Foundations of Maintenance and Reliability Engineering

Introduction

Mechanical systems are created to perform useful functions over time: a pump must deliver fluid, a gearbox must transmit torque, and a production line must convert inputs into products safely and economically. Their performance, however, is never permanent. Components wear, materials degrade, controls become obsolete, operating conditions change, and human decisions influence whether equipment remains capable of performing its intended function.

Maintenance and reliability engineering provides the concepts and methods needed to manage this performance throughout an asset’s life. It connects design, procurement, installation, operation, inspection, maintenance, modification, and disposal. The subject is therefore broader than repairing failed equipment. It concerns the systematic control of risk, performance, cost, safety, and resource use.

Reliability engineering primarily asks whether an item will perform its required function for a specified time under specified conditions. Maintenance engineering asks how the item can be restored, preserved, inspected, modified, or improved. Together, they support dependable and sustainable industrial operations.

Learning objectives

After studying this chapter, the reader should be able to:

1. Define reliability, maintainability, availability, dependability, failure, defect, downtime, and asset life cycle.
2. Distinguish related maintenance and reliability terms accurately.
3. Explain the principal objectives and organisational roles of maintenance.
4. Describe how failures produce safety, environmental, operational, financial, and reputational consequences.
5. Explain the relationship between design, operation, maintenance, and life-cycle cost.
6. Apply basic reliability and availability measures to simple engineering situations.
7. Evaluate an equipment problem using a structured maintenance and reliability perspective.

1. Maintenance and reliability as engineering disciplines

Maintenance is the combination of technical, administrative, and managerial actions intended to retain an item in, or restore it to, a state in which it can perform a required function. This definition includes preventive inspection, lubrication, adjustment, replacement, repair, testing, software updates, condition monitoring, and improvement—not only emergency repair.

Reliability is an attribute designed into an asset but demonstrated during operation. A reliable machine is not merely one that has not yet failed; it is one whose probability of performing its required function is acceptably high under stated conditions. Those conditions include load, speed, temperature, environment, maintenance policy, operator behaviour, and mission duration.

Maintenance and reliability are related but not identical. Maintenance can restore a failed pump, but it cannot fully compensate for an unsuitable pump design, excessive process demands, poor installation, or inadequate operator training. Conversely, a well-designed machine may still be unreliable if it is operated beyond its rating or maintained incorrectly.

A useful engineering view is that asset performance results from an interaction:

Design → installation and commissioning → operation → maintenance → modification or replacement

Decisions made early in this sequence constrain later performance. For example, a designer who provides no access for inspection has created a maintainability problem that technicians may never be able to eliminate economically.

2. Fundamental terminology

2.1 Item, asset, function, and performance standard

An item is any part, component, device, subsystem, or system that can be individually considered. An asset is an item that provides value to an organisation. An asset may be physical, such as a compressor, or may include supporting infrastructure and information.

A function is what an asset is required to do. A function should be expressed with sufficient precision. “Operate a pump” is incomplete; “deliver 120 m³/h of water at 45 m head for continuous operation” is more useful. A functional failure occurs when an item is unable to perform a required function to the specified performance standard.

Performance standards should identify capacity, quality, speed, accuracy, containment, safety, availability, and operating conditions where relevant. Without a clear function, it is impossible to define failure consistently.

2.2 Failure

A failure is the termination of an item’s ability to perform a required function. Failure is a state or event relative to a specified requirement. A bearing that produces excessive vibration may have failed if the machine can no longer meet its vibration limit, even though it continues to rotate.

Failures may be classified in several ways:

• Functional failure: the required function is lost or degraded.
• Potential failure: a detectable condition indicates that functional failure is approaching.
• Evident failure: failure is apparent during normal operation.
• Hidden failure: failure is not apparent during normal operation, as may occur in a standby protection system.
• Partial failure: some required performance remains, but below the specified standard.
• Catastrophic failure: sudden failure involving extensive damage, serious hazard, or major loss of function.

Failure cause and failure mode should be distinguished. A failure mode describes how failure occurs, such as shaft fracture, seal leakage, insulation breakdown, or software loss of control. A failure cause explains why it occurred, such as fatigue, contamination, overload, incorrect assembly, corrosion, or inadequate design.

2.3 Defect

A defect is a condition that does not necessarily constitute functional failure but indicates a departure from an acceptable state. Examples include a small oil leak, a loose guard fastener, abnormal temperature, or a crack below a critical size. Defects are important because they may provide an opportunity for planned intervention before failure.

The distinction is practical. A defect is an observed condition; failure is judged against the required function. A defect may be harmless, may require monitoring, or may already represent failure if the applicable performance standard is strict.

2.4 Reliability

Reliability, denoted by R(t), is the probability that an item performs its required function without failure for time t, under stated operating conditions.

Here, T is the random time to failure and P(T > t) is the probability that failure time exceeds t. Reliability is dimensionless and lies between 0 and 1. The expression applies to a defined population or probabilistic model and a clearly specified mission and operating environment.

The complementary probability is the unreliability or cumulative failure probability:

For a constant failure rate λ, a common model is the exponential distribution:

The constant-rate assumption is appropriate only for a limited class of situations, often representing the useful-life portion of a hazard profile. It should not automatically be applied to components dominated by wear-out, infant mortality, or changing operating conditions. Here, λ has units of inverse time, so λt is dimensionless.

The failure rate or hazard rate describes the instantaneous tendency to fail, conditional on survival to time t. In general:

where f(t) is the probability density of failure time. Failure rate is not identical to the proportion of items that fail in a calendar period.

2.5 Maintainability

Maintainability is the probability that a maintenance action can be performed within a stated time, using specified procedures and resources, under stated conditions. It reflects accessibility, modularity, diagnosis, tools, documentation, skills, spare parts, lifting facilities, and task design.

If corrective repair time is represented by a random variable M, maintainability is:

For a constant repair rate μ, the exponential repair-time model gives:

The parameter μ has units of inverse time. This model assumes statistically independent repair times with a constant repair rate and is a simplification. Actual repair time may include diagnosis, waiting, access, isolation, repair, testing, and recommissioning, each with different causes of delay.

2.6 Availability

Availability is the ability of an item to be in a state to perform a required function when required, under given conditions. It incorporates both failure and restoration effects and is therefore influenced by reliability and maintainability.

For a repairable item in a steady state, with mean time between failures (MTBF) and mean time to repair (MTTR), the commonly used inherent or operational approximation is:

MTBF and MTTR must use the same time units. This expression assumes a stable repairable system, statistically representative averages, and a definition of MTTR that is consistent with the availability being calculated. If logistics delays, administrative waiting, or planned maintenance are excluded from MTTR, the result is not total operational availability.

Availability measures should therefore be labelled carefully:

For example, if MTBF is 900 h and MTTR is 6 h:

This does not mean that 99.34% of every short interval will be available, nor does it guarantee production availability. It is a long-run average under the stated assumptions.

2.7 Dependability

Dependability is the ability to deliver service that can justifiably be trusted. It is a broad concept incorporating attributes such as reliability, availability, maintainability, safety, integrity, and security, depending on the system and application.

Reliability asks whether the function continues without failure. Availability asks whether the function is ready when required. Maintainability asks how readily the item can be restored. Dependability combines these concerns with the consequences of incorrect or unsafe service.

2.8 Downtime

Downtime is the time during which an item is unavailable for performing its required function. It may be caused by failure, inspection, preventive maintenance, waiting for materials, lack of personnel, isolation, testing, or external constraints.

Downtime is usefully separated into:

• Active maintenance time: time spent directly diagnosing, repairing, adjusting, or testing.
• Logistic delay: time waiting for tools, people, spares, transport, or access.
• Administrative delay: time awaiting permits, decisions, production release, or scheduling.
• External downtime: time caused by utilities, upstream or downstream equipment, weather, or other systems.

The distinction supports improvement. Reducing hands-on repair time will not solve a problem caused mainly by spare-parts delay.

2.9 Asset life cycle and life-cycle cost

The asset life cycle comprises need identification, concept and design, acquisition, manufacture, installation, commissioning, operation, maintenance, modification, decommissioning, and disposal or renewal. Reliability and maintainability should be addressed from the earliest stages, when design changes are relatively inexpensive.

Life-cycle cost includes more than purchase price. A general representation is:

where C₀ is acquisition cost, C_install is installation and commissioning cost, C_operating is operating cost, C_maintenance is maintenance cost, C_downtime is the cost of lost production or service, C_disposal is end-of-life cost, and V_residual is residual value. The terms must be evaluated over a consistent boundary and, for financial decisions over long periods, may require discounting. The equation is a cost framework rather than a universal accounting rule.

Figure: The asset life cycle and the feedback of operational experience into design and improvement.

3. Objectives and types of maintenance

The principal objectives of maintenance are to:

• preserve required function and production capability;
• protect people, the environment, and property;
• control life-cycle cost and energy consumption;
• achieve required quality and regulatory compliance;
• reduce the probability and consequence of failure;
• restore service safely and promptly;
• improve maintainability, reliability, and operational knowledge.

Corrective maintenance

Corrective maintenance restores an item after a fault or failure has been detected. It may be planned when a defect is accepted temporarily and work is scheduled, or unplanned when emergency intervention is required. Corrective maintenance is not inherently poor practice: run-to-failure can be appropriate for inexpensive, non-critical items whose failure has acceptable consequences and whose replacement is straightforward.

Preventive maintenance

Preventive maintenance is performed at predetermined intervals or according to prescribed criteria to reduce the probability of failure or degradation. Examples include lubrication, calibration, scheduled replacement, cleaning, and inspection. Time-based replacement is effective only when age is meaningfully related to failure risk and when the replacement interval is technically and economically justified.

Predictive and condition-based maintenance

Condition-based maintenance uses information about actual condition to decide whether intervention is needed. Vibration analysis, thermography, oil analysis, ultrasound, motor-current analysis, and performance trending can reveal potential failure. The method requires a detectable degradation process, a suitable measurement technique, sufficient warning time, and competent interpretation.

Improvement maintenance

Improvement maintenance changes the asset, process, procedure, or control system to remove a recurring cause or improve performance. Examples include redesigning a seal arrangement, improving filtration, changing alignment methods, or installing better guarding. Repeated replacement without addressing the underlying cause is usually a weak maintenance strategy.

4. Organisational roles and interfaces

Maintenance performance is a shared organisational responsibility. Typical roles include:

• Design and engineering: define functions, loads, margins, interfaces, maintainability, and protection systems.
• Operations: operate within limits, identify abnormal conditions, provide equipment history, and participate in basic care.
• Maintenance technicians and supervisors: inspect, diagnose, repair, test, document, and improve equipment.
• Reliability engineers: analyse failure data, identify dominant losses, model risk, and coordinate improvement programmes.
• Planning and scheduling: prepare work packages, estimate labour, reserve materials, and coordinate access.
• Stores and procurement: manage critical spares, supplier quality, obsolescence, and lead times.
• Safety, health, and environment personnel: control hazards, permits, isolation, waste, and legal compliance.
• Management: set priorities, allocate resources, define risk tolerance, and evaluate life-cycle performance.

The interface between operations and maintenance is especially important. Operators often detect changes first, while maintainers possess specialised diagnostic and repair skills. A culture that treats maintenance as a separate department’s problem encourages poor information flow and delayed intervention.

Work should be governed by clear authority, competent personnel, controlled procedures, and effective communication. Safety-critical maintenance requires verification that the correct equipment has been isolated, energy has been controlled, the work has been completed correctly, and protective functions have been restored before return to service.

5. Consequences of failure

Failure consequences depend on both probability and severity. A small component can create a major event if it disables a protection system, while a large machine may have modest consequences if a redundant standby is available.

Important categories include:

1. Safety consequences: injury, fatality, fire, explosion, pressure release, entanglement, or loss of protective function.
2. Environmental consequences: emissions, spills, contamination, noise, thermal discharge, or waste.
3. Operational consequences: reduced output, quality loss, unstable process conditions, or interruption of essential service.
4. Financial consequences: repair cost, lost production, energy waste, contractual penalties, and inventory losses.
5. Reputational and social consequences: loss of public confidence, customer dissatisfaction, or disruption to communities.
6. Regulatory consequences: non-compliance, enforcement action, licence restriction, or legal liability.

Risk is often represented conceptually as a combination of likelihood and consequence. However, risk matrices are screening tools, not substitutes for engineering analysis. They may conceal uncertainty, encourage artificial precision, or treat qualitatively different hazards as equivalent. Critical decisions should use appropriate consequence modelling, safeguards analysis, inspection evidence, and professional judgement.

6. Relationship between design, operation, and maintenance

Design influence

Design determines load capacity, material selection, lubrication arrangements, access, modularity, redundancy, monitoring points, isolation provisions, and protection. Design for maintainability includes safe access, standard fasteners, lifting points, replaceable modules, clear diagnostic interfaces, and adequate space for tools and personnel.

Reliability can be improved by reducing stress, controlling contamination, selecting suitable materials, providing protection against overload, and eliminating unnecessary complexity. Redundancy may increase availability, but it also introduces additional components, testing requirements, common-cause vulnerabilities, and maintenance obligations. It is not automatically the best solution.

Operation influence

Operating practices affect degradation. Overload, excessive speed, poor alignment, contaminated lubricants, frequent starts, thermal cycling, unsuitable process chemistry, and bypassed alarms can accelerate failure. Operators therefore need defined operating envelopes and training in abnormal-condition response.

Maintenance influence

Maintenance quality affects both immediate restoration and future reliability. Incorrect torque, contamination introduced during assembly, poor alignment, wrong lubricant, incomplete testing, or inadequate documentation can create repeat failures. Maintenance must therefore be treated as a controlled engineering process rather than simply a manual activity.

Feedback and learning

Failure history should be converted into knowledge. Useful information includes failure mode, cause, detection method, operating state, parts replaced, repair duration, temporary actions, permanent actions, and post-maintenance test results. A database with many records but inconsistent definitions is less valuable than a smaller, accurate dataset.

Common analytical methods include failure modes and effects analysis (FMEA), fault-tree analysis, root-cause analysis, reliability-centred maintenance (RCM), criticality analysis, and trend analysis. These methods should lead to decisions: redesign, condition monitoring, task revision, operating change, spare strategy, or acceptance of risk.

7. Worked application: availability and maintenance delay

A cooling-water pump operates in a service where its average operating time between functional failures is 1,200 h. The average hands-on repair time is 5 h, but the average wait for a seal kit and permit is 19 h.

If only hands-on repair time is included, the apparent availability is:

If total restoration time includes the 19 h delay, then total downtime per failure is 24 h:

The difference is approximately 1.55 percentage points. The engineering conclusion is not necessarily to make technicians repair faster. The dominant opportunity may be spare-parts availability, permit coordination, or isolation planning. This example also demonstrates why availability figures must state their boundary and definitions.

The approximation assumes a stable repairable process and uses average times. It does not capture common-cause failures, planned maintenance, changing demand, standby operation, or uncertainty in the estimates.

8. Illustrative case study: recurring conveyor gearbox failures

Background

The following is a fictional but realistic illustrative case. A mineral-processing plant experiences repeated failures of a conveyor gearbox driving a heavily loaded belt. Each failure causes approximately 14 h of production interruption. The maintenance department replaces bearings and seals several times per year.

Problem

The equipment appears to receive regular preventive maintenance, yet failures continue. Management initially proposes shortening the bearing replacement interval.

Analysis

A cross-functional team reviews work orders, vibration records, installation procedures, operating data, and spare-part history. The team finds:

• failures occur mainly after periods of high belt loading;
• coupling alignment is checked with a low-precision method;
• the gearbox breather is exposed to dust and becomes blocked;
• replacement bearings are stored without adequate contamination control;
• the original design provides limited access for alignment verification;
• work orders record “bearing failed” but not the failure mode or cause.

The bearing replacement interval is therefore not the principal issue. Overload, contamination, alignment error, and weak information quality interact to produce recurring damage.

Approach and solution

The plant establishes a defined operating limit, improves belt loading control, installs a protected breather, introduces clean-storage requirements, and uses a more accurate alignment procedure. A condition-monitoring route is created for vibration and oil analysis. Work-order codes are revised to distinguish fatigue, contamination, misalignment, and overload. A design modification provides better inspection access during the next planned shutdown.

Outcome

In the illustrative scenario, gearbox-related downtime decreases substantially over the following operating period. The improvement does not result from one replacement task; it comes from coordinated changes to design, operation, maintenance practice, spares, and data quality.

Learning points

1. Repeated component replacement does not prove that the replacement interval is correct.
2. Failure mode and failure cause must be recorded separately.
3. Reliability improvement is often cross-functional.
4. Maintainability features can reduce both repair time and maintenance error.
5. Condition monitoring is useful only when measurements lead to timely, technically sound decisions.

9. Challenges, governance, and contemporary practice

Maintenance decisions involve uncertainty. Failure data may be incomplete, operating conditions may change, and apparent trends may reflect reporting practices rather than physical degradation. Engineers should avoid false precision and should state assumptions, data quality, confidence, and decision boundaries.

Ethical practice requires that safety and environmental protection are not traded casually for short-term production. Engineers must report credible hazards, respect competence limits, maintain accurate records, and avoid concealing failures or bypassing safeguards. Regulatory requirements may specify inspection intervals, competency, pressure equipment controls, electrical safety, lifting practices, emissions, or management of hazardous materials. Compliance is a minimum obligation; responsible engineering also considers foreseeable misuse and cumulative risk.

Digital technologies increasingly support maintenance through sensors, wireless monitoring, digital work management, digital twins, and machine-learning models. Artificial intelligence can identify patterns in vibration, temperature, or process data, but it does not remove the need for physics, data validation, domain expertise, and human accountability. Models can generate false alarms, miss novel failures, reproduce biased historical decisions, or become invalid after process changes. A sound implementation defines the decision to be supported, validates sensor quality, protects data and access, monitors model performance, and retains human control over safety-critical actions.

Good practice includes:

• defining functions and performance standards before selecting maintenance tasks;
• ranking assets by consequence and criticality;
• using failure data consistently;
• selecting tasks based on detectable failure behaviour and warning time;
• controlling spares, tools, procedures, and competence;
• measuring both technical outcomes and maintenance-process performance;
• conducting post-maintenance testing and learning from repeat failures;
• reviewing life-cycle cost rather than purchase price alone;
• integrating safety, environmental, cybersecurity, and regulatory requirements.

Chapter Summary

Maintenance and reliability engineering manages the ability of assets to deliver required functions safely, economically, and consistently throughout their life cycle. Reliability is the probability of failure-free performance for a specified time and condition. Maintainability concerns the probability of restoring an item within a specified time. Availability expresses readiness for service and depends on both failure behaviour and restoration performance. Dependability is the broader property of trustworthy service.

Failure is loss of required function, whereas a defect is an abnormal condition that may precede or exist without functional failure. Downtime includes active repair as well as logistical, administrative, and external delays. Asset life-cycle thinking shows that design, operation, maintenance, modification, and disposal are interconnected, and that early decisions strongly influence later cost and risk.

Maintenance may be corrective, preventive, condition-based, or improvement-oriented. Organisational performance depends on cooperation among design, operations, maintenance, reliability, planning, procurement, safety, and management. Failures can have safety, environmental, operational, financial, reputational, and regulatory consequences. Effective engineering therefore combines sound definitions, valid data, appropriate analysis, human competence, ethical governance, and continuous learning.

Key Terms

Reliability: Probability that an item performs its required function without failure for a specified time and condition.

Maintainability: Probability that a maintenance action can restore an item within a specified time and condition.

Availability: Ability of an item to be in a state to perform its required function when required.

Dependability: Ability to deliver service that can justifiably be trusted.

Failure: Termination of an item’s ability to perform a required function.

Functional failure: Inability to meet a defined performance standard.

Potential failure: Detectable condition indicating that functional failure is approaching.

Defect: Abnormal condition that may not yet constitute functional failure.

Failure mode: Manner in which an item fails.

Failure cause: Physical, human, environmental, or organisational reason for a failure.

Downtime: Time during which an item is unavailable for its required function.

MTBF: Mean time between failures for a repairable item under a defined operating context.

MTTR: Mean time to repair, according to a stated boundary and definition.

Corrective maintenance: Maintenance that restores an item after a fault or failure.

Preventive maintenance: Maintenance performed at predetermined intervals or criteria to reduce failure probability or degradation.

Condition-based maintenance: Maintenance selected from evidence of actual equipment condition.

Life-cycle cost: Total cost of acquiring, operating, maintaining, supporting, and disposing of an asset, less residual value where applicable.

Criticality: Relative importance of an asset or failure based on consequences and, where appropriate, likelihood.

Maintainability design: Design that enables safe, efficient inspection, repair, adjustment, and restoration.

Review Questions

Short Answer Questions

1. Define reliability and explain why operating conditions must be stated.
2. Distinguish a defect from a functional failure.
3. What is the difference between active maintenance time and logistic delay?
4. Name four organisational groups that influence asset reliability.
5. Why should life-cycle cost be considered when selecting equipment?

Long Answer Questions

1. Explain the relationship among reliability, maintainability, availability, and dependability. Include the assumptions behind the common availability expression.
2. Compare corrective, preventive, condition-based, and improvement maintenance, giving one suitable application for each.
3. Discuss how design decisions influence maintenance workload and operational reliability.
4. Explain the main categories of failure consequence and how they should influence maintenance priorities.
5. Describe how a maintenance organisation can convert recurring failure information into reliability improvement.

Analytical / Application Questions

1. A machine has an MTBF of 700 h, an active repair time of 4 h, and an average logistics delay of 10 h per failure. Calculate the availability using active repair time alone and then using total restoration time. Interpret the difference.
2. A pump repeatedly experiences seal leakage after maintenance. Propose a structured investigation that separates failure mode, failure cause, maintenance practice, operating condition, and design contribution.
3. An emergency shutdown sensor is rarely demanded and its failure is not obvious during normal operation. Explain why it may represent a hidden failure and propose an appropriate maintenance or testing strategy.

Multiple Choice Questions with Answers

Question 1

Which statement best defines reliability?

A. The probability that an item performs its required function without failure for a specified time and condition

B. The time required to repair an item after failure

C. The percentage of maintenance work completed on schedule

D. The purchase cost divided by operating hours

Correct Answer: A. The probability that an item performs its required function without failure for a specified time and condition

Question 2

A condition that has not yet prevented required function but indicates degradation is best called a:

A. Catastrophic failure

B. Defect

C. Logistic delay

D. Redundancy

Correct Answer: B. Defect

Question 3

For a repairable item with MTBF = 900 h and MTTR = 6 h, the approximate availability is:

A. 98.00%

B. 98.67%

C. 99.34%

D. 99.99%

Correct Answer: C. 99.34%

Question 4

Which factor is most directly associated with maintainability?

A. Ease of access for inspection and repair

B. Maximum product selling price

C. Number of customers served

D. Ambient pressure only

Correct Answer: A. Ease of access for inspection and repair

Question 5

A hidden failure is one that:

A. Occurs only in a new asset

B. Is not apparent during normal operation

C. Always causes catastrophic damage

D. Can never be detected by testing

Correct Answer: B. Is not apparent during normal operation

Question 6

Which maintenance approach uses evidence of actual equipment condition to determine intervention?

A. Condition-based maintenance

B. Unplanned corrective maintenance only

C. Procurement maintenance

D. Administrative maintenance

Correct Answer: A. Condition-based maintenance

Question 7

In the expression R(t) = e^(−λt), the constant λ has units of:

A. Time

B. Time squared

C. Inverse time

D. Dimensionless percentage only

Correct Answer: C. Inverse time

Question 8

Which action most directly addresses a dominant logistic delay?

A. Increasing the equipment operating speed

B. Improving spare-parts availability and permit coordination

C. Removing all condition monitoring

D. Reducing the equipment nameplate rating

Correct Answer: B. Improving spare-parts availability and permit coordination

Question 9

Why is life-cycle thinking important in maintenance engineering?

A. Purchase price is always the largest cost

B. Early design decisions influence later reliability, maintainability, risk, and cost

C. It eliminates the need for maintenance records

D. It guarantees that no component will fail

Correct Answer: B. Early design decisions influence later reliability, maintainability, risk, and cost

Question 10

Which statement about artificial intelligence in maintenance is most appropriate?

A. AI eliminates the need for engineering judgement

B. AI is valid even when sensor data are unreliable

C. AI can support diagnosis but requires validated data, monitoring, and human accountability

D. AI should automatically bypass safety controls when confidence is high

Correct Answer: C. AI can support diagnosis but requires validated data, monitoring, and human accountability