09_Reliability_Centered_Maintenance_Handbook_(S9081-AB-GIB-010).pdf

S9081-AB-GIB-010

REVISION 1

Reliability-Centered

Maintenance (RCM)

Handbook

This manual supersedes S9081-AB-GIB-010, dated October 1983, which shall be destroyed.

DISTRIBUTION STATEMENT A: APPROVED FOR PUBLIC RELEASE.

DISTRIBUTION IS UNLIMITED.

PUBLISHED BY DIRECTION OF COMMANDER, NAVAL SEA SYSTEMS COMMAND

0910-LP-106-0318 18 April 2007 iii

FORWARD

If you require assistance, contact distance support (anchor desk) via the web http://www.anchordesk.navy.mil/) and via the toll free number

(1-877-4-1-TOUCH [86824]).

Ships, training activities, supply points, depots, Naval Shipyards and Supervisors of Shipbuilding are requested to arrange for the maximum practical use and evaluation of NAVSEA technical manuals. All errors, omissions, discrepancies, and suggestions for improvement to NAVSEA technical manuals shall be forwarded to:

COMMANDER

CODE 310 TMDER, BLDG 1388

NAVSURFWARCENDIV NSDSA

4363 MISSILE WAY

PORT HUENEME CA 93043-4307

on NAVSEA Technical Manual Deficiency/Evaluation Report, form NAVSEA 4160/1. All feedback comments shall be thoroughly investigated and originators will be advised of action resulting there from. Three copies of NAVSEA 4160/1 are included at the end of each separately bound technical manual 8-1/2 x 11 inches or larger. Copies of form NAVSEA 4160/1 may be downloaded from:

https://nsdsa2.phdnswc.navy.mil/tmmp/forms/TMDER_BLANK_REV_3-2001.doc

Users are encouraged to transmit deficiency submittals via the Naval Systems Data Support Activity Web page located at:

https://nsdsa2.phdnswc.navy.mil/tmder/tmder.htm

Urgent priority TM deficiencies shall be reported by Naval message to Port Hueneme Division, Naval Surface Warfare Center (Code 311), Port Hueneme, CA. Local message handling procedures shall be used. The message shall identify each TM deficiency by TM identification number and title.

iv

—THIS PAGE INTENTIONALLY LEFT BLANK—

v

Reliability-Centered Maintenance (RCM) Handbook

Contents

Section Title Page

1 Introduction to RCM 1-1

1.1 Purpose 1-1

1.2 History of RCM 1-1

1.3 Basis for RCM 1-3

1.4 RCM Certification 1-4

1.5 Summary 1-5

2 Fundamentals of Maintenance Engineering (ME)

2-1

2.1 Failures Happen 2-1

2.2 Not All Failures Have the Same

Probability

2-2

2.3 Not All Failures have the Same

Consequences

2-4

2.4 Simple Components Wear-Out, Complex

Systems Break Down

2-8

2.5

Good maintenance provides required functionality for lowest practicable cost

2-8

2.6 Maintenance Can Only Achieve Inherent

Design Reliability

2-9

2.7 Hidden Functions Require Special

Treatment

2-9 vi

2.8 Unnecessary Maintenance Takes

Resources Away From Necessary

Maintenance

2-10

2.9 Good Maintenance Programs Undergo

Continuous Improvement

2-11

2.10 Summary 2-13

3 What Comprises Maintenance? 3-1

3.1 Definition of Maintenance 3-1

3.2 Three Categories of Maintenance 3-2

3.3 Five Types of Preventive Maintenance

Tasks

3-4

3.4 Summary 3-10

4 The Rules of RCM 4-1

4.1 Three Hypotheses 4-1

4.2 All Maintenance Tasks Must Be

Applicable (Relevant)

4-3

4.3 All Maintenance Tasks Must Be Effective

(Have Value)

4-4

4.4 Why We Need RCM Methodology 4-5

4.5 Summary 4-5

5 Two RCM Processes 5-1

5.1 Summary 5-3

vii

6 The Classic RCM Process 6-1

6.1 Analyst Selection and Training 6-1

6.2 Information Collection 6-2

6.3 System Partitioning and Identification 6-2

6.4 Systems Analysis 6-3

6.5 The RCM Methodology 6-3

6.6 Designing an RCM-Based Preventive

Maintenance Program (MIL-P-24534A)

6-4

6.7 Phase 1: Functional Block Diagram and

the Master Systems and Subsystems Index

(OPNAV Form 4790/114)

6-5

6.8 Phase 2: Functional Failure Analysis

(OPNAV Form 4790/116)

6-7

6.9 Phase 3: Additional Functionally

Significant Item Selection (OPNAV Form

4790/117)

6-8

6.10 Phase 3 Continued: Functionally

Significant Items Index (OPNAV Form

4790/118)

6-9

6.11 Phase 4: Failure Modes and Effects

Analysis (OPNAV Form 4790/119)

6-9

6.12 Phase 5: Decision Logic Tree Analysis

(OPNAV Form 4790/120)

6-10

6.13 Phase 6: Servicing and Lubrication

Analysis (OPNAV Form 4790/121)

6-16 viii

6.14 Periodicity Considerations 6-16

6.15 Phase 7: Audit and Preparation of the

Maintenance Requirement: Maintenance Requirement Index (MR) (OPNAV Form

4790/123)

6-16

6.16 Phase 8: Method Study and Procedure

Evaluation for the New Tasks and Revised Maintenance Requirement Cards (MRCs)

(OPNAV Form 4790/130)

6-16

6.17 Phase 9: Maintenance Requirement Task

Definition (OPNAV Form 4790/124)

6-17

6.18 Phase 10: Inactive Equipment

Maintenance (OPNAV Form 4790/129)

6-17

6.19 Phase 11: Unscheduled Maintenance 6-18

6.20 Phase 12: MRC/MIP Preparation

(OPNAV Forms 4790/82, 83, 84, and 85)

6-18

6.21 eRCM 6-18

6.22 Summary 6-19

7 The Backfit Process 7-1

7.1 Analyst Selection and Training 7-1

7.2 Information Collection 7-1

7.3 The Backfit RCM Methodology 7-2

ix

7.4 Conducting a Backfit RCM Analysis of an

Existing PMS Program

7-2

7.5 Finalizing PMS Task Revisions 7-7

7.6 Summary 7-8

Appendix A Glossary A-1

Appendix B MIL-P-24534A (NAVY) Data Forms B-1

Appendix C NAVSEA/SPAWAR Technical Manual Deficiency/Evaluation Report (TMDER)

C-1 x

1-1

Section 1—Introduction to RCM

1.1 Purpose

This handbook provides Navy maintenance practitioners a reference document for applying the principles of Reliability-Centered Maintenance (RCM) to the evaluation of both new and well-established ship planned maintenance system (PMS) preventive maintenance requirements. It is intended as a supplement to the Naval Sea Systems Command Reliability-Centered Maintenance-based certification program for those who develop, modify, review and authorize Planned Maintenance System tasks for Navy ships, systems, and equipment. The handbook introduces basic principles of maintenance and illustrates how these principles establish rules for good maintenance tasks. It describes how these rules are used to develop maintenance requirements and associated documentation for new systems and equipment as well as to evaluate and improve the quality and effectiveness of well-established maintenance programs.

Requirements for analysis of preventive maintenance requirements and for development of associated support documentation are defined in MIL-P-24534A, “Planned Maintenance System: development of Maintenance Requirement Cards, Maintenance Index Pages, and Associated Documentation.”

1.2 History of RCM

Logical methods for designing preventive maintenance programs started about a half century ago. This brief history sets the stage for detailed discussion of the Reliability-Centered Maintenance methodology mandated by the Chief of Naval Operations for ship maintenance. OPNAVINST 4700.7(series) “Maintenance Policy for U.S. Navy Ships,” and OPNAVINST 4790.4(series) Ships’ Maintenance and Material Management (3-M) System Policy require the use of RCM for the development of maintenance programs and associated maintenance tasks.

Efforts to look deeply into the effectiveness of preventive maintenance as a process for avoiding failure began in the late 1950s. Those involved in establishing preventive maintenance requirements before then may have been so convinced of the value of their actions that they saw no need to prove its truth.

They reacted almost entirely to each event as it occurred rather than to generalize their experience.

The introduction of jet aircraft fleets led the airline industry to apply a growing expertise in the process of analysis to improving the effectiveness of preventive maintenance for transport aircraft. Since the underlying reason for preventive maintenance is the belief that reliability of hardware decreases with use, the first efforts examined the relationship between reliability and age (use). This was done by application of techniques already used by life insurance actuaries.

In 1967, the airline industry’s Maintenance Steering Group (MSG) first applied decision tree logic – a series of questions that lead to a supportable maintenance task decision – to the problem of identifying required preventive maintenance tasks. This predecessor of RCM proved an efficient approach, since it focused directly on the impact of unreliability on operations and safety. The following year, decision tree logic formed the basis for design of the initial Boeing 747 maintenance program. In the ensuing

1-2 years, the airline industry updated the MSG approach and applied it to such aircraft as the Boeing 767 and beyond.

In the early 1970s, this work attracted the attention of the Office of the Secretary of Defense, the Naval Air Systems Command, the Air Force, and the Army. The Navy was the first to apply this new philosophy and an improved methodology called Reliability-Centered Maintenance (RCM) to both new design and in-service aircraft.

In 1978, United Airlines published a book prepared under contract to the Office of Assistant Secretary of Defense (Manpower, Reserve Affairs and Logistics) detailing the RCM methodology. This book, Reliability-Centered Maintenance by Stanley Nowlan and Howard Heap and the work leading to its publication formed the basis for the Naval Sea Systems Command application of RCM to ship maintenance.

The prototype application of RCM to surface ship preventive maintenance was installed in USS ROARK (FF-1053) in 1978. Following evaluation of results on four additional FF-1052 class ships, the Chief of Naval Operations directed that RCM be used for development of scheduled maintenance tasks for all new and in-service naval ships in accordance with MIL-P-24534A, Planned Maintenance System.

This application is now called “Classic RCM.”

The principles of RCM were also applied to ship depot level maintenance in the Phased Maintenance Program (PMP) developed initially for Combat Support Ships in the early 1980s. This program, which was later expanded to all auxiliary ships, amphibious ships and certain combatant ships, served as the catalyst for Continuous Maintenance (CM). CM is a nearly continuous process of identifying, screening, authorizing, planning, and accomplishing maintenance at all maintenance levels (organizational, intermediate, and depot). The underlying policy for both PMP and CM is called Condition-Based Maintenance (CBM), which is the CNO’s policy for ship maintenance as called out in OPNAVINST 4700.7(series). CBM is defined as maintenance based on objective evidence of need.

OPNAVINST 4790.16, “Condition-Based Maintenance Policy,” requires use of RCM to determine evidence needed to select appropriate maintenance.

Benefits derived from the use of RCM to establish maintenance programs eroded over the years because the Navy tended to overlook an important part of the RCM process – periodic analysis of operational and maintenance feedback to continuously improve the periodicity and scope of prescribed maintenance tasks. This became critical in the 1990s, when construction of new ships fell prey to the cost of maintaining in-service ships.

Recognizing that adopting change based on operating experience is a major component and benefit of RCM, the Naval Sea Systems Command developed an RCM methodology to be used to improve existing preventive maintenance programs in 1996. A prototype of the so-called RCM ‘Backfit’ Methodology was applied to the USS YORKTOWN (CG-48) Planned Maintenance System (PMS) as part of the Smart Ship Program. Results included a 46.7 percent reduction in ship’s force PMS workload without adverse impact on safety, mission, or the environment. This approach has been refined and is applied to all shipboard systems by Navy In-Service Engineers (ISEs) in conjunction with ship’s force personnel experienced in operation and maintenance of the systems evaluated. Both the ISEs and involved ship’s force personnel are provided training in Backfit RCM prior to conducting their analyses of existing PMS tasks during an evaluation called Ship Maintenance Effectiveness Review

(SHIPMER).

1-3

This handbook addresses both the ‘Classic’ RCM approach for development of new maintenance requirements for systems and equipment and the RCM ‘Backfit’ approach for validation of maintenance requirements for in-service systems and equipment.

1.3 Basis for RCM

RCM is derived from careful consideration of basic questions such as the following:

• What functions does the system perform?

• What functional failures might occur?

• Which of the functional failures are likely to occur?

• Are the functional failures evident to the operating crew?

• What are the consequences of failure on safety, mission, and cost?

• What is the relative risk of failure in terms of probability of failure and severity of failure?

• What, if anything, can be done to prevent likely failures?

• What is the cost of trying to prevent failures?

RCM is reliability-centered. Its objective is to maintain the inherent reliability of the system or equipment design, recognizing that changes in inherent reliability may be achieved only through design changes. Within this constraint, RCM functions to determine what preventive maintenance is required to:

• Ensure safety of personnel, protect the environment and equipment; and

• Provide reasonable assurance of being able to accomplish the ship’s mission at a cost less than that of correcting the failure the preventive task is trying to prevent.

Classic RCM develops maintenance tasks for new systems by:

• Partitioning the ship into systems and subsystems that require analysis;

• Identifying functionally significant items;

• Determining any needed maintenance requirements (tasks) for each significant item based on analysis of its functions (both evident and hidden), its dominant failure modes, and management of risk associated with functional failure;

• Determining when, how, and by whom each task should be accomplished; and

• Identifying any need for design change:

– When safety is threatened by a failure for which there is no applicable and effective preventive task, or

– When inherent reliability proves to be less than adequate.

Backfit RCM validates existing maintenance tasks by using information from operations and maintenance in a streamlined analysis, and adjusting task intervals and task content where appropriate.

The identification and application of fundamental maintenance concepts (e.g., risk management) provides the maintenance practitioner a foundation for future application.

1-4

1.4 RCM Certification

The Chief of Naval Operations (CNO) has mandated that Condition-Based Maintenance (CBM) practices be implemented in all Navy maintenance decisions involving ships, aircraft, and infrastructure in OPNAVINST 4790.16. The instruction tasks NAVSEA to “provide procedures and training for the implementation of CBM…." The objective of condition based maintenance is that maintenance is performed based on objective evidence of the need. Reliability Centered Maintenance (RCM) is the process that is used to develop the maintenance tasks needed to implement CBM. To achieve the CNO’s CBM goal, NAVSEA 04RM established the RCM Certification Program. The Naval Sea Systems Command certifies all those who develop, review, or approve scheduled maintenance requirements to ensure that RCM principles and approved methodology are properly employed in maintenance plan development for Organizational, Intermediate, and Depot (OI&D) levels. RCM certification is required for all PMS practitioners in accordance with Appendix J of NAVSEAINST 4790.8(series), Ship’s Maintenance and Material Management (3-M) Manual. NAVSEA 04RM manages the RCM Certification Program and associated training.

Each equipment system placed in a U.S. Navy ship has an experienced shore-based In-Service Engineer (ISE) responsible for monitoring lifecycle as well as individual problems in that system. The RCM Certification Program affects all individuals who either develop new PMS, or make changes to existing PMS. Those persons, whether they are Navy employees or contractors must achieve and maintain the appropriate level of RCM certification relevant to the work they are performing.

ISEs are challenged daily with design and maintenance issues. NAVSEA’s RCM Certification Program ensures that they are equipped with the skills required to develop, review or approve changes to PMS.

RCM Certification comprises three certification levels.

Level I - Backfit RCM for Practitioners: Level I certification is required for ISEs, Commodity Specialists or contractors who modify existing Planned Maintenance System (PMS) maintenance tasks.

The course is two days in duration and taught by an approved SEA 04RM instructor. This course provides training in the fundamental CBM topics and methodologies to be employed when performing an engineering review of existing PMS maintenance tasks. Upon completing certification, individuals apply this methodology to every maintenance related assessment they perform. Level I certification is valid for three years and can be renewed by attending another Level I training class, a Ship Maintenance Effectiveness Review (SHIPMER) session, or by satisfactorily completing an online recertification examination.

Level II – Classic RCM for PMS Requirement Developers: Level II certification applies to ISEs, Commodity Specialists or contractors who develop or approve maintenance requirements for new systems or equipment. The course begins with an introduction to maintenance engineering and RCM fundamentals and leads to detailed instruction in the MIL-P-24534A (Navy) process for developing maintenance task requirements for Navy equipment. The student is guided through multiple practical

1-5 application examples to further illustrate the RCM process and to prepare for the certification examination. The course is five days in duration and taught by an approved NAVSEA 04RM instructor.

Level II certification is valid for three years and can be renewed by attending another Level II training class or by satisfactorily completing an online recertification examination.

Level III – Navy Backfit RCM Trainer: Level III certification applies to a very small number of ISEs who have been selected to teach the Level I certification course within their command or activity.

Individuals selected for Level III training must be certified in both Level I and Level II and demonstrate a high level of proficiency in RCM as outlined in MIL-P-24534A (Navy). The Level III course is composed of two major elements and is two weeks in duration. The first week is devoted to an intensive review of RCM theory and application including a comprehensive examination of RCM fundamentals and instructor oriented training in the RCM Level I curriculum. The second week is an introductory course on adult training techniques and includes opportunity to practice the presentation of the RCM Level I course material in a controlled environment. Level III certification requires the demonstration of instruction proficiency of the RCM Level I course material and is valid for one year. Renewal of certification is contingent upon successful demonstration of continued proficiency to NAVSEA 04RM or designated technical warrant holder.

1.5 Summary

In 1967 the airline industry’s Maintenance Steering Group (MSG) applied the first decision logic tree for development of preventive maintenance. This work grew out of a comprehensive review of maintenance practices begun in the late 1950s. By the 1970s, a systematic approach for the development of maintenance had become known as “Reliability-Centered Maintenance” (RCM). The fundamental goals of RCM were to maintain system functionality by ensuring all maintenance actions were designed to maximize system reliability at minimum cost. To accomplish this, the RCM process is a structured approach that requires the analyst to justify maintenance requirements by answering a series of questions.

• What functions does the system perform?

• What functional failures might occur?

• Which of the functional failures are likely to occur?

• Are the functional failures evident to the operating crew?

• What are the consequences of failure on safety, mission, and cost?

• What is the relative risk of failure in terms of probability of failure and severity of failure?

• What, if anything, can be done to prevent likely failures?

• What is the cost of trying to prevent failures?

The US Navy began applying RCM to surface ship maintenance in 1978. Within a few decades RCM has become a fundamental part of the Navy’s Conditioned Based Maintenance (CBM) Policy. To ensure RCM principles are understood and applied across all maintenance decisions during a ship’s entire life cycle, Naval Sea Systems Command has instituted a certification program. All personnel associated with the development, review, or approval of scheduled maintenance requirements are required to undergo training in RCM fundamentals as they apply to existing maintenance (Level I) or new maintenance (Level II).

1-6

2-1

Section 2―Fundamentals of Maintenance Engineering (ME)

This section briefly explains nine fundamental maintenance concepts that govern the development, implementation, execution, and continuous improvement of ship maintenance programs. These are core principles of RCM. These nine fundamental concepts are:

• Failures happen.

• Not all failures have the same probability

• Not all failures have the same consequences

• Simple components wear out, complex systems break down

• Good maintenance provides required functionality for lowest practicable cost

• Maintenance can only achieve inherent design reliability

• Hidden functions require special treatment

• Unnecessary maintenance takes resources away from necessary maintenance

• Good maintenance programs undergo continuous improvement.

These concepts form a progression of maintenance engineering thought from the most fundamental concept ⎯ failures happen ⎯ to the greatest challenge for any maintenance program: establishment of a process for continuous improvement.

2.1 Failures Happen

It is appropriate to begin discussion with the fact that functional failures -- unsatisfactory conditions in which intended functions are not adequately provided -- happen.

Maintenance planners know intuitively, if not by personal experience, that unexpected failures happen.

Nevertheless, they tend to approach the business of maintenance program planning from one of two very different directions. One direction is to attempt to eliminate all failures. The second direction is risk management. There are many factors in equipment reliability that can cause a failure. They include stress, corrosion and fatigue. A maintenance program can be used to prevent or preclude such failures.

However, there are failures caused by things beyond the control of the maintainer. These include unplanned or random events (sudden high-energy impact, lightning strike, etc.), sometimes called acts of God. There is nothing that can be done to prevent these failures, especially from an economic standpoint. There are failures caused by poor manufacturing quality. These types of failures cannot be prevented by maintenance.

We know that not all failures can be prevented. Good maintenance programs don’t try to prevent all failures; they couldn’t if they tried! Good maintenance programs will minimize the number of failures, however, perhaps even to the point where they appear to have been totally eliminated.

The paradigm of attempting to prevent all failures has lead to extremely costly maintenance programs that do not succeed because all failures cannot be prevented. The important thing is to focus resources on failures that can be prevented.

2-2

2.2 Not All Failures Have the Same Probability

Reliability engineers define reliability as the probability that equipment will provide its desired function over a specified period of time. When maintainers speak of probability of failure, they really mean the rate at which a piece of equipment will fail over a period of time, i.e. the frequency, or rate, of failure.

Things to consider about probability of failure are: Is that probability, or rate, of failure a function of time? Is it increasing in frequency? (a bad thing), Is it decreasing in frequency? (a good thing), or is it constant. A constant failure rate is often called random failures. It is often said that complex equipment experiences random failures. So how do real-world equipment failure rates compare to our preconceived notions about probability of failure?

The pioneering work on Reliability-Centered Maintenance performed by United Airlines began with a study of hardware age-reliability characteristics. When United disassembled engines for repair, it found some engines whose internal components were in good condition and some engines whose internal components were in poor condition. Further, when United began to explore the age-reliability relationship by introducing these engines back into service after minimal repair, it found that some engines with internal components in good condition exhibited lower reliability in service than engines with components in worn condition. United realized that it did not have as firm a grip on the relationship between time in service (or “age”) and reliability as it thought. Reliability degradation is a reduction of an item’s resistance to failure.

United Airlines applied the same technique used by insurance actuaries when developing mortality curves for human beings. This technique was used to explore the relationship between the failure-rate or conditional probability of failure and some measure of operating age for aircraft hardware. The conditional probability of failure is the failure rate of the component with respect to time. A constant conditional probability of failure means that the failure rate is not increasing or decreasing with respect to time, it means that we have random failures.

The results of United’s initial exploration of this relationship are shown in Figure 2.1 in column marked UAL, which were developed from a study of 139 aircraft components and equipment. These results show both wear out (age-related failures, curves A, B, and C) and random failures (curves D, E, and F).

The figure also includes similar results from later studies.

2-3

(Figure 2.1.) Age-Reliability Characteristics Curves.

Analysis showed that:

1. Random failures predominate compared to age-related failures. Infant mortality (i.e., high initial probability of failure, decreasing with age) is common.

2. Infant mortality persists (significant time is required to transition out of infant mortality to steady state conditional probability of failure).

3. The conditional probability of failure is never zero.

4. The shape of the age-reliability characteristic curve is highly configuration-dependent.

5. Simple items tend to exhibit wear out (curves A, B and C) whereas complex items tend to exhibit random failures (curves D, E and F).

Curves D, E, and F should not be interpreted to mean that some items never degrade or wear out. These curves simply show the life of some items came to an end before wear out was evident, perhaps because they were removed for restoration, or were replaced with modified or upgraded items. Everything will eventually degrade with time, but some items degrade so slowly that wear out is not a concern since the degradation will not adversely affect performance during the life of the ship. Degradation of the glass in the bridge windows, for example, does not trouble the maintenance planner, who needs to focus on items whose degradation is of real concern.

Age-reliability characteristics curves can tell a maintenance planner whether or not wear out (increase in the conditional probability of failure) exists. If there is no evidence of wear out, there

Age

No Evidence of Wearout

Evidence of Wearout

A ge

-R el at ed

R an do m

A. 4% 3% 3% 6%

B. 2% 1% 17%

C. 5% 4% 3%

D. 7% 11% 6%

E. 14% 15% 42% 60%

F. 68% 66% 29% 33%

89% 92% 77% 93%

UAL Broberg MSDP Studies SSMD 1973 1983 1993Pf

Age

No Evidence of Wearout

Evidence of Wearout

A ge

-R el at ed

R an do m

A. 4% 3% 3% 6%

B. 2% 1% 17%

C. 5% 4% 3%

D. 7% 11% 6%

E. 14% 15% 42% 60%

F. 68% 66% 29% 33%

89% 92% 77% 93%

UAL Broberg MSDP Studies SSMD 1973 1983 1993Pf

2-4 is no good basis for a time-directed life-renewal task. It makes no sense to spend maintenance resources to renew the life of an item whose reliability has not degraded or which may have actually improved with age.

It is not necessary to develop age-reliability characteristics curves for every item of equipment. In fact, these curves should be limited to a small set of high-value items where the investment cost of curve development will be returned in reduced life-cycle maintenance requirements. Maintenance planners need to keep these curves ⎯ and the lessons they teach ⎯ in mind as they review maintenance for the items for which they are responsible.

Rapid increase in the conditional probability of failure of an item may be associated with a component, subassembly, or assembly within an item ⎯ or it may be typical of the complete item. The more complex an item, the less likely it is that all its constituent elements will have the same age-to-failure characteristics.

2.3 Not All Failures have the Same Consequences

Maintenance is an investment, an indirect cost that organizations are willing to bear with the expectation they will receive a benefit in the form of sustained personnel safety and system reliability that exceeds the magnitude of their investment. The consequence of failure of an item is a very important consideration when considering appropriate levels of maintenance. In RCM we refer to the consequence as severity of failure.

We saw in Section 2.2 that the conditional probability of failure is never zero. Because there is always a chance of a component or a system failing, the severity of that failure is very important. It is also important not merely to focus on what happens to a component when it fails, but what happens at higher levels. Severity should be considered at the highest level, i.e. what happens to the ship when the component fails? A good example is a bearing in a ship’s fire pump. If the bearing fails, we know that we will need a replacement. An immediate reaction would be to consider this a critical failure, but we must look at the overall effects. How many fire pumps are onboard and how many are required for system pressure? i.e. How much redundancy is there? Is there a cross connect from the seawater service system? It is important for the analyst to consider the overall effect to the ship, to see the big picture.

The consequence of any failure may be categorized as impacting Personnel, Safety, Ship Mission and All Others. Obviously, the highest level of severity is a Sailor being injured or killed; this is a Safety related failure. The next level down in importance is the Mission level. Will the failure affect the mission of the ship? If the failure does not affect safety or mission, than it falls into the “all others” category and preventing the failure is primarily an economic decision; the cost of doing the preventive maintenance must be less than the cost of correcting the failure that the maintenance is trying to prevent.

It is important to recognize that not every failure results in death or injury; not every failure will impact the ship’s ability to perform its mission. As a result, we must always make a realistic assessment of each failure’s consequences. The probability and the severity of a failure factor into the concept of risk.

Everyone involved with maintenance should understand the fundamental concepts of risk in order to deal with it objectively and effectively. Maintenance practitioners need to know what risk is, how it can be assessed, and how it can be managed as part of a sound maintenance program.

2-5

Most people asked to define “risk” identify it as a threat, a hazard, or a very undesirable situation. They clearly associate “risk” with adverse circumstances. This may be correct, but it presents only one aspect of risk.

Risk is composed of two factors: the probability of failure (Pf) of an item and the severity of failure consequences (Sf). Thus, from a mathematical perspective:

This definition of risk is very important. The natural tendency is to focus on the severity of failure, but not all potentially adverse failures are likely to occur.

The probability of a failure is at least as important a consideration in the objective assessment of its risk as is failure severity. The design of a well-balanced and effective maintenance program requires that both these factors be considered. Neglecting Sf could result in overlooking needed coverage for a failure mode that could injure or kill a member of the operating crew, damage the environment, or cause loss of a critical mission capability. Neglecting Pf could result in the unnecessary expenditure of resources to “prevent” a failure that may never occur.

A well-designed maintenance program uses the concept of risk to allocate resources where they will provide the greatest benefit. It assesses the risk of failure that confronts individual ship systems and equipment, and it allocates resources to prevent failure on the basis of that risk assessment. It recognizes the “opportunity cost” of each maintenance decision: unnecessary maintenance that reduces resources available to accomplish needed work in other areas that must then be deferred for lack of resources.

Now that we know what “risk” is, we need some way to evaluate or assess it. We can begin by looking at dominant failures modes.

Dominant failure modes occur frequently or have serious consequences

There are two specific considerations used to identify dominant failure modes:

• First, the failure mode should appear relatively often in material condition analyses. It should constitute a relatively large proportion of the total number of failures, or it should appear in significantly degraded condition in a large proportion of diagnostic tests or inspections. The sample size must be adequate to establish confidence in the results.

• Second, a failure that does actually occur should be considered dominant if it has serious consequences, regardless of frequency of occurrence. Serious consequences could be such things as personnel injury or death, loss of mission capability, or expensive repairs.

A 19th century Italian economist named Pareto once stated that 20 percent of Italy’s population controlled 80% of the nation’s wealth. This became known as the Pareto Principle, and has many applications, including maintenance. As shown in Figure 2.2, dominant failure modes account for the

Risk ≡ Pf x Sf

2-6 majority of failures. Those dominant failure modes that don’t fit the description of frequent occurrence are failures that have such severe consequences that they must be considered for preventive action despite their low probability of occurrence. The curve of Figure 2.2 also applies to the application of maintenance resources. Twenty percent of systems or equipment consumes about 80 percent of maintenance resources.

100 200 300 400 500 600

Number of Different Equipment

Pe rc en t o f R es ou rc es

700 900800

20% of Failure Modes Cause About 80% of Failures; in addition, 80% of Maintenance Resources Are Applied to About 20% of Equipment.

(Figure 2.2)

Note: This graph displays theoretical data derived from the Pareto Principle.

Failure modes that create personnel safety hazards dominate consideration by the severity of their potential consequences. All failures that involve safety must be considered no matter how remote the probability of failure: no one puts a price on the life or limb of the operating crew.

Failure modes whose effects are severe with respect to such things as mission or expense should also be considered, accounting for severity as a risk factor.

Failure modes that are most likely to occur dominate the population of failures by their frequency of occurrence. This consideration accounts for probability of failure as a risk factor, as it focuses on failure modes of most likely concern. The maintenance analyst doesn’t need statistically valid figures for Pf accurate to two or three decimal places to proceed with the analysis. Navy maintenance data collection

2-7 and analysis systems have not yet reached the point where discrete, widely accepted values for system and equipment Pf are available. Fortunately, broad or general values for Pf can and should be used until more detailed figures are developed. Even the use of such broad values of Pf permits comparing the relative risk of different work items where tradeoff decisions involving funding must be made.

Next, we consider the second element of risk assessment: severity of failure.

The Classic RCM analysis process evaluates the severity of failure at three different levels.

The first step is the component level, which is an evaluation of failure severity at the local (or equipment) level. The principal concern at this point is whether the failure mode will cause injury or death to operating personnel in the vicinity at the time of failure. A secondary concern at the local level is whether the failure will cause complete loss of the equipment, i.e., require rebuild or replacement of the equipment or simply replacement of a component. The answer has implications for material procurement and corrective effort.

The next step, the subsystem/system level, is to examine severity of failure within the system boundaries. The concern here is loss of system functionality, including safety issues related to loss of function. Some systems have redundant subsystems; loss of one subsystem has no immediate system impact, thereby reducing the severity of failure. The steering system is one example of this, as it has both port and starboard subsystems. Loss of one side is compensated for by operation of the other.

The final step examines severity of failure outside the bounds of the system. The concern here is loss of mission capability. Ship systems are the basic building blocks of ship mission capabilities; loss of a ship system can mean loss or significant degradation of a mission capability. The effect of failure on ship mission capability can be highly varied. Loss of the electronic countermeasures system may jeopardize the ship in a combat situation, but loss of the potable drinking water system will put the ship out of business in any situation.

So failures have different consequences: safety or regulatory, mission, and economic. The maintenance analyst and the maintenance planner must evaluate these consequences to determine the exact nature of severity of failure as it relates to risk.

Both risk factors, Pf and Sf, are considered in the decision to either conduct a risk management analysis or to eliminate it as a source of concern. If in doubt, the default decision is to carry the functional failure and the failure mode forward for analysis. The entire purpose of risk assessment is to decide whether to eliminate failures of little significance from further preventive maintenance analysis.

We recognize that it simply is not possible to prevent all failures. Further, it is not even worthwhile to prevent all failures. It is possible to reduce the number of failures to an acceptable level and then to manage the risk assessment of that small number. That is what good maintenance is all about.

Excessive maintenance that attempts to prevent all failures does not fit the definition of “good maintenance.”

2-8

2.4 Simple Components Wear Out, Complex Systems Break Down

This was mentioned briefly in the previous discussion on age-reliability curves, but we will now discuss this concept in greater detail. A “simple” component is something that has relatively few failure modes.

Some examples are the timing belt in an automobile, the roller bearing on a drive shaft, the cable on a crane. Simple items often exhibit some particular sign of distress before they fail (i.e., they provide evidence of potential failure). For example, a radiator hose in your car most often fails by rupturing. A soft or bulged hose is an indication that failure is near. For such failures, a knowledgeable planner can often design a task to detect the evidence of potential failure and take corrective action prior to failure.

When we say “wear out,” we mean that there is a marked increase in the conditional probability of failure. As can be seen in Figure 2.1, the conditional probability of failure for curves A and B show this.

In each case, the conditional probability of failure has a “wear out” age. If the maintenance planner knows the typical wear out age for a component, it is easy to develop a schedule to replace the component before failure.

Complex items tend to breakdown through random failure. Complex items contain many simple components, each having its own failure modes. Because there are so many components in a complex item, no single failure mode tends to dominate. Because complex items have a large variety of failure modes, they typically do not exhibit a wear out age. Their failures do not tend to be a function of age.

Failures of complex systems most often occur randomly. Their conditional probability of failure is generally a constant. These constant conditional probability of failure equipments are represented by curves E and F in Figure 2.1.

These differences between simple and complex components are important with respect to maintenance strategies. Because we do not have a wear out age for complex items, the practice of performing time-driven overhauls on such items is wasteful of resources. One spends money fixing things that do not need to be fixed! Only when we can show that an item exhibits a wear out age does performing an overhaul or replacement of that component at a particular time or age make sense.

2.5 Good maintenance provides required functionality for lowest practicable cost

We previously defined maintenance as action to ensure that items provide their intended functions when required. Good maintenance, therefore, serves to maintain or restore function.

Our definition of maintenance is helpful but does not help the analyst or manager to differentiate good maintenance from other maintenance. To do that, the analyst needs to evaluate the cost of performing preventive maintenance against the cost and other consequences of not performing the maintenance. A good maintenance program delivers required system, subsystem, and equipment functions for least cost, i.e., least total expenditure of resources.

The statement that “good” maintenance delivers “required” functionality means that it delivers functions essential for satisfactory operation when they are needed. Good maintenance satisfies actual, but not necessarily theoretical, demands.

The essential nature of good maintenance is to ensure that the hardware does what is required of it when required, not that it be wholly capable of achieving its full design capability at all times.

2-9

Hardware may be capable of doing much more than what is required of it. Maintenance efforts to ensure that hardware produces greater capability than needed are a waste of resources. We do not maintain our automobiles so they are capable of running at their maximum rated speed every day; we maintain them so they will deliver normal driving speeds. Cost effective maintenance serves to preserve the functions required by the user.

2.6 Maintenance Can Only Achieve Inherent Design Reliability

A very important point to consider about maintenance and design capability is that maintenance can only restore or preserve the hardware’s inherent design reliability and performance characteristics. If the design’s inherent reliability or performance is poor, doing more maintenance will not help. To improve poor reliability or performance attributable to inadequate design, one must change the design.

Systems and equipment do not always have to “look like the blueprint” or meet design specifications to deliver required functionality. There is no need to invest maintenance resources to attain greater performance than needed. The design engineer typically incorporates various margins in the design specifications. These margins provide protection from such conditions as corrosion and erosion in the operating environment and abnormal levels of stress from performance outside prescribed operating ranges. They also take into account dimensional variations from manufacturing processes, differences in the material composition of constituent parts that comprise the complete item, and other factors. In addition, functional performance parameters, such as pump pressure and throughput, required by procurement specifications or available in commercial off the shelf (COTS) equipment may be greater than required for adequate system performance. The shipbuilder may select an available 250 gpm pump, for example, rather than procure a new pump design to provide 225 gpm required by design specifications.

2.7 Hidden Functions Require Special Treatment

We must now ask the question, what is a hidden function, and why does it require special treatment?

Some organizations perform mostly corrective maintenance, i.e. fix-when-fail, opting not to take equipment offline to perform preventive maintenance. But what happens when you have a failure that is not visible or evident to the operating crew? For example, an emergency device like a fire sprinkler system, alarm system, or an overspeed shutdown device may have failed. How will you know whether those devices have failed until such time as their function is required?

The answer is that you will not know about a hidden functional failure until it is too late, unless you perform a procedure to test for that hidden failure.

A maintenance task to find a hidden failure, a failure finding task, does not actually prevent that failure.

It seeks to find a failure that has already happened that you do not know about. The rest is troubleshooting and corrective maintenance. The goal is to find that failure before it becomes a larger problem. For example, you don’t want to find out your smoke detector has failed by discovering you have an out of control fire in your home. You don’t want to find out that your boiler safety relief valve has failed by the fact that your boiler has just ruptured. You do the failure finding test to help ensure the availability of those protective or safety devices when they are needed.

2-10

What failure finding tasks do prevent are multiple or cascading failures. Cascading failures have a very small probability of occurring because they depend upon several events, each with its own probability happening at the same time. For example, even though they have a small probability of failure, some cascading failures are worth trying to prevent because their consequences are so severe. Most people will agree it is good to inspect for hidden failures that involve safety or mission consequences.

However, it must be understood that the very act of testing some safety or protective devices causes undue wear on them and may result in premature failure. For maintenance associated with hidden failures, it is important to consider not only the consequence of failure, but the consequence of over-testing.

2.8 Unnecessary maintenance takes resources away from necessary maintenance

There is almost always more maintenance to do than resources to do it with. If one wastes resources doing unnecessary maintenance, there won’t be enough available to perform the truly necessary maintenance. Inadequate investment in maintenance has adverse consequences. It is not difficult to understand that deferring or eliminating valid maintenance will increase an item’s probability of failure.

On the other hand, excessive investment in maintenance also has adverse consequences. There are several ways this takes place.

First, many maintenance actions require systems or equipment to be off-line when maintenance is performed. Every minute an item is off-line for maintenance is a minute not available for operations.

This time off-line penalizes the ship with decreased operational availability (Ao).

Second, some maintenance actions induce “infant mortality” [i.e., increased conditional probability of failure after maintenance] as a result of such factors as human error, the use of defective material, or errors in technical documentation. If the maintenance were not mandated, there would not be the exposure to infant mortality and no penalty in decreased operational availability would result.

Third, excessive maintenance requirements may result in maintenance accomplishment that is less thorough than required because maintainers don't have the time necessary to accomplish all tasks properly. This can also occur when maintenance personnel don’t believe in the value of the tasks they are performing and perform tasks in a superficial manner. The odds of performing maintenance properly increase when maintainers are convinced of its intrinsic value.

Fourth, an adverse mismatch between maintenance task workload and maintenance manpower available may present the work center supervisor aboard ship a difficult choice:

• Perform every task as best as possible with the manpower available, taking shortcuts where necessary but “accomplishing” every task;

• Perform some tasks thoroughly with no shortcuts and defer the remainder; or

• Some combination of the two.

Thus, too much maintenance is not necessarily a better situation than not enough maintenance. The most desirable maintenance program matches the investment of resources to the priority of requirements as closely as possible. This approach provides the most effective use of maintenance resources to achieve necessary operational availability for any level of maintenance resource allocation.

2-11

2.9 Good Maintenance Programs undergo Continuous Improvement

The most effective maintenance programs are dynamic: they are changing and improving regularly to make ever better use of resources the longer they are in operation. They regularly test and explore the boundaries of maintenance, obtaining increasing knowledge of the items being maintained as they proceed, while accepting unexpected failures as the price of progress.

The Navy in peacetime should be a laboratory for the Navy at war. The Navy in peacetime should learn all it can about system and equipment failure and the effectiveness of maintenance tasks so that these lessons are available for use in war.