Reliability, Maintenance and Asset Management

Introduction

Industrial systems create value only when their equipment performs safely, consistently and economically. A production line with high nominal capacity but frequent failures may deliver less output than a smaller, more reliable line. Reliability engineering therefore connects engineering design, operations management, maintenance planning, safety, quality and financial performance.

Reliability concerns the probability that an item performs its required function for a specified time under stated conditions. Maintenance is the set of technical, administrative and managerial actions intended to retain or restore an asset to a condition in which it can perform its required function. Asset management extends this view across the entire life cycle: planning, acquisition, operation, maintenance, renewal and disposal.

The central managerial challenge is not to eliminate every failure at any cost. It is to achieve an acceptable balance among performance, risk, safety, environmental protection and life-cycle cost.

Learning Objectives

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

• define reliability, availability, maintainability, failure rate and mean-time measures;
• calculate reliability and unreliability for common probability models;
• analyze series and parallel systems;
• distinguish corrective, preventive, predictive and condition-based maintenance;
• apply failure mode and effects analysis and fault-tree analysis;
• explain total productive maintenance and life-cycle asset management;
• select maintenance policies using technical, economic and risk-based reasoning;
• recognize limitations of reliability models and the importance of data quality.

1. Fundamental Concepts and Measures

1.1 Failure and functional performance

A failure occurs when an item is unable to perform a required function within specified performance limits. The definition must identify the function, operating context and acceptance criteria. For example, a pump may be considered failed when it stops completely, when its flow falls below the process requirement, or when vibration exceeds a safety limit.

A failure can be classified as:

• critical or catastrophic, when it causes immediate danger or major system loss;
• degraded, when the item continues operating but below specification;
• incipient, when evidence of future failure exists although the function is currently available;
• overt, when the loss of function is directly observable.

The distinction between a failure and a fault is useful. A fault is an abnormal condition or defect; failure is the inability to perform the required function. Several faults may exist before a failure occurs.

1.2 Reliability and unreliability

Reliability is the probability that an item survives without functional failure to time t:

where T is the random time to failure. Reliability is dimensionless and lies between 0 and 1. The complementary probability is unreliability, or the cumulative distribution function:

These expressions apply to a defined population of similar items operating under specified conditions. Reliability is not an intrinsic constant: load, environment, maintenance, manufacturing variation and operating practice all affect it.

1.3 Probability density and hazard rate

If failure time is continuous, the probability density function is f(t), where:

The instantaneous failure rate, or hazard rate, is:

The unit of λ(t) is inverse time, such as failures per hour. It is the conditional rate of failure at time t, given survival to t. It is not generally the same as the proportion of all units that fail per unit time.

The cumulative hazard is:

and therefore:

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

This model assumes a constant hazard, statistical independence of failure events, and operation under unchanged conditions. It is often useful during the random-failure portion of a product’s life, but it should not be applied automatically to every asset.

1.4 Mean time to failure and mean time between failures

For a non-repairable item, the mean time to failure is:

For an exponential lifetime distribution:

For a repairable item, mean time between failures (MTBF) is the average operating time between successive failures. In simple steady-state situations, MTBF is often approximated as the reciprocal of the failure rate. However, MTBF is not a universal measure of useful life and should not be interpreted as a guaranteed service interval.

Mean time to repair (MTTR) is the average active repair time. In practice, total downtime may also include diagnosis, waiting for parts, authorization, access, cooling, testing and return to service. These delays should be reported separately where possible.

Worked Example 1: Exponential reliability

A motor has a constant failure rate of 0.0002 failures per operating hour. Calculate its reliability over 1,000 hours and its MTTF.

Using the exponential model:

Thus, the probability of surviving 1,000 hours is approximately 81.9%, and unreliability is 18.1%. The mean time to failure is:

The result is a population average, not a prediction that every motor will operate exactly 5,000 hours.

2. Availability and Maintainability

2.1 Availability

Availability is the probability that an item is in a state to perform its required function when demanded. It depends on both failures and restoration.

For a repairable item in steady operation, inherent availability is commonly expressed as:

This assumes that MTBF and MTTR are stable averages and that downtime consists only of active corrective repair. Operational or achieved availability should include relevant logistics, preventive maintenance and administrative delays. A general long-run measure is:

Availability and reliability are different. A highly repairable machine can have high availability despite frequent failures, while a non-repairable safety device may require high reliability even though availability is less meaningful.

Worked Example 2: Availability

A packaging machine operates, on average, 360 hours between failures. Active repair takes 12 hours. Its inherent availability is:

Therefore, inherent availability is approximately 96.8%. If waiting for spare parts adds 8 hours per failure, operational availability based on the same cycle becomes:

The difference demonstrates why reducing logistical delay can be as valuable as reducing repair time.

2.2 Maintainability

Maintainability is the probability that a failed item can be restored to a specified condition within a specified time, using stated procedures and resources. If repair time M has distribution G(t), then:

For an exponential repair-time model with repair rate μ:

where μ has units of repairs per hour and mean active repair time is 1/μ. The exponential repair model assumes a constant repair rate and is most suitable when repair tasks have no pronounced learning, setup or wear-out pattern.

Design measures that improve maintainability include modular replacement, accessibility, standard fasteners, clear diagnostics, lifting provisions, isolation points, built-in test facilities and safe work instructions.

3. Reliability of System Configurations

3.1 Series systems

In a series system, every component must function for the system to function. If component failures are statistically independent:

For n identical components, each with reliability R(t):

For independent exponential components with rates λᵢ, the system rate is:

The series result shows why adding components can reduce system reliability. The weakest component, common-cause failures and maintenance configuration must also be considered.

3.2 Parallel systems

In a parallel-redundant system, the system succeeds if at least one component functions. For independent components:

For two identical components:

This benefit disappears if failures are caused by a common power supply, environment, design defect or maintenance error. Redundancy may also introduce switching, testing and common-cause risks.

Worked Example 3: Series and parallel configurations

Two independent components each have reliability 0.90 over a mission. In series:

In active parallel redundancy:

The parallel arrangement improves reliability from 81% to 99%, provided that independence and adequate load sharing or switching are valid assumptions.

4. The Bathtub Curve and Failure Behaviour

A traditional bathtub curve describes three broad regions:

1. Early-life failures: decreasing failure rate, often associated with manufacturing defects, installation errors or commissioning problems.
2. Useful life: approximately constant failure rate, commonly associated with random stresses.
3. Wear-out period: increasing failure rate due to fatigue, corrosion, erosion, insulation aging or other degradation.

The curve is a conceptual model rather than a universal law. Some products do not display all three regions, and modern electronics may have complex failure distributions. Maintenance intervals should therefore be based on evidence about failure mechanisms, not on the curve alone.

5. Maintenance Strategies

5.1 Corrective maintenance

Corrective maintenance restores an asset after a detected failure. It is suitable when consequences are low, failure is predictable in its effects, repair is rapid and spare capacity exists. Unplanned corrective maintenance creates production disruption, safety exposure and emergency procurement.

5.2 Preventive maintenance

Preventive maintenance is performed at predetermined intervals or according to prescribed criteria. Time-based replacement is effective when age is strongly related to failure probability and the replacement task is economical. Excessive replacement can introduce infant mortality, maintenance errors and unnecessary cost.

5.3 Predictive and condition-based maintenance

Predictive maintenance uses measurements and analysis to estimate future condition or remaining useful life. Condition-based maintenance performs action when evidence indicates that intervention is required. Typical technologies include vibration analysis, thermography, oil analysis, ultrasonic inspection, motor-current analysis and performance trending.

A condition-monitoring program requires:

• a failure mode that produces a detectable physical signal;
• a useful interval between detectable degradation and functional failure;
• reliable sensors and data collection;
• competent interpretation;
• a defined decision threshold and work process.

Artificial intelligence can support anomaly detection, image inspection and remaining-useful-life estimation. It does not remove the need for representative data, explainable decisions, sensor validation, cybersecurity and engineering review. A model trained on normal operation may fail when process conditions, product mix or equipment configuration changes.

5.4 Reliability-centred maintenance

Reliability-centred maintenance (RCM) selects maintenance tasks according to functions, functional failures, failure modes, effects and consequences. A typical RCM analysis asks:

• What should the asset do?
• How can it fail to perform that function?
• What causes each functional failure?
• What happens when the failure occurs?
• Why does the failure matter?
• What proactive task is technically and economically justified?
• What should be done if no proactive task is effective?

RCM emphasizes consequence-based decisions rather than equal treatment of every component.

6. Failure Mode and Effects Analysis

Failure mode and effects analysis (FMEA) is a structured, bottom-up method. For each item, the team identifies functions, failure modes, causes, local effects, system effects, existing controls and recommended actions.

Traditional FMEA sometimes uses severity S, occurrence O and detection D rankings to calculate a risk priority number:

Each ranking is usually ordinal, often on a 1–10 scale. The product is not a probability and should not be treated as a statistically precise risk measure. A low RPN can conceal a high-severity failure if occurrence or detection scores are low. Organizations should therefore use severity gates, action thresholds and engineering judgment in addition to RPN.

FMEA is most useful when performed by a multidisciplinary team including design, operations, maintenance, quality and safety personnel. It should be updated after field experience, design change or significant incidents.

7. Fault-Tree Analysis

Fault-tree analysis (FTA) is a top-down deductive method. It begins with an undesired top event, such as “loss of cooling,” and logically decomposes it into contributing events using AND and OR gates.

For independent basic events A and B:

• OR gate probability: P(A OR B) = P(A) + P(B) − P(A)P(B);
• AND gate probability: P(A AND B) = P(A)P(B).

For rare, independent events, the OR expression is often approximated by the sum of the individual probabilities, but this approximation must be checked against the exact expression. FTA identifies minimal cut sets, common-cause vulnerabilities and opportunities for redundancy. It is especially valuable for safety and high-consequence systems.

FMEA is generally bottom-up and component-focused; FTA is generally top-down and system-focused. They are complementary rather than competing techniques.

8. Total Productive Maintenance

Total productive maintenance (TPM) seeks maximum equipment effectiveness through participation across the organization. Its major principles include autonomous maintenance by operators, planned maintenance, focused improvement, early equipment management, quality maintenance, training, safety and office support.

Overall equipment effectiveness (OEE) is commonly defined as:

Availability is often calculated as operating time divided by planned production time. Performance compares actual output rate with ideal rate, and quality is good units divided by total units. The exact definitions must be agreed locally; otherwise comparisons become misleading.

Worked Example 4: OEE

A line has availability 0.90, performance 0.95 and quality 0.98:

OEE is therefore approximately 83.8%. The loss tree should then identify whether the greatest opportunity lies in downtime, speed loss or defective output. OEE is a diagnostic indicator, not a complete measure of asset value or maintenance excellence.

9. Life-Cycle Asset Management

Life-cycle asset management integrates technical performance, risk and cost from concept to disposal. A life-cycle cost model may be expressed as:

where C₀ is acquisition cost, Cᵢ installation and commissioning cost, Cₑ energy cost, Cₘ maintenance and spare-parts cost, Cₒ operating cost, C_d disposal cost, and S residual or salvage value. If costs occur at different times, they should be discounted to present value:

where C_t is a future cash flow, r is the discount rate per period and t is the number of periods. Risk-related costs, production losses and environmental liabilities should be included when they are material and can be estimated credibly.

Asset management decisions should align with organizational objectives, risk tolerance, regulatory requirements and stakeholder needs. Important practices include an asset register, criticality ranking, maintenance strategy, work management, spare-parts control, competence management, configuration control and performance review.

10. Illustrative Case: Cooling-Water Pumps in a Process Plant

Background

A fictional process plant operates two cooling-water pumps. Either pump can supply the minimum required flow, but both are needed during peak demand. The pumps have experienced repeated seal failures, causing emergency maintenance and occasional production reduction.

Problem

The maintenance department proposes replacing seals every six months. Operations argues that the failures are associated with misalignment and dry running during low-level conditions. The plant must choose an effective maintenance approach without creating unnecessary planned work.

Analysis

The team constructs an FMEA. The principal failure mode is mechanical-seal leakage. Causes include shaft misalignment, inadequate cooling, contaminated fluid and operation below the minimum stable flow. Effects range from minor leakage to pump unavailability and environmental release. The failure has safety and production consequences, so purely run-to-failure maintenance is unsuitable.

Vibration data show increasing axial vibration before some failures, while seal-temperature measurements identify low-flow operation. A fault tree for “loss of cooling capacity” identifies pump failure, unavailable standby power and closed isolation valves as major branches. The analysis reveals that the two pumps share a vulnerable power-control panel, so nominal parallel redundancy does not provide full independence.

Approach and solution

The plant implements the following controls:

• alignment verification after installation and major maintenance;
• a minimum-flow interlock and operator response procedure;
• weekly inspection of leakage and temperature;
• monthly vibration trending;
• quarterly functional testing of the standby pump and automatic changeover;
• critical seal and bearing spares stored on site;
• a documented work order closeout requiring failure-cause coding;
• review of the shared electrical panel as a common-cause risk.

Seal replacement is no longer performed solely at a fixed six-month interval. It is triggered by condition evidence or planned outage opportunity, subject to safety and environmental requirements.

Outcome

In this illustrative case, emergency pump work decreases, planned maintenance becomes more targeted and the plant identifies a shared electrical vulnerability that was not visible from component-level reliability figures. The most important improvement is not a single predictive instrument; it is the integration of design, operating discipline, condition monitoring, spares and learning from failure data.

Learning points

• Redundancy must be examined at system level, including common-cause failures.
• A maintenance interval should reflect the actual failure mechanism.
• Condition monitoring is valuable only when measurements lead to timely action.
• Accurate failure coding is essential for continuous improvement.

11. Challenges, Governance and Best Practice

Reliability calculations are limited by inaccurate failure definitions, small samples, censored data, changing operating conditions and dependence between components. MTBF can be distorted when minor failures are not recorded or when overhaul events are treated inconsistently. Statistical estimates should include uncertainty where decisions are significant.

Maintenance work also involves ethical and legal responsibilities. Safety-critical equipment must not be deferred merely to improve production indicators. Isolation, lockout/tagout, competence, permits, contractor control and regulatory inspection must be embedded in the management system. Digital monitoring introduces cybersecurity, privacy and data-governance concerns. Automated recommendations should remain auditable, and personnel must be able to challenge unsafe or implausible outputs.

Good practice includes:

• defining asset functions and failure criteria explicitly;
• using criticality and consequence to prioritize resources;
• combining reliability data with engineering knowledge;
• separating active repair time from logistical delay;
• validating sensors, alarms and analytical models;
• reviewing maintenance after failures and near misses;
• designing for maintainability, inspectability and safe isolation;
• evaluating alternatives using life-cycle cost and risk rather than purchase price alone;
• treating operators as essential sources of condition information;
• continually improving the asset register and failure taxonomy.

Chapter Summary

Reliability is the probability that an item performs its required function for a specified time and condition. Unreliability, failure density and hazard rate describe different aspects of failure behaviour. Exponential models are useful under constant-hazard assumptions, while real assets may exhibit early-life, random and wear-out behaviour.

Availability combines operating time and restoration performance. Maintainability describes the probability of restoring a failed item within a specified time. Series systems become less reliable as required components are added; parallel redundancy can improve reliability but is weakened by common-cause failures.

Corrective, preventive, predictive and condition-based maintenance should be selected according to failure mechanisms, consequences, data quality and cost. FMEA provides a bottom-up analysis, whereas FTA works top-down from an undesired event. TPM connects operator involvement, planned maintenance and improvement, while OEE separates availability, performance and quality losses.

Life-cycle asset management extends decisions from acquisition to disposal. Effective practice integrates engineering, operations, safety, finance, data governance and human competence. Reliability improvement is therefore a socio-technical and life-cycle discipline, not merely a calculation of failure rates.

Key Terms

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

Unreliability: Probability that an item has failed by a specified time.

Failure rate: Conditional rate at which failures occur among surviving items.

Hazard rate: Another term for instantaneous failure rate.

MTTF: Mean time to failure for a non-repairable item.

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

MTTR: Mean active time required to repair an item.

Availability: Probability or proportion of time that an item is able to perform its function.

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

Series system: System that functions only when every required component functions.

Parallel system: System that functions when at least one redundant component functions.

Common-cause failure: Failure of multiple components due to one shared cause.

Corrective maintenance: Maintenance performed to restore function after failure.

Preventive maintenance: Maintenance performed at planned intervals or predetermined criteria.

Predictive maintenance: Maintenance based on analysis of condition or degradation data.

FMEA: Bottom-up analysis of failure modes, causes, effects and controls.

FTA: Top-down logical analysis of an undesired system event.

TPM: Organization-wide approach to maximizing equipment effectiveness.

OEE: Product of availability, performance and quality factors.

Life-cycle cost: Total relevant cost of an asset over its life.

Review Questions

Short Answer Questions

1. Distinguish reliability from availability.
2. What assumptions support the exponential reliability model?
3. Why can parallel redundancy fail to provide the expected reliability improvement?
4. Define maintainability and MTTR.
5. State one difference between FMEA and FTA.

Long Answer Questions

1. Explain the bathtub curve and discuss why maintenance intervals should not be based on it alone.
2. Compare corrective, preventive, predictive and condition-based maintenance, including suitable applications and limitations.
3. Explain how TPM and OEE support continuous improvement in a manufacturing plant.
4. Describe the stages of a reliability-centred maintenance analysis.
5. Discuss life-cycle asset management and explain why acquisition price is an incomplete basis for asset selection.

Analytical / Application Questions

1. Two independent components each have reliability 0.92 for a mission. Calculate the reliability of their series arrangement and active parallel arrangement. Interpret the results.
2. A machine has MTBF of 480 hours and active MTTR of 16 hours. Calculate inherent availability. If logistics add 10 hours per failure, calculate the corresponding operational-cycle availability and explain the difference.
3. Develop a concise FMEA for a centrifugal pump that loses required flow. Identify at least three failure modes, their likely causes and effects, and recommend an appropriate maintenance or design action for each.

Multiple Choice Questions with Answers

Question 1

A motor has a constant failure rate of 0.0001 failures per hour. What is its reliability after 1,000 hours under the exponential model?

A. 0.1000

B. 0.3679

C. 0.9048

D. 0.9999

Correct Answer: C. 0.9048

Question 2

For two independent components in series, each with reliability 0.90, the system reliability is:

A. 0.09

B. 0.81

C. 0.90

D. 0.99

Correct Answer: B. 0.81

Question 3

For two independent components in active parallel, each with reliability 0.90, the system reliability is:

A. 0.81

B. 0.90

C. 0.95

D. 0.99

Correct Answer: D. 0.99

Question 4

A repairable machine has MTBF of 200 hours and MTTR of 10 hours. Its inherent availability is approximately:

A. 90.0%

B. 95.2%

C. 97.5%

D. 99.0%

Correct Answer: B. 95.2%

Question 5

Which measure is most directly associated with the probability of restoring a failed item within a specified time?

A. Reliability

B. Maintainability

C. Failure density

D. Unreliability

Correct Answer: B. Maintainability

Question 6

Which maintenance strategy uses measurements such as vibration and thermography to determine intervention needs?

A. Run-to-failure maintenance

B. Time-based replacement only

C. Condition-based maintenance

D. Emergency maintenance only

Correct Answer: C. Condition-based maintenance

Question 7

Which statement correctly describes FMEA?

A. It is normally a top-down analysis beginning with a system hazard.

B. It identifies failure modes, causes, effects and controls, usually from component level upward.

C. It calculates availability directly from MTBF and MTTR only.

D. It applies only to financial assets.

Correct Answer: B. It identifies failure modes, causes, effects and controls, usually from component level upward.

Question 8

In a fault tree with independent basic events A and B connected by an AND gate, the top-event probability is:

A. P(A) + P(B)

B. P(A) − P(B)

C. P(A)P(B)

D. 1 − P(A)P(B)

Correct Answer: C. P(A)P(B)

Question 9

If availability is 0.90, performance is 0.95 and quality is 0.98, OEE is approximately:

A. 0.8379

B. 0.9000

C. 0.9310

D. 2.8300

Correct Answer: A. 0.8379

Question 10

Which item is most characteristic of life-cycle asset management?

A. Selecting equipment solely by lowest purchase price

B. Ignoring disposal and energy costs

C. Considering performance, risk and cost from acquisition through disposal

D. Replacing every component at the same fixed interval

Correct Answer: C. Considering performance, risk and cost from acquisition through disposal