recording the entire
problem. When questioned at any point in the repair process, the computer gives
all the information recorded up to the point of inquiry. For example, during
preflight inspection the crew chief discovers a brake leaking. Through a radio
link-up to a remote computer input device, the crew chief inputs all he knows
about the defect: such things as the status (red cross, red diagonal, etc.),
when he discovered it, aircraft identification, work unit code of the brake,
and a narrative of the defect.8
The computer then reproduces this information in job control. Job control
should recognize the potential problem to the daily flying schedule. If for
some reason job control does not react to schedule corrective action, the
computer will respond because it too has the flying schedule. The computer “scans”
not only the daily flying schedule but all operations, maintenance, and
training events automatically at frequent intervals, recognizes conflicts, and
prints notices to the appropriate agency.9
In our example, job control asks the computer if the hydraulic shop has people available for dispatch.10 With a real-time computer capturing control data on a minute-by-minute basis, the computer will have the most current information about specialist availability.
Having a hydraulic specialist available, job control assigns the job a start and stop time. This action causes a work order to be produced on the remote output device in the field maintenance shop area. This work order is unique. It has all the information input by the crew chief and job control, plus some information the computer data bank supplied. Previous maintenance actions by the hydraulic shop on this particular brake system or special tools or equipment or special items of inspection might be included on the work order.11
In our example, however, the specialist proceeds to the job and calls in a
job start. This action satisfies the computer. Had no job start been recorded
within preset time parameters, the computer would have notified job control.12
Among the many things the computer does with job start information is to
decrease the dispatching shop’s personnel availability and initiate actions to
monitor job completion.13
The repairs are completed, and a job completion is radioed in for recording
on the computer remote. By this single input the men are made available for
further dispatch, the aircraft status is upgraded (if no other work is
outstanding), maintenance man-hours are computed, aircraft history is updated,
maintenance data collection information is extracted and stored, to name just a
few actions accomplished automatically, efficiently, and correctly. Even AFM
65-110 equipment status reporting required for many higher headquarters
products, although not in MMICS, is automatically captured and can be
automatically produced by this system.14
viewed from job control
In the current system, job control is the functional element charged with
managing the ongoing minute-to-minute maintenance operation. That office
controls maintenance by authorizing and assigning jobs, work priorities, and
start and stop times. Job control is required to maintain visual aids depicting
each aerospace vehicle, selected ground equipment, maintenance in progress,
munitions installed, specialist availability, and a host of other things.15
The information flowing throughout the current system, in attempting to comply
with these requirements, is overwhelming.16
In MMICS, job control uses the computer as its communications link with the
total ongoing maintenance effort. The computer serves as the storehouse of all
information, handles all the routine work following, and functions to alert job
control when things are not going as scheduled. The daily maintenance and
flying schedules are loaded into the computer at some predetermined time. These
preplanned maintenance actions require no action or monitoring from job
control—so long as the actual jobs start, stop, and progress as planned. Job
control does not get involved if preplanned jobs are not started on time
(within preset parameters) because the computer notifies the shop or activity
first. If, however, the shop or activity supervisor does not correct the
situation within preset time limits, job control is notified by the computer.17
Since all maintenance actions involve computer inputs, job control has at
its request detailed information about the entire fleet. The individual
aircraft configuration, time compliance technical order (TCTO compliance,
delayed discrepancies, specialist availability, equipment location, and many,
many more items of current information are instantly available. As the work day
progresses, unscheduled maintenance is handled by job control’s calling up
various programs, assessing the situation with up-to-the-minute information,
and assigning jobs and priorities based on complete knowledge. The
time-consuming grease-board posting; information recording and re-recording;
chasing, via telephone and radio, the person with the latest information; and
completing the myriad forms and reports—all are a thing of the past in the
computer-assisted system. Job control now spends its effort on the
time-sensitive items that truly require human decision-making; the computer
does the laborious, routine tasks. Job control can now look ahead, anticipate
events, and, in general, control maintenance instead of following it.18
viewed from plans and scheduling
In any well-organized planning and scheduling activity are filing cabinets full of detailed information that must be consulted to preplan cyclic maintenance. Maintenance delayed for some reason fills delayed discrepancy files. Walls are covered with status boards of all types. Reports of all types consume valuable planning time.
In the computer-assisted system, the entire history of each vehicle, along
with detailed cyclic maintenance requirements, is loaded in the data bank. Such
things as technical order compliance are recorded during the normal ongoing
control cycle, and the aircraft histories are updated automatically. Delayed
maintenance may be reviewed in an instant by recalling that program. For
reports, the planner needs only to ask for the information in the desired
format. Internal computer programs generate tentative maintenance plans, taking
cyclic requirements into consideration along with operational requirements.
Once tentative maintenance plans and schedules are confirmed by the planner,
the computer produces the plan and issues the work orders—both former
time-consuming tasks. These work orders are produced containing much of the
information previously researched by the mechanic. Freed from the laborious
manual tasks, the plans and scheduling people now have the time and information
to plan more effectively.19
viewed from supervisory levels
Base-level maintenance has suffered from a glut of information, yet, because
of manual processing methods, timely decisions based on up-to-date knowledge
were often lacking.20
In the computer-assisted system, questions that once required hours of research and yielded questionable products are answered by the computer on line and in real time. For example, the chief of maintenance asks the status of a certain modification. The inquiry is answered with electronic speed and computer accuracy. Since modification is an ongoing maintenance action scheduled and controlled through computer action, job completion automatically updates the modification status information within the computer. Intermediate supervisors might call up training requirements/schedules to plan work shifts. A simple inquiry returns the current, up-to-the-minute information. These, like hundreds more day-to-day questions now requiring manual research and reporting, are handled in microseconds. But more than that, the answers are the most current obtainable because all ongoing actions update system statuses and equipment histories. Through internal computer programs these ongoing control data are captured and massaged into whatever format is needed not only for base consumption but for off-base use as well. The entire data bank of the maintenance complex—from personnel to equipment to schedules to problems—is instantly available to supervisors at all levels.
problems
Lest I suggest that the computer-assisted system is a panacea for all base-level maintenance management ills, some formidable problems must be addressed. Maintenance has traditionally rejected formal job standards. For this computer system to work, fairly accurate job standards must be developed, implemented, and constantly updated. Without valid job standards, computer planning and scheduling become a farce. (The same problem exists in today’s manual system, but we get around it by not following the schedule with any degree of accuracy.) Associated with this problem is the operational flying schedule. Again the number and type of sorties must be given to maintenance well ahead of the time needed for input to the computer.
Another problem that must be addressed is the control of ongoing tasks. As indicated in the RAND studies, job control has never really controlled the ongoing tasks—the job control people have followed the work, and only the most flagrant “conflicts” came to their attention. With the computer monitoring all tasks—maintenance jobs, training programs, etc.—any deviation will be noted and addressed by the computer. With literally hundreds of tasks being accomplished hour by hour, nonreporting of starts, stops, and deviations will inundate supervisors and job control with problem notices. The entire maintenance complex must become attuned to getting instructions from the computer and telling the computer what they are doing and when they are doing it. In other words, maintenance people MUST agree to extremely close control.
Moreover, using the computer to capture ongoing data and using the data for submission of off-base reports means that the actual work and training accomplished, inspection and analyses performed, and equipment kept in readiness go automatically to higher headquarters. Many management goals may have to be adjusted once such accurate and timely data become available.
Finally, as with any computer system, the data retrieved are only as good as the data input. MMICS is a dynamic system, highly dependent upon accurate and timely inputs. This may prove the most critical problem of all.
Base-Level maintenance management is in for a revolution. The Air Force Data Design Center is implementing MMICS in several “packages.” As each package is implemented, it will require adjustment in the way we do business. Until the last package is implemented, there will be some duplication of effort. This short-time duplication will be more than offset by the increased capability brought about by each package. Moreover, the knowledge that full MMICS implementation will result in a fully integrated data system, furnishing information on-line and in real time not only for the control of ongoing maintenance but also for off-base reports, should lend enthusiasm for the system. MMICS is a big step in closing the gap between our modern weapon systems and our manual management system.
Air War College
Notes
1. Interview with Major James W. Frank, Project Officer for MMICS, AF Data Design Center, Gunter AFS, AL.
2. Major General L. F. Tanberg, “Maintenance in the Seventies,” Air University Review, XXI, 5 (July-August 1970), p. 18.
3. Philip J. Kiviat, “Computer-Assisted Maintenance Planning,” RAND Corporation, Santa Monica, California, July 1965, RM-4562PR, p. 8.
4. AFM 66-1, Maintenance Management, vol. 2, “Chief of Maintenance (Aircraft and Missile),” 1 August 1972, Department of the Air Force, para 2-11.
5. “Maintenance Management Information and Control System, Project Plan,” Hq USAF, November 1970, p. 1.
6. S. M. Drezner and R. L. Van Horn, “Design Considerations for CAMCOS—A Computer-Assisted Maintenance Planning and Control System,” RAND Corporation, Santa Monica, California, July 1967, RM-5255PR, p. v (hereafter referred to as CAMCOS).
7. Ibid., p. 15.
8. AFM 66-176 (Test), Maintenance Management Information and Control System (MMICS), vol. 3, 1 September 1971, Department of the Air Force, p. 44-1-5.
9. Ibid., p. 30-1-14.
10. Ibid., p. 64-1.
11. Ibid., p. 17-1-11.
12. Ibid., p. 30-1.
13. Ibid., p. 40-1.
14. CAMCOS, p. 14.
15. AFM 66-1, p. 2-2.
16. CAMCOS, p. iii.
17. AFM 66-276, p. 30-1-14
18. Maintenance Management Information and Control System (MMICS), System Description, 1 August 1969, AF Data Design Center, p. 2-27-99.
19. Ibid., p. 2-100-154.
20. CAMCOS, p. 15.
Lieutenant Colonel Monroe T. Smith (M.S., George Washington University) is Commander, 22d Organizational Maintenance Squadron, March AFB, California. An ex-career NCO, he has served in maintenance organizations in SAC, TAC, and PACAF, including tours as Atlas F missile crew commander and maintenance control officer, Cam Ranh Bay, RVN. He has been a faculty member, Air Command and Staff College. Colonel Smith is a graduate of ACSC (1968) and Air War College (1973).
Disclaimer
The conclusions and opinions expressed in this
document are those of the author cultivated in the freedom of expression,
academic environment of Air University. They do not reflect the official
position of the U.S. Government, Department of Defense, the United States Air
Force or the Air University.
Air
& Space Power Home Page | Feedback? Email the Editor