Maintenance, Reliability and Asset Management

Introduction and context

Production systems create value only when people, materials, information and equipment are available at the required time and performance level. Maintenance is therefore not merely a repair activity. It is a coordinated management process that preserves or restores the function of physical assets while controlling safety, quality, cost and environmental risk.

In modern production management, maintenance decisions influence capacity, delivery reliability, inventory, labour productivity, product quality and return on assets. A poorly maintained machine may cause an immediate breakdown, but it can also create hidden losses through reduced speed, excessive scrap, energy waste and unstable processes. Effective maintenance integrates engineering knowledge with planning, scheduling, risk analysis, performance measurement and organisational behaviour.

Reliability describes the probability that an item performs its required function for a specified time under stated conditions. Availability describes the proportion of time an asset is capable of performing its function. Asset management extends these ideas across the entire life cycle: acquisition, operation, maintenance, modification and disposal.

Learning objectives

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

• distinguish corrective, preventive, predictive and condition-based maintenance;
• select maintenance policies according to failure behaviour, consequence and cost;
• calculate reliability, failure rate, maintainability and availability measures;
• apply Total Productive Maintenance (TPM) and Overall Equipment Effectiveness (OEE);
• use risk-based methods to prioritise maintenance resources;
• develop practical maintenance plans, work orders and performance indicators;
• evaluate technology, human, ethical and governance issues in asset management.

1. Essential terminology

An asset is an item, system or resource that provides value to an organisation. A production asset may be a machine, production line, building, utility system, software-controlled device or associated infrastructure.

An item is any component, subsystem or system considered for analysis. Function means the performance required from an item, including output, quality, speed, accuracy and safety conditions. A failure is the termination of the ability of an item to perform a required function. A fault is an abnormal condition that may lead to failure but does not necessarily mean that the required function has already stopped.

Downtime is the period during which an asset is unavailable for production. It may include waiting for diagnosis, parts, labour, approval or access, not only the time spent physically repairing the asset. Mean Time Between Failures (MTBF) is commonly used for repairable equipment, whereas Mean Time To Failure (MTTF) is generally used for non-repairable items. Mean Time To Repair (MTTR) represents average active repair time; organisations should define whether logistics and waiting time are included.

2. Maintenance strategies

2.1 Corrective maintenance

Corrective maintenance restores an asset after a fault or failure has been detected. It may be:

• run-to-failure, where the asset is deliberately operated until failure;
• deferred corrective maintenance, where a detected defect is safely postponed;
• immediate corrective maintenance, where repair is required promptly because continued operation is unsafe or damaging.

Run-to-failure is reasonable for inexpensive, non-critical and easily replaceable items whose failure has little effect on safety, quality or production. It is inappropriate for assets whose failure can injure people, contaminate products, damage other equipment or cause major interruption.

2.2 Preventive maintenance

Preventive maintenance is planned maintenance performed before functional failure, usually at fixed time, usage or production intervals. Examples include lubrication every 500 operating hours, belt replacement after a specified number of cycles and inspection at the start of each week.

Its advantages are planning convenience, reduced uncertainty and better control of critical deterioration. Its limitations include unnecessary replacement, maintenance-induced defects and the possibility that a time-based interval does not match the actual failure mechanism. Preventive replacement is effective when failure probability increases predictably with age. It is less effective for random failures.

2.3 Predictive and condition-based maintenance

Predictive maintenance uses measurements and analysis to estimate the developing condition of an asset and determine when intervention is needed. Condition-based maintenance (CBM) is the broader policy of making maintenance decisions from observed condition rather than solely from calendar time. Predictive maintenance is commonly treated as an advanced form of CBM because it attempts to forecast future condition or remaining useful life.

Typical techniques include:

CBM requires reliable sensors, appropriate alarm limits, competent interpretation and a response process. A measurement that produces no maintenance decision is not useful merely because it is technically sophisticated.

2.4 Selecting a maintenance policy

Selection should consider failure consequences, failure predictability, warning time, repair cost, spare-part lead time, safety, environmental exposure, quality impact and the economics of intervention. A practical decision sequence is:

1. Define the asset's required functions and performance standards.
2. Identify functional failures and their physical causes.
3. Determine the consequences of each failure.
4. Assess whether degradation is detectable before functional failure.
5. Compare preventive, condition-based, corrective and redesign options.
6. Specify the task, interval, trigger, resources and escalation rule.
7. Review results using actual failure and maintenance data.

Reliability-Centred Maintenance (RCM) formalises this logic. It does not mean that every asset receives more maintenance; it means that each task is justified by a failure consequence and an effective failure-management strategy.

3. Reliability and failure behaviour

3.1 Reliability function

Let T be the random time to failure and let t be a specified operating time. Reliability is:

where R(t) is reliability at time t, and P(T > t) is the probability that the item survives beyond t. Reliability is dimensionless and lies between 0 and 1. The expression assumes a clearly defined population, operating condition and failure criterion.

The cumulative distribution function is:

F(t) is the probability of failure by time t. If a probability density exists, f(t) is related to reliability by:

3.2 Hazard or failure rate

The instantaneous hazard rate is:

It represents the conditional rate of failure at time t, given survival to that time. It is not simply the percentage of all units that fail at that instant. Its unit is inverse time, such as failures per operating hour.

The classic “bathtub curve” describes three broad regions: early failures caused by manufacturing or installation defects, a relatively stable useful-life period with approximately random failures, and wear-out failures whose rate increases with age. The curve is a useful conceptual model, not a universal law. Many electronic components exhibit approximately random failures, while some mechanical components exhibit wear-out behaviour.

For a constant failure rate λ, the exponential model applies:

The model assumes independent operating exposure and a constant hazard rate. Its mean time to failure is:

For a repairable asset operating under stable conditions, MTBF is often estimated as:

This estimate is meaningful only when operating time, failure definitions and the observation period are consistently recorded. MTBF is not a guarantee that each failure will occur after the average interval.

3.3 Reliability of systems

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

where R_i is the reliability of component i. For two independent components in parallel, the system works if at least one works:

These formulas assume correctly defined system success, statistical independence and no common-cause failure. Shared power, environment, software or maintenance errors can invalidate the simple parallel calculation.

Worked example 1: exponential reliability

A pump has an estimated constant failure rate of 0.0002 failures per operating hour. Estimate its reliability over 1,000 operating hours.

The estimated probability of surviving 1,000 hours is approximately 81.9%, and the probability of failure by that time is approximately 18.1%. The result is valid only if the constant-rate assumption is reasonable.

4. Maintainability and availability

Maintainability is the probability that a failed item can be restored to a specified condition within a given time under stated procedures and resources. For an exponentially distributed active repair time with repair rate μ:

where M(t) is maintainability by repair time t, and μ has units of repairs per unit time. Under this assumption:

Availability has several forms. Inherent availability focuses on active corrective maintenance and excludes logistics and administrative delays:

Operational availability includes realistic downtime such as waiting for parts, technicians and approvals:

The denominator and data definitions must be consistent. Using a short repair-time definition with a long downtime denominator can create a misleading performance measure.

Worked example 2: availability

A packaging machine operates for 7,200 hours during a period and experiences 12 failures. Total active repair time is 96 hours. Estimate MTBF and inherent availability.

Here MTTR = 96/12 = 8 hours. If an additional 144 hours were spent waiting for parts and access, operational availability would be:

The difference shows why stores, planning and maintenance coordination affect asset availability even when technical repair skill is unchanged.

5. Total Productive Maintenance and OEE

Total Productive Maintenance is an organisation-wide approach that seeks maximum equipment effectiveness throughout the asset life cycle. It involves operators, maintenance personnel, engineers, supervisors, managers and support functions. Common TPM pillars include autonomous maintenance, planned maintenance, focused improvement, quality maintenance, early equipment management, training, safety and office TPM.

Autonomous maintenance gives operators responsibility for basic cleaning, inspection, lubrication, abnormality detection and standard care. It does not transfer specialist work without training; rather, it creates early detection and shared ownership.

Overall Equipment Effectiveness

OEE decomposes productive effectiveness into availability, performance and quality:

Using the usual production definitions:

All three factors are ratios. The numerator and denominator must use compatible time and unit definitions. OEE should not be used to compare unrelated processes without considering product mix, planned stops, bottleneck position and data quality.

Worked example 3: OEE

A line has 480 minutes of planned production time. Downtime is 60 minutes. It produces 800 units, including 32 defective units. The ideal cycle time is 0.5 minutes per unit.

The loss is not explained by downtime alone: speed loss and defects also reduce effective output.

6. Risk-based asset management

Risk-based asset management directs resources toward the combination of likelihood and consequence that matters most. A simple qualitative model is:

This multiplication is appropriate only when the scales and interpretation are defined. A typical five-level risk matrix may score likelihood from 1 to 5 and consequence from 1 to 5, producing a score from 1 to 25. The score is a prioritisation aid, not a precise probability of loss.

A stronger analysis may estimate expected annual loss:

For safety-critical assets, a low-frequency event with catastrophic consequences may require controls even if its numerical score is lower than a frequent minor stoppage. Regulatory requirements and the hierarchy of controls override simplistic cost ranking.

A practical criticality assessment considers:

• personnel safety and legal compliance;
• environmental release or community impact;
• product quality and customer risk;
• production loss and recovery time;
• damage propagation and redundancy;
• spare-part availability and supplier dependence;
• detectability and warning time;
• maintainability and access.

Failure Mode and Effects Analysis (FMEA) identifies failure modes, effects, causes, existing controls and recommended actions. A traditional risk priority number may multiply severity, occurrence and detection rankings, but rankings are ordinal and multiplication can hide important differences. Teams should therefore discuss the individual ratings rather than relying only on the composite number.

7. Maintenance planning and control

A professional maintenance system connects work identification to feedback. The normal workflow is:

1. detect a defect or receive a preventive plan;
2. screen safety and production consequences;
3. create a clear work request;
4. approve and prioritise the work;
5. plan labour, permits, tools, drawings, parts and access;
6. schedule the job with operations;
7. execute using safe work procedures;
8. test and return the asset to service;
9. close the work order with failure, cause, remedy, time and materials;
10. analyse the result and update the maintenance strategy.

A work order should identify the asset, functional location, symptom, priority, required isolation, estimated labour, craft, parts, risk controls, completion evidence and failure code. Backlog should be measured by age, labour hours, criticality and overdue risk, not merely by the number of open work orders.

Useful indicators include schedule compliance, planned versus reactive work, emergency work percentage, repeat failures, preventive maintenance compliance, stock-outs, wrench time, MTBF, MTTR, availability, OEE and maintenance cost per unit. No single indicator is sufficient. For example, very high preventive maintenance compliance may coexist with poor reliability if tasks are ineffective or failure coding is weak.

8. Illustrative case: beverage filling line

Background

A beverage plant operates a filling line containing a conveyor system, filler, capper, labeller and packaging machine. The line is a bottleneck. Operators report frequent short stops, while maintenance records show relatively few major breakdowns.

Problem

Monthly production data show availability of 91%, performance of 88% and quality of 98%. The calculated OEE is:

The plant initially proposes replacing the filler, but the evidence does not establish that the filler is the dominant loss source.

Analysis

A loss tree separates planned stops, breakdowns, setup, minor stops, speed loss and defects. Time-stamped observations reveal that 47% of unplanned loss consists of stops shorter than five minutes at the capper. Operators also report inconsistent cap torque. Vibration measurements identify increasing bearing energy, and inspection finds poor lubrication practice and a misaligned guide.

The team conducts an FMEA. The capper is highly critical because its failure interrupts the bottleneck and can create quality defects. However, the selected controls are relatively inexpensive: standard lubrication, alignment, torque verification, operator inspection and a condition alarm.

Approach and solution

The plant implements the following actions:

• a centreline and cleaning standard for the capper;
• operator training in abnormality tagging and basic inspection;
• scheduled lubrication using the correct quantity and lubricant;
• vibration checks at an appropriate interval;
• a torque trend with defined escalation limits;
• a planned replacement of the bearing during the next controlled shutdown;
• revised failure codes and daily review of minor stops.

Outcome

After two months, availability increases to 95%, performance to 93% and quality to 99%. OEE becomes:

The improvement is not attributed to one technology. It results from combining operator involvement, condition information, basic equipment care, planned intervention and better data.

Learning points

The case demonstrates that major replacement is not always the best first response. Maintenance strategy should address the actual loss mechanism, distinguish short stops from major failures, and combine technical and organisational controls. OEE indicates where losses occur; it does not by itself identify the engineering cause.

9. Technology, data and contemporary practice

Computerised Maintenance Management Systems (CMMS) and Enterprise Asset Management systems provide asset registers, preventive schedules, work orders, inventory links, history and cost information. Sensors and Industrial Internet of Things platforms can stream vibration, temperature, current, pressure and energy data. Analytics may detect deviations, classify failure patterns or estimate remaining useful life.

Artificial intelligence can support anomaly detection, image-based inspection, natural-language work-order analysis and spare-parts forecasting. It does not eliminate the need for engineering judgement. Models require representative data, stable sensor configuration, clear failure labels and controls against false alarms. A prediction that cannot be acted upon may increase workload without improving reliability.

Digital systems also create cybersecurity and governance risks. Connected assets should use access control, network segmentation, patch management, backup, change control and incident response. Data ownership, retention, auditability and vendor access should be specified contractually. Automated recommendations affecting safety must remain subject to competent human approval.

10. Safety, ethics and governance

Maintenance work involves stored energy, moving machinery, electrical hazards, pressure, chemicals, height and confined spaces. Lockout/tagout or equivalent isolation procedures, permit-to-work systems, guarding, verification of zero energy, appropriate personal protective equipment and post-maintenance testing are essential. Production pressure never justifies bypassing a safety control.

Ethical asset management requires truthful failure reporting. Suppressing bad news, manipulating downtime definitions or closing work orders without evidence produces misleading performance and increases risk. Decisions should balance cost with worker safety, environmental protection, product integrity and community obligations.

Organisations should maintain an asset hierarchy, standard failure codes, competence requirements, inspection records, calibration controls, audit trails and management-of-change procedures. Regulatory obligations may apply to pressure systems, lifting equipment, electrical installations, environmental emissions and safety-critical controls. Local legislation and recognised technical standards must be consulted for specific applications.

11. Limitations and best practices

Reliability measures are sensitive to definitions and data quality. A rising MTBF may result from fewer recorded failures rather than genuine improvement. OEE can be distorted by changing ideal cycle times or excluding inconvenient downtime. Risk matrices may create false precision. Predictive models may fail when operating conditions change. Preventive maintenance can introduce errors if tasks are poorly designed.

Best practice is to:

• define functions, boundaries and failure criteria clearly;
• use a hierarchy of maintenance strategies rather than one universal policy;
• combine operator knowledge, engineering analysis and historical data;
• prioritise safety and high-consequence failures;
• design maintenance for access, modularity and error prevention;
• measure both leading indicators and outcomes;
• review tasks after actual failure evidence;
• control spare parts according to criticality and lead time;
• train and authorise personnel appropriately;
• audit data quality and close the learning loop.

Chapter Summary

Maintenance is a production-management system that preserves asset function, not simply a repair department. Corrective maintenance is appropriate when consequences are low or failure is the chosen strategy; preventive maintenance uses planned intervals; predictive and condition-based maintenance use measured deterioration and analysis. Reliability is the probability of survival to a specified time, while availability depends on both failure frequency and restoration time. Constant-rate exponential models are useful only under stated assumptions.

System reliability depends on configuration and component behaviour. Maintainability concerns restoration within a specified time. TPM creates shared ownership and continuous improvement, while OEE decomposes effectiveness into availability, performance and quality. Risk-based asset management prioritises safety, environmental, quality and production consequences, using tools such as RCM, FMEA and criticality assessment. Effective maintenance requires planning, scheduling, competent execution, accurate work-order data, safe governance and continual review. Digital and AI technologies can improve detection and decision support, but they require trustworthy data, cybersecurity and human accountability.

Key Terms

Asset: A physical or information-based item that provides organisational value.

Availability: The proportion of time an asset is capable of performing its required function.

Condition-based maintenance: Maintenance triggered by observed asset condition.

Corrective maintenance: Maintenance performed to restore function after a defect or failure.

Failure: Loss of ability to perform a required function.

Failure rate: Conditional rate at which failures occur, usually expressed per unit time.

FMEA: A structured analysis of failure modes, causes, effects and controls.

Inherent availability: Availability based mainly on active repair time and failure intervals.

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

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

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

MTTR: Mean time to repair or restore an item, according to the defined data convention.

OEE: Product of availability, performance and quality.

Predictive maintenance: Condition-based maintenance that uses analysis to anticipate future degradation or intervention need.

Preventive maintenance: Planned maintenance performed before functional failure.

Reliability: Probability that an item performs its required function for a specified time.

Reliability-Centred Maintenance: A structured process for selecting effective failure-management tasks.

Risk: Combination of likelihood and consequence of an unwanted event.

Total Productive Maintenance: Organisation-wide approach to maximising equipment effectiveness and participation.

Work order: Controlled instruction and record for maintenance work.

Review Questions

Short Answer Questions

1. Distinguish preventive maintenance from condition-based maintenance.
2. What is the difference between MTBF and MTTR?
3. State the three components of OEE.
4. Why can a parallel system reliability calculation be misleading when common-cause failures exist?
5. Identify four factors used in asset criticality assessment.

Long Answer Questions

1. Explain how an organisation should select between corrective, preventive, predictive and condition-based maintenance.
2. Discuss the assumptions and practical limitations of the exponential reliability model.
3. Explain how TPM changes the roles of operators, maintainers and managers.
4. Describe a complete maintenance work-management process from defect identification to feedback.
5. Evaluate the benefits and governance risks of using AI-enabled predictive maintenance.

Analytical / Application Questions

1. A machine operates for 4,000 hours, fails 10 times and requires 75 hours of active repair. Calculate MTBF, MTTR and inherent availability. State the assumptions behind your calculation.
2. A production line has availability of 0.90, performance of 0.92 and quality of 0.97. Calculate OEE and identify which factor should be investigated first if the largest improvement opportunity is required.
3. A safety-critical pump has infrequent failures but a long spare-part lead time and severe environmental consequences. Recommend a maintenance and asset-management strategy, explaining why run-to-failure would be unsuitable.

Multiple Choice Questions with Answers

Question 1

Which maintenance strategy is triggered primarily by measured equipment condition?

A. Run-to-failure maintenance

B. Condition-based maintenance

C. Corrective maintenance after breakdown

D. Calendar-only replacement

Correct Answer: B. Condition-based maintenance

Question 2

If the constant failure rate is λ, which expression represents exponential reliability at time t?

A. R(t) = λt

B. R(t) = 1 − e^(−λt)

C. R(t) = e^(−λt)

D. R(t) = t/λ

Correct Answer: C. R(t) = e^(−λt)

Question 3

A repairable asset operates for 2,000 hours and fails five times. Its estimated MTBF is:

A. 5 hours

B. 25 hours

C. 400 hours

D. 10,000 hours

Correct Answer: C. 400 hours

Question 4

Which expression is the usual inherent availability formula?

A. MTBF/(MTBF + MTTR)

B. MTTR/(MTBF + MTTR)

C. MTBF × MTTR

D. MTBF − MTTR

Correct Answer: A. MTBF/(MTBF + MTTR)

Question 5

OEE is calculated as:

A. Availability + Performance + Quality

B. Availability × Performance × Quality

C. Availability ÷ Performance ÷ Quality

D. Performance − Availability − Quality

Correct Answer: B. Availability × Performance × Quality

Question 6

A line has availability 0.80, performance 0.90 and quality 0.95. Its OEE is approximately:

A. 0.684 or 68.4%

B. 0.720 or 72.0%

C. 0.855 or 85.5%

D. 2.65 or 265%

Correct Answer: A. 0.684 or 68.4%

Question 7

What is the principal purpose of a CMMS work order?

A. To replace all engineering judgement

B. To record and control maintenance work and its evidence

C. To guarantee zero equipment failures

D. To eliminate the need for spare parts

Correct Answer: B. To record and control maintenance work and its evidence

Question 8

In a simple independent series system, system reliability is:

A. The sum of component reliabilities

B. The largest component reliability

C. The product of component reliabilities

D. Always greater than every component reliability

Correct Answer: C. The product of component reliabilities

Question 9

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

A. Failure can release hazardous chemicals

B. Failure can injure operators

C. The item is inexpensive, non-critical and easily replaced

D. Failure stops a bottleneck production line

Correct Answer: C. The item is inexpensive, non-critical and easily replaced

Question 10

What is a major limitation of a risk matrix score?

A. It cannot include consequence

B. It always gives a precise probability of loss

C. Ordinal rankings may create false precision

D. It is usable only for financial assets

Correct Answer: C. Ordinal rankings may create false precision