Reliability, Maintenance and Asset Management

Introduction and context

Operations performance depends not only on how quickly or cheaply a process runs, but also on whether its equipment is capable of performing consistently. A production line that frequently stops, a hospital diagnostic device that gives unreliable results, or a water-treatment pump that fails during peak demand creates quality, safety, financial and reputational consequences. Reliability, maintenance and asset management therefore connect engineering decisions with total quality management (TQM).

Reliability concerns the probability that an item performs its required function for a stated period under stated conditions. Maintenance concerns the actions used to preserve or restore that ability. Asset management is the coordinated life-cycle discipline through which an organisation obtains value from physical and intangible assets while managing risk, cost and performance. These ideas are related but not identical: maintenance is one contributor to asset performance, whereas asset management includes investment, operation, renewal and disposal decisions.

The central managerial question is not simply, “How do we repair equipment?” It is, “What level of performance, risk and life-cycle cost is appropriate for the value and criticality of this asset?”

Learning objectives

By the end of this chapter, the learner should be able to:

1. Explain reliability, maintainability, availability and asset-management principles.
2. Interpret common failure patterns and select suitable maintenance policies.
3. Calculate basic reliability, mean time metrics and overall equipment effectiveness.
4. Distinguish corrective, preventive, predictive and proactive maintenance.
5. Explain total productive maintenance (TPM) and the role of operators.
6. Evaluate asset life-cycle decisions using cost, risk, quality and sustainability criteria.
7. Design practical performance measures and governance controls for maintenance systems.

1. Core concepts and terminology

Reliability

Reliability is a conditional concept. A statement such as “the pump is 98% reliable” is incomplete unless the time interval, operating environment, load and definition of failure are specified. If T is the random variable representing time to failure, the reliability function is:

Here, R(t) is reliability at time t, P(T > t) is the probability of surviving beyond t, and F(t) is the cumulative distribution function of failure time. The equation applies to a defined population of similar items operating under specified conditions.

The failure density is f(t), and the hazard rate is:

The hazard rate represents the instantaneous risk of failure at time t, conditional on survival until t. It is not itself a probability and has units of inverse time.

For an exponential failure model, the hazard rate is constant at λ:

This model assumes a constant failure rate and is most appropriate for the useful-life period of some repairable or electronic items. It should not automatically be used for ageing mechanical equipment.

Maintainability and maintenance supportability

Maintainability is the probability that a failed item can be restored to a specified condition within a given time, using prescribed procedures and resources. It is influenced by accessibility, modular design, diagnostic information, skills, tools and spare-parts availability. Supportability extends the view to logistics, documentation, training and infrastructure.

Availability

Availability expresses the proportion of time an asset is capable of performing its required function. It must be defined carefully because calendar time, planned shutdowns, waiting time and performance losses may be treated differently.

For a simple steady-state repairable system with constant failure rate λ and repair rate μ:

where A is steady-state inherent availability, λ is the failure rate and μ is the repair rate. Since mean time between failures (MTBF) is approximately 1/λ and mean time to repair (MTTR) is approximately 1/μ under exponential assumptions:

This is a simplified relationship. Operational availability may also include delays for permits, diagnosis, parts and operators. Thus, an asset can have good inherent availability but poor operational availability.

Asset management

Asset management integrates asset data, risk, performance, cost, people and decisions across the life cycle. ISO 55000-based thinking emphasises value, alignment, leadership, assurance and continual improvement. Value is context-dependent: a low-cost asset may be unacceptable if its failure creates major safety or environmental consequences.

A useful hierarchy is:

2. Failure behaviour and failure analysis

The bathtub curve

The traditional bathtub curve presents three broad regions:

1. Infant mortality: high early failure caused by design defects, manufacturing errors, installation problems or commissioning weaknesses.
2. Useful life: approximately random failures, often associated with external events, overload or hidden defects.
3. Wear-out: increasing failure risk caused by fatigue, corrosion, erosion, ageing, contamination or loss of material properties.

The curve is a useful teaching model, not a universal law. Many modern products have different failure distributions, and the appropriate maintenance policy depends on the actual failure mechanism.

Failure modes, causes and effects

A failure mode describes how an item fails, such as “seal leaks” or “sensor gives a false low reading.” A failure cause is the underlying reason, such as chemical incompatibility, vibration or incorrect calibration. A failure effect describes what happens at the next system level. A consequence describes the organisational result, such as lost production, unsafe release or customer complaint.

Failure Mode and Effects Analysis (FMEA) structures this reasoning. The team identifies functions, functional failures, failure modes, effects, causes, existing controls and recommended actions. Traditional risk priority numbers multiply severity, occurrence and detection ratings, but such multiplication can conceal important differences. A high-severity safety failure should receive attention even if its numerical product is not the highest. Risk matrices and criticality analysis should therefore support, not replace, engineering judgement.

Fault tree analysis (FTA) works deductively from a top event, such as “loss of cooling,” toward combinations of causes using logical AND and OR relationships. FMEA is generally bottom-up; FTA is top-down. Together they help reveal single points of failure, common-cause failures and inadequate safeguards.

Failure-pattern implications

Age-related failures are most effectively managed by age-based replacement or restoration when a clear wear-out mechanism exists. Random failures are usually better managed by condition monitoring, functional checks, redesign or run-to-failure for low-consequence items. Replacing every item at a fixed age is not automatically preventive; it can introduce infant-mortality risk and waste useful life.

3. Maintenance strategies

Corrective maintenance

Corrective maintenance restores an item after a detected failure. It may be planned, when consequences are acceptable and resources are prepared, or unplanned, when failure creates an emergency. Planned corrective maintenance is often rational for inexpensive, non-critical components with simple replacement. Unplanned corrective work generally causes higher disruption, overtime, secondary damage and safety exposure.

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, inspection, calibration, cleaning, replacement and functional testing. Time-based tasks are appropriate when failure probability increases with age and the task is technically effective. Excessive or poorly designed preventive work can create maintenance-induced failures.

Predictive and condition-based maintenance

Predictive maintenance uses information about actual condition to determine when intervention is needed. Condition indicators include vibration, oil particles, thermography, motor current, ultrasound, pressure, temperature and performance trends. Condition-based maintenance is effective when:

• a measurable degradation indicator exists;
• the indicator changes sufficiently before functional failure;
• inspection frequency is adequate;
• an intervention is technically and economically feasible.

A warning interval is the time between detection of a potential failure and functional failure. If inspection intervals exceed this interval, the monitoring programme may not prevent failure. Data quality, sensor placement, alarm limits and competent interpretation are as important as the sensor itself.

Proactive maintenance

Proactive maintenance removes root causes rather than repeatedly treating symptoms. Examples include eliminating contamination, correcting misalignment, improving lubrication practices, redesigning a weak component and controlling operating loads. Root-cause analysis should distinguish immediate cause, contributing factors and systemic causes such as poor specifications, inadequate training or production pressure.

Reliability-centred maintenance

Reliability-centred maintenance (RCM) asks seven practical questions: What are the asset functions? How can it fail to perform them? What causes each failure? What happens when failure occurs? Why does each consequence matter? What task can predict or prevent it? What should be done if no suitable task exists? RCM selects tasks according to failure consequences and technical feasibility rather than tradition.

4. Quantitative measures and worked applications

Mean time measures

MTBF is commonly used for repairable assets as the average operating time between inherent failures. For n observed operating intervals:

MTTR is the average active restoration time:

The definitions must state whether waiting, diagnosis, parts and administrative delays are included. For non-repairable products, mean time to failure (MTTF) is generally more appropriate than MTBF.

Worked example: availability

A packaging machine operates for 1,800 hours during a study and experiences six failures. Total active repair time is 30 hours. Assuming the study records six complete operating intervals and excludes logistical delays:

MTBF = 1,800 ÷ 6 = 300 hours.

MTTR = 30 ÷ 6 = 5 hours.

Estimated inherent availability is:

If the machine also waits 42 hours for parts and authorisation, operational downtime is 72 hours rather than 30 hours. A management dashboard showing only 98.36% could therefore create a misleading impression of service performance.

Series and parallel reliability

For independent components in series, the system works only if every component works:

For two independent components in parallel, the system works if at least one works:

These equations assume clearly defined system success and statistical independence. Common-cause failures, shared power supplies and shared environments can invalidate the simple parallel calculation.

Overall equipment effectiveness

OEE separates three losses:

Availability is normally planned production time minus unplanned and setup losses, divided by planned production time. Performance compares actual output rate with ideal output rate. Quality is good units divided by total units. All three are ratios expressed as decimals or percentages.

Worked example: planned production time is 420 minutes. Downtime is 60 minutes, actual output is 1,600 units, ideal cycle time is 0.20 minutes per unit, and 80 units are rejected.

Availability = (420 − 60) ÷ 420 = 0.8571.

Performance = (0.20 × 1,600) ÷ 360 = 0.8889.

Quality = (1,600 − 80) ÷ 1,600 = 0.9500.

Therefore:

OEE = 0.8571 × 0.8889 × 0.9500 = 0.7238, or approximately 72.38%.

OEE is a diagnostic measure, not a universal target. Increasing OEE by producing unnecessary inventory, bypassing quality checks or postponing essential maintenance is not genuine improvement.

5. Total productive maintenance

TPM is a company-wide approach that seeks maximum equipment effectiveness through employee involvement, preventive systems and continuous improvement. Its principles commonly include focused improvement, autonomous maintenance, planned maintenance, quality maintenance, early equipment management, training and education, safety/environment and administrative TPM.

Autonomous maintenance does not turn operators into unqualified technicians. It gives operators structured responsibility for cleaning, inspection, lubrication where appropriate, abnormality identification and basic restoration. Technicians retain specialist diagnosis, statutory work and complex repairs. The boundary must be explicit and risk-controlled.

A mature TPM system typically uses visual standards, abnormality tags, clean-and-inspect routines, small-group improvement, planned maintenance schedules, skills matrices and loss trees. Leadership must protect time for these activities; otherwise TPM becomes an additional slogan imposed on production staff.

6. Asset life-cycle decisions

Asset decisions begin before purchase. A technically impressive machine may be unsuitable if it has high energy consumption, scarce parts, poor maintainability, incompatible controls or excessive skill requirements. Life-cycle planning includes need definition, concept selection, design, procurement, installation, commissioning, operation, maintenance, renewal and disposal.

A simplified life-cycle cost model is:

C₀ is initial acquisition and installation cost; C_t is operating cost in year t; C_m is maintenance cost; C_e is energy or consumables cost; C_d is expected downtime or consequence cost; C_r is renewal or major overhaul cost; S_t is residual or salvage value; i is the discount rate; and t is the year. All monetary values must use a consistent price basis and time period. The model is a decision aid, not a prediction: uncertainty, non-financial benefits, safety and environmental obligations must also be considered.

Criticality analysis ranks assets according to consequences and likelihood. Consequences may include personnel safety, regulatory compliance, environment, product quality, customer service, production, finance and reputation. Criticality should determine resources, redundancy, inspection depth, spare strategy and escalation rules.

Replacement should not be based only on age. Relevant evidence includes rising failure frequency, declining efficiency, obsolete controls, unacceptable risk, unavailable parts, excessive life-cycle cost and changed demand. A replacement proposal should compare “repair and retain,” “modify,” “replace,” “add redundancy” and “change the process.”

7. Applied case: illustrative food-packaging line

Background

A fictional food manufacturer operates a filling and sealing line. Customer complaints increased because of under-filled packs and intermittent seal defects. The line recorded acceptable OEE on average, but stoppage data were incomplete.

Problem and analysis

The cross-functional team mapped losses and discovered that operators routinely adjusted temperature alarms to avoid nuisance stoppages. Maintenance records described many events as “machine fault,” preventing analysis by failure mode. A Pareto analysis showed that sealing-jaw temperature variation, film-feed misalignment and delayed changeovers caused most quality and availability losses.

FMEA identified seal failure as a high-severity quality failure. The causes included worn bearings, contaminated temperature sensors, incorrect setup and weak alarm governance. A process capability review also showed that the nominal temperature was not stable under changing line speeds.

Approach and solution

The organisation introduced a standard setup sheet, sensor cleaning and calibration intervals based on condition, vibration checks for the jaw drive, operator autonomous inspection and a controlled alarm-change authorisation process. Technicians redesigned the bearing-seal arrangement and installed a trend display for temperature and line speed. The maintenance planner added failure codes and required a cause-and-action field in the work-order system.

The team trained operators to identify abnormal noise, contamination and temperature drift, but prohibited them from defeating safety or quality interlocks. Weekly reviews linked OEE losses with customer complaints and corrective actions.

Outcome

After three months, the illustrative site recorded fewer emergency stoppages, a lower seal-defect rate and more accurate downtime information. The improvement did not result from a single sensor or a larger maintenance budget. It resulted from combining data quality, operator involvement, equipment modification, standard work and management control.

Learning points

1. OEE alone cannot explain a quality problem.
2. Poor failure coding prevents reliability learning.
3. Alarm suppression may conceal process risk.
4. Maintenance and quality ownership must be integrated.
5. The best intervention may be a design or operating change rather than more preventive tasks.

8. Governance, safety and contemporary technology

Maintenance work must comply with applicable legislation, permits, isolation and lockout/tagout requirements, competence rules, pressure-vessel or electrical standards, environmental controls and record-retention obligations. Safety-critical inspections must not be cancelled merely to protect production targets. Digital records should preserve traceability: who performed the task, what was found, what was changed and whether the asset was returned to service safely.

Computerised maintenance management systems (CMMS) support work orders, preventive schedules, inventory, failure codes and history. Enterprise asset-management platforms extend this view into finance, procurement, projects and portfolio decisions. Sensors, industrial internet of things (IIoT) systems and machine-learning models can detect patterns, but prediction is not automatically prevention. Models require representative data, stable operating contexts, validated alerts and human accountability.

Important risks include cyberattack on connected assets, poor data governance, algorithmic false alarms, undocumented model changes and excessive dependence on vendors. AI should support diagnosis and prioritisation while engineers retain responsibility for safety-critical decisions. A model that predicts bearing failure without indicating confidence, data quality and lead time may be operationally unsafe.

9. Implementation principles and limitations

A practical improvement programme should:

• define asset functions and criticality before selecting maintenance tasks;
• establish accurate equipment hierarchies and failure codes;
• remove obvious defects before adding sophisticated monitoring;
• use standard job plans, permits and quality checks;
• measure both leading indicators and outcomes;
• review maintenance-induced failures and overdue work;
• balance planned work with emergency response capacity;
• involve operations, maintenance, engineering, quality, finance and safety;
• test whether a task actually changes failure risk;
• periodically review strategies as equipment, products and risks change.

Common limitations include inaccurate MTBF caused by inconsistent failure definitions, misleading OEE caused by changing planned-time rules, over-maintenance, sensor data without action thresholds, excessive spare-parts inventories and pressure to defer essential work. Metrics must be interpreted in context. A high completion percentage can coexist with poor reliability if the schedule contains ineffective tasks. Likewise, a low OEE may be appropriate for a flexible line designed for frequent small batches.

Chapter Summary

Reliability is the probability that an asset performs its required function for a specified time and under specified conditions. Maintainability concerns restoration, while availability combines failure and repair behaviour. Failure analysis uses tools such as FMEA, FTA, criticality analysis and root-cause analysis to connect technical events with operational consequences. Corrective, preventive, predictive, condition-based and proactive maintenance are selected according to failure mechanisms, consequence, detectability and economic feasibility. RCM provides a structured method for making these choices.

MTBF, MTTR, availability and OEE provide useful quantitative views, but their definitions and data boundaries must be controlled. TPM extends responsibility across the organisation through operator care, planned maintenance, focused improvement, training and safe work. Asset management broadens the perspective to life-cycle value, risk, sustainability, renewal and disposal. CMMS, sensors and AI can improve decisions when supported by reliable data, cybersecurity, competent people and accountable governance. The best maintenance system is not the one with the most tasks; it is the one that delivers required performance and risk control at acceptable life-cycle cost.

Key Terms

Reliability: Probability that an item performs its required function for a stated time and conditions.

Failure: Loss of ability to perform a required function.

Failure mode: Specific manner in which an item fails.

Failure cause: Underlying reason for a failure mode.

Maintainability: Probability of restoring an item within a specified time.

Availability: Proportion of time an asset is capable of performing its function.

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

MTTR: Mean active time required to restore an item.

MTTF: Mean time to failure, generally used for non-repairable items.

Hazard rate: Conditional instantaneous rate of failure.

Preventive maintenance: Planned work intended to reduce failure probability or degradation.

Predictive maintenance: Maintenance based on analysis of actual or forecast condition.

Condition-based maintenance: Maintenance triggered by measured condition evidence.

Proactive maintenance: Action directed at eliminating underlying failure causes.

FMEA: Structured bottom-up analysis of functions, failure modes, effects and controls.

Fault tree analysis: Top-down logical analysis of causes leading to a defined top event.

OEE: Product of availability, performance and quality ratios.

TPM: Company-wide approach to maximising equipment effectiveness through participation.

Life-cycle cost: Total discounted cost of acquiring, operating, maintaining and disposing of an asset.

Criticality: Relative importance of an asset based on consequence and likelihood of failure.

Review Questions

Short Answer Questions

1. Define reliability and explain why time and operating conditions must be specified.
2. Distinguish maintainability from availability.
3. Identify the three regions of the traditional bathtub curve.
4. State the three components of OEE.
5. Explain one responsibility of operators under autonomous maintenance.

Long Answer Questions

1. Compare corrective, preventive, predictive and proactive maintenance, giving an appropriate application for each.
2. Explain how FMEA and fault tree analysis differ and how they can be used together.
3. Discuss why high preventive-maintenance completion does not necessarily indicate high asset reliability.
4. Explain the role of TPM in connecting maintenance, production, quality and employee development.
5. Discuss the factors that should be considered in an asset replacement decision.

Analytical / Application Questions

1. A machine operates for 2,400 hours, has eight failures and requires 48 hours of active repair. Calculate MTBF, MTTR and inherent availability, stating assumptions.
2. A manager proposes replacing every motor annually, regardless of condition. Evaluate this policy using failure-pattern and life-cycle-cost reasoning.
3. Design a maintenance strategy for a critical pump whose failure creates safety consequences, whose vibration can be monitored, and whose replacement requires a 12-hour shutdown. Justify task frequency, escalation and governance.

Multiple Choice Questions with Answers

Question 1

Which statement best defines reliability?

A. The percentage of planned work completed on time

B. The probability that an item performs its required function for a specified time under stated conditions

C. The average time required to repair a failed item

D. The total cost of owning an asset

Correct Answer: B. The probability that an item performs its required function for a specified time under stated conditions

Question 2

For a repairable asset with MTBF of 300 hours and MTTR of 5 hours, what is the estimated inherent availability?

A. 83.33%

B. 95.00%

C. 98.36%

D. 99.50%

Correct Answer: C. 98.36%

Question 3

Which maintenance approach uses measured equipment condition to determine when intervention is required?

A. Predictive or condition-based maintenance

B. Unplanned corrective maintenance

C. Fixed-interval replacement of every component

D. Product inspection only

Correct Answer: A. Predictive or condition-based maintenance

Question 4

In FMEA, a failure mode is best described as:

A. The financial value of an asset

B. The specific manner in which an item fails

C. The time spent waiting for spare parts

D. The discount rate used in life-cycle costing

Correct Answer: B. The specific manner in which an item fails

Question 5

OEE is calculated as:

A. MTBF × MTTR × quality

B. Availability × performance × quality

C. Failure rate ÷ repair rate

D. Output ÷ maintenance expenditure

Correct Answer: B. Availability × performance × quality

Question 6

Which situation most strongly supports planned run-to-failure maintenance?

A. Failure has severe safety consequences

B. The item is inexpensive, non-critical and readily replaceable

C. A failure causes regulatory non-compliance

D. A clear wear-out mechanism requires age-based replacement

Correct Answer: B. The item is inexpensive, non-critical and readily replaceable

Question 7

What is a principal purpose of autonomous maintenance in TPM?

A. To eliminate all specialist maintenance roles

B. To give operators responsibility for basic care and abnormality detection

C. To permit operators to bypass safety interlocks

D. To transfer all maintenance costs to suppliers

Correct Answer: B. To give operators responsibility for basic care and abnormality detection

Question 8

Which cost belongs in a life-cycle cost evaluation?

A. Only the purchase price

B. Only annual maintenance expenditure

C. Acquisition, operation, maintenance, consequence, renewal and disposal-related costs

D. Only the resale value

Correct Answer: C. Acquisition, operation, maintenance, consequence, renewal and disposal-related costs

Question 9

For two independent components arranged in parallel, system reliability is:

A. R₁ × R₂

B. R₁ + R₂

C. 1 − (1 − R₁)(1 − R₂)

D. (1 − R₁)(1 − R₂)

Correct Answer: C. 1 − (1 − R₁)(1 − R₂)

Question 10

Why can an AI-based predictive-maintenance model be unsafe if used without governance?

A. AI models always increase MTTR

B. The model may generate false alarms or miss failures when data and operating contexts are unsuitable

C. Sensors cannot collect any maintenance data

D. Predictive maintenance never requires human judgement

Correct Answer: B. The model may generate false alarms or miss failures when data and operating contexts are unsuitable