A Blueprint for the Material Readiness Index

Over two years ago, an important request for assistance went out to the Atlantic and Pacific Fleet Commanders from the Office of the Chief of Naval Operations and the Chief, Bureau of Ships. The Commander of the Atlantic Fleet and the Commander of the Pacific Fleet called, in turn, on their Type Commanders for assistance. CINCPACFLT summed up the problem succinctly in his letter of 11 September 1961:

During the Pacific Fleet Maintenance Conference held in San Diego on 29 and 30 August, the Chief, Bureau of Ships and the Director of the Ships’ Material Readiness Division, OPNAV, discussed the urgent requirement for an index, or series of indices, which would measure accurately the material condition of ships of the Fleet, the over-all effectiveness of the various ship maintenance programs, and the funding therefor. Such an index is needed to buttress justifications for budget requests for ship maintenance and support programs.

A conservative estimate would be that 150 senior officer-engineers, on 15 major naval staffs, subsequently worked on this problem, and at least 35 of them had advanced degrees.

Yet, the suggestions received by the Office of the CNO proved to be of little value. In most cases, the readiness index was rejected as being impractical to obtain, and the letters devoted themselves to analytical discussions of possible substitutes.

That the desired answer—a blueprint for a readiness index—was not forthcoming is not surprising. A search of the technical literature reveals that this subject, insofar as it relates to ships, has received negligible attention. Nor has any attempt been made to relate maintenance and repair expenditure to material readiness and reliability. Our Merchant Marine, for instance, spends 120 million dollars a year for maintenance and repair¹ without questioning, in the literature at least, whether a 10 per cent spending reduction eventually would reduce reliability 2 per cent or 20 per cent.

Still, despite all of the foregoing, the ability to develop a material readiness index is well within our grasp.

Naval Planners should be able to know at all times the material readiness, or responsive capability, of any and all naval ships. This information is now partially available through the CASREPT System. When a ship suffers a material casualty or derangement to a component which significantly affects the ship’s ability to perform any or all of the missions and tasks, the ship must report to all interested commands. The CASREPT system reports the “go/no go” condition of major installed components, not of the whole ship. The ship’s capability will be reduced significantly, but this does not make the whole ship “no go.” The questions remain. How much is the ship’s total capability reduced as a result of the non-functioning component? And what is the ship’s percentage reduction from total “go”?

The “go/no go” information on the components does not quantitatively describe the reduction in capability which the ship has suffered. In the simple case of a boiler casualty in a destroyer, the ship’s fighting capability is certainly reduced, but by how much? A snap answer might be 25 per cent, since one of four boilers failed. Yet, that boiler would merely have increased the speed about 2 knots. For shore bombardment purposes, the ship’s capability would not be reduced. Conversely, for duty with a fast carrier task force, the ship’s capability would have been reduced an unacceptable amount. For escort duty with an 18-knot convoy, the destroyer’s capability may or may not have been severely affected, depending on the condition of the other boilers on the ship and the probability of encountering a fast nuclear submarine.

It becomes apparent that a knowledge of the readiness or responsive capability of a ship depends on more than just a knowledge of the failed equipment on board ship. It is also necessary to know the resulting loss of capability for each of the various missions, and the chances of failure of the other installed equipment. A material reliability index of a ship which could give this information might be defined as a number which represents:

“The integration of the failure probability, times the mission essentiality for all of the components installed in the ship.”

The ship’s readiness to perform each of the various missions will differ, but the general readiness index would probably use the aggregate of all missions. For example, consider a hypothetical destroyer in perfect condition except for one boiler which has a 50 per cent chance of failure. Assume it has been determined that a boiler failure would reduce by 20 per cent the capability of the destroyer to accomplish the aggregate of all her missions. This destroyer, perfect except for one 50 per cent reliable boiler, is then reduced from the fully ready condition by half of 20 per cent, or 10 per cent, and the readiness index, then, is 90 per cent or just 90. Of course, at present we do not have a method of establishing the chances of failure of each component on each ship. We do not even know quantitatively how much the capability of the ship is reduced by failures of components. That is, we do not know whether the loss of one boiler will reduce the aggregate mission capability 30, 20, or 10 per cent.

If readiness indices are established, it should also be practicable to establish the “minimum index acceptable” for normal operations, for deploying, etc. It might be established, for instance, that a deploying destroyer must have a readiness index of 95 to insure reasonable reliability during deployment. In the case of the hypothetical destroyer with an index of 90 as a result of the marginal boiler, the ship would require some repair attention prior to deployment. If the cost were known for restoring the boiler from its existing 50 per cent reliability to a reliability of 75 per cent, giving a readiness index of 95, the picture would be complete. In general terms, if the readiness index of every ship were known, and the cost of any incremental improvement in this index for each ship were known, both the naval planner and the budget analyst would have a very complete and valuable system.

The task force commander with weekly reports on the readiness indices of the ships of the force could better determine the equivalent numbers of his fully effective ships. He could also better determine the best mission for each ship. During the Cuban crisis of 1962 both the Commanders of Task Forces 135 and 136 must have had difficulty after six weeks of heavy operations in assessing the readiness of the units collectively or individually. Periodic reports on the computed index of each ship for each mission would have been extremely helpful, not only to a task force commander, but also to a type commander. If the crisis had continued, the type commanders would have had to establish rotational schedules based on the restoration requirements of the individual ships. A continuously generated readiness index on each ship would have been invaluable in determining which units most urgently needed restorative assistance first, and which would have been able to hold the line longer while waiting their turn.

The index envisioned in this discussion has, as an inherent part, the man-hours and material requirements for restoring each component to reasonably reliable condition. The readiness index at the start of the Cuban crisis minus the index at the conclusion would give the drop in readiness due to the high usage rate. The cost of restoring the index of each ship to pre-crisis levels would be the wear and tear cost information desired by the budget analyst.

Our naval forces are not always confronted with a hot crisis, but they are continually operating in a climate of “power and presence.” Readiness to react or respond is the power that gives presence its effectiveness. There is no end in sight to this era of power and presence, so we must husband our resources very carefully. We cannot afford to attempt to attain perfect readiness. Therefore, we must be able to measure readiness—unit, group, force, and Navy readiness—with some exactitude. We must determine what degree or per cent of perfect readiness is acceptable and achieve that degree of readiness—not only for the sake of the economy of the country, but also, for the sake of the protection of the country.

The need, then, for a readiness index is valid. Fortunately, the task is no longer as hopelessly unapproachable as it appeared to be two years ago. Agencies of the Navy are now working on various programs which, if properly co-ordinated and combined, can produce the index for all U. S. Navy ships.

The programs which can be combined to produce the readiness index are:

• Planned Maintenance System (PMS)

• Military Essentiality Coding (MEC)

• Maintenance Data Collection System (MDCS)

These three programs must be combined in such a way as to give both a method of establishing the present material condition and of predicting the future condition of each installed component; and a mathematically developed military essentiality number for each component.

The Planned Maintenance System is being gradually introduced into all U. S. naval ships. It is designed to insure comprehensive maintenance of all shipboard equipment. And, perhaps of equal importance, it will insure a high degree of uniform maintenance, reproducible maintenance, among the various ships. This is extremely important in analytical derivations.

Military Essentiality Coding came into being during the preparation of repair part load lists for Polaris submarines.² With limited stowage space, it became necessary to examine critically the mission essentiality of each component and assign numbers designating relative importance. Equipment of less importance in fulfilling the mission had a lower priority demand on repair part stowage space.

The same technique used so successfully in coding essentiality of submarine components was of little value in coding destroyer component essentiality. This can probably be explained by the multitudinous missions assigned to a destroyer and the difficulty of separating and weighing both the aggregate mission and the various individual missions when considering the essentiality of any particular component.

There has now been devised a mathematical technique for coding components known as METRI (Military Essentiality Through Readiness Indices). With this technique all of the various missions are listed and, by logical and simple single steps, the various systems, sub-systems, components, and bits and pieces are interrelated. The essentiality relationships at the subsystem and component level are similar in any ship as, for example, the fuel oil service pump and piping essentiality for a boiler. So sub-system and component relationships need to be derived only once. The result is an essentiality assignment with precise mathematical significance for each part for each mission.

A destroyer in the Atlantic Fleet is being coded at the present time. The job will not be completed for many months, but when it is complete, it will be applicable in toto to a large number of ships. And a large percentage of the sub-systems will be adaptable to a majority of our naval ships, reducing significantly the remaining task of coding the rest of the ships.

The Maintenance Data Collection System is still in the formative stages. The pilot run was introduced into Destroyer Squadron 32 and the destroyer tender USS Sierra (AD-18) on 1 August 1963. By a technique of time accounting at the working level, data will be collected from each ship for use by the maintenance managers at the various levels of the Navy Department. It is expected that this data will be of such detail that the operating and repair history of each component on each ship can be recorded. A mass of carefully selected data has obvious value in maintenance management. The type of reports to be prepared from these data for various levels of management is now being formulated. Although the use of these data for maintenance management is receiving intensive study, no known program exists for studying a method of using these data for readiness evaluation purposes. The recently published results of an analytical study, however, show promise in pointing the way toward the use of these data in attaining the second part of the readiness index—the present and future material condition of every component.

As a result of the CINCLANTFLT letter and the CINCPACFLT letter partially quoted at the beginning of this article, there was initiated in October 1961 a massive data collection program within the Destroyer Force, U. S. Atlantic Fleet. The object was the derivation of some mathematical relationships for a readiness index. Initially, it was hoped that sufficient data existed to derive readiness indices for major machinery components. In July 1962, the study became a joint effort of COMCRUDESLANT and the Institute of Naval Studies.

The study was reoriented owing to insufficient detailed information on the machinery component level. It was decided to try deriving mathematical expressions which would predict the total amount of repairs required to restore a ship to a thoroughly overhauled condition (i.e., the arresting of deterioration and the accomplishment of those repairs required to insure reasonably reliable operation during the overhaul cycle). The published results of the study are beyond the scope of this article except to say that the results were quite successful. In fiscal year 1963, the equations predicted within 2 per cent the 8.5 million dollars total cost of overhauling 16 CRUDESLANT ships in naval shipyards. The ships ranged in size from diesel destroyer escorts to a destroyer leader. The initial overhaul costs of a DLG, DDG, and CLG were considered to include too much debugging to be valid examples, and were not included.

The significance of this successful investigation is that, given enough data, predictive relationships can almost certainly be formulated in a similar manner for the components. At the outset of the investigation, it was felt that the need for repairs was primarily the result of corrosion, erosion, and fatigue which are manifestations of age and usage. This assumed that any improper operation or random failures would not to any significant degree adversely affect the result of the analysis. This original belief was verified subsequently. The programed analysis revealed that the most important variables in predicting the cost of repairing a ship were her age, displacement, time since last overhaul, and the fuel oil burned since last overhaul.

Since the ship as a whole is essentially a mechanical system, the relationship that is true for the whole should be true for the lesser mechanical subsystems, including components. Hence, for mechanical components, if sufficient data exist on the age, relative size, time since last repairs and amount of usage, it should be possible to predict the effort in man-hours and material to restore the component to a fully repaired condition. The amount of repair effort required is indicative of the material condition which exists, and this in turn can be related to failure probability. Electronic parts follow different laws of deterioration, and ordnance equipment usage is not reflected too accurately by the amount of fuel oil a ship has used. In computing the cost of total repairs during overhauls, the poor applicability of the equations to electronic and ordnance equipment does not significantly affect the accuracy of the computations, since these costs are small. With the data which will be available from the collection program, predictive relationships will be obtainable for both electronic and ordnance components.

The data to be collected on each component should include: first the amount of maintenance man-hours and materials used in preventative maintenance, corrective maintenance, and more extensive repairs; and second, the amount of usage between failures (a data collection technique should be used to insure that failures are related to fuel oil burned, hours used, or some other usage index).

The derivation of accurate equations predicting probability of failure and cost of repairs may require several years of experience, but first approximations are obtainable almost immediately from the large data cross sections available from the great numbers of every component in various stages of repair throughout the Navy. As in a fire control computer, the initial solution can be regenerated and refined as additional data are introduced.

The final system would function as follows: For each ship in the Navy, there would be stored in a central location a record of all its components with the derived essentiality for every mission. The condition of each component would be continuously generated as the periodic (probably weekly) experience information is received. The resulting readiness index would be continuously generated, to be reported out both periodically, and when a predetermined low limit had been reached or an alarming deteriorating rate had developed. Concurrently, the estimated manhours and material required to restore the component to a previously determined upper level would also be generated.

This would not be a prohibitive amount of information to be stored and analyzed weekly, or even daily. Our approximately 850 ships, with 200 to 300 components each, could be handled with ease by present-day computers.

No consideration of readiness can ignore the impact of personnel manning and competency. The condition of the individual ship is inextricably tied to the leadership and competence of the personnel. We are talking in terms of probabilities however, when making any type of analytical assessment. An insurance company will ensure the life of a 40- year-old man on the probability that he will live to 75 years of age. The fact that a particular man does not take proper care of his health is naturally of concern, but does not vitiate the premise that statistically he will live to the age predicted. Also, of equal importance, as curative and preventative medicine improves the life span, the continually generating statistical solutions on expectancy will adjust accordingly.

The same is true for the readiness index. Any further Navy-wide reduction of manning levels or any improvement will be picked up in the continually regenerating relationships between usage, age, etc., on the one hand and failure probability on the other hand. Also the competency of the individual crews will follow normal distribution laws so that the probability will be that a particular ship will behave in accordance with the predicted behavior for the majority.

Therefore, although the human factor is of vital importance to readiness, the readiness index will take into account automatically the probable competency which exists; because the index is derived from the average personnel competency which exists throughout the U. S. Navy.

It is appropriate in concluding to quote from CINCPACFLT’s letter of 11 January 1962 which replied to the 1961 CNO request, in part:

It is evident, as stressed by some of the Type Commanders, that any procedure which hopes to arrive at a completely accurate index of the material condition of each ship in the Pacific Fleet would have to be supported by efforts in men, money and time far disproportionate to the value of the index obtained.

Although this article is not proposing a “completely accurate index” a statistical index of useful accuracy can now be obtained without disproportionate efforts. Since the January 1962 CINCPACFLT letter was written, three large programs PMS, METRI, and MDCS, have been initiated for other purposes, which supply most of the “effort” required for obtaining the index.

The only effort still required involves the decision to proceed and the assigning of the project manager to combine these three programs into the index—the index which is needed more every day (to quote CINCPACFLT’s initial letter again) to “buttress justifications for budget requests for ship maintenance and support programs.”