Planned Maintenance Management

Introduction and Context

Planned maintenance is the systematic preparation, authorization, execution and improvement of maintenance work before asset failure occurs or before a known maintenance opportunity is lost. In Total Productive Maintenance (TPM), planned maintenance complements autonomous maintenance, focused improvement, quality maintenance, training and safety management. Its purpose is not simply to increase the number of preventive tasks. It is to achieve the required equipment function at acceptable risk and life-cycle cost.

A mature planned-maintenance system answers six questions:

1. Which assets and failure modes matter most?
2. What maintenance task is technically justified?
3. What resources, materials, permits and skills are required?
4. When should the work be performed?
5. How will completion and quality be verified?
6. What evidence will be used to improve the program?

The central managerial challenge is to balance prevention, correction, production continuity, safety, regulatory compliance and cost. Excessive maintenance can create induced failures, consume scarce resources and increase cost. Insufficient maintenance increases breakdowns, hazards, quality losses and emergency work.

Learning Objectives

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

• distinguish preventive, predictive, corrective, shutdown and opportunity maintenance;
• rank equipment using asset criticality and failure consequences;
• select maintenance tasks from failure history, failure modes and condition information;
• design job plans, work orders, schedules and shutdown programs;
• control backlog, priorities, compliance and maintenance standards;
• evaluate program performance using reliability, availability, schedule and cost measures;
• critically assess the limitations of calendar-based maintenance and the role of digital and AI-enabled tools.

1. The Planned-Maintenance System

Planned maintenance is a socio-technical system rather than a collection of schedules. It connects asset information, engineering knowledge, operations, stores, contractors, safety controls, execution and feedback.

A useful lifecycle is:

Identify → assess → plan → schedule → execute → verify → close → learn.

Identification begins with a defect, inspection finding, recurring failure, statutory requirement or scheduled task. Assessment determines risk and urgency. Planning defines the work method and resources. Scheduling places work in a feasible production and resource window. Execution follows the approved job plan and permit controls. Verification confirms that the asset is safe, functional and returned to service. Closure records actual work, findings, measurements and recommendations. Learning updates the maintenance strategy, bill of materials, standards or training.

The system should distinguish three levels of control:

Work management should not be confused with maintenance strategy. A perfectly executed job is still wasteful if the task does not address a credible failure mode.

2. Maintenance Policies and Task Types

Preventive maintenance

Preventive maintenance is performed at predetermined intervals or according to prescribed criteria to reduce the probability of failure or degradation. It includes inspection, cleaning, lubrication, adjustment, replacement and functional testing.

Time-based replacement is appropriate when failure probability increases predictably with age and the consequences justify replacement. It is inappropriate when failures are largely random or when replacement introduces significant installation risk.

Predictive and condition-based maintenance

Condition-based maintenance uses observable evidence—such as vibration, thermography, oil debris, electrical signatures, process performance or ultrasonic readings—to decide when intervention is necessary. It is most effective when:

• a measurable degradation phase exists;
• the condition indicator is related to functional failure;
• inspection intervals provide sufficient warning;
• the cost of monitoring is justified by avoided consequences.

Predictive maintenance is therefore a decision system, not merely the purchase of sensors.

Corrective maintenance

Corrective maintenance restores an asset after a defect or failure has been identified. It may be planned or unplanned. Planned corrective work is often preferable to immediate intervention when the defect is safe to defer and resources can be coordinated. Unplanned corrective maintenance is emergency or reactive work caused by failure or unacceptable deterioration.

Shutdown and turnaround maintenance

Shutdown maintenance is concentrated work performed during a planned production interruption. It may include statutory inspection, major overhaul, campaign changeover, modification and accumulated corrective work. Shutdown planning requires a defined scope freeze, network schedule, critical-path analysis, contractor strategy, materials readiness, isolation plan and start-up assurance.

Opportunity maintenance

Opportunity maintenance uses an already available production or access window to perform additional work. Examples include replacing a coupling while a motor is isolated for another job or inspecting a vessel during an adjacent unit shutdown. Opportunity work is beneficial only if it does not compromise the primary job, restart date or risk controls.

Run-to-failure

Run-to-failure is a deliberate policy, not an absence of policy. It can be suitable for non-critical, inexpensive, easily replaceable items where failure has no unacceptable safety, environmental, quality or production consequence. It is not suitable merely because preventive maintenance is inconvenient.

3. Asset Criticality and Maintenance Strategy Selection

Criticality analysis ranks assets according to the consequences and likelihood of undesirable events. A practical assessment considers:

• personnel safety and process safety;
• environmental and legal consequences;
• production loss and customer impact;
• product quality;
• repair cost and spare-part lead time;
• redundancy and the ability to operate at reduced capacity;
• detectability and restoration time.

A simple semi-quantitative risk model is:

where R is risk score, P is an assigned likelihood category and C is an assigned consequence category. This is a prioritization aid, not a physically exact probability model. Categories must be defined consistently, and high-severity events should not be hidden by averaging or scoring alone.

For more detailed analysis, expected annual loss may be represented as:

where pᵢ is the probability of consequence scenario i during a specified period and Cᵢ is its monetary or consequence value. The expression assumes scenarios are defined consistently and that their expected values are decision-useful; it does not capture all catastrophic or interdependent risks.

Criticality should lead to differentiated strategies. A high-criticality compressor may require condition monitoring, redundancy review, tested protective devices and a stocked strategic spare. A low-criticality guard fan may use inspection and run-to-failure. The correct question is not “How often should everything be serviced?” but “What policy best manages each credible failure mode?”

Reliability-centred reasoning

Failure Mode and Effects Analysis (FMEA) or Reliability-Centred Maintenance (RCM) asks:

1. What functions must the asset perform?
2. How can it fail to perform those functions?
3. What causes each functional failure?
4. What happens when the failure occurs?
5. What are the consequences?
6. What task can predict, prevent or detect the failure?
7. If no task is effective, should the design be changed, redundancy added or failure accepted?

A maintenance task is justified when it is technically effective, operationally feasible and economically or risk-wise defensible. FMEA ratings should support discussion rather than create false precision. A high numerical ranking is not automatically a reason for a calendar task.

4. Designing Preventive-Maintenance Programs

A preventive-maintenance program should be built from asset functions, failure modes, manufacturer information, operating context, legal requirements and local failure history. Manufacturer recommendations are useful starting evidence but must be adapted to duty, environment, load, contamination and consequence.

Each task should specify:

• asset and component identification;
• failure mode or risk addressed;
• task type and method;
• interval or trigger;
• acceptance criteria;
• required competence and tools;
• isolation, permit and personal protective equipment requirements;
• estimated labour and duration;
• materials and calibration requirements;
• data to record;
• escalation rules when results are abnormal.

Intervals should be reviewed using evidence. If a task repeatedly finds no deterioration and has no compliance justification, its value should be challenged. If failures occur before the task interval, the team should examine whether the task is ineffective, the interval is too long, the operating context changed or the failure mechanism was misunderstood.

Reliability and availability measures

For repairable assets, mean time between failures is commonly calculated as:

Mean time to repair is:

The simple inherent-availability approximation is:

Here, operating time and repair time must use the same units, and the approximation assumes stable average rates, statistically comparable cycles and no additional logistics or preventive downtime. For operational availability, waiting time, supply delays, planned downtime and administrative delays may also need to be included; therefore, the simple formula should not be used to represent every availability definition.

Worked example. A pump operates for 7,200 hours and experiences 6 failures. Average corrective repair time is 8 hours per failure.

MTBF = 7,200 ÷ 6 = 1,200 h

MTTR = 8 h

A_i = 1,200 ÷ (1,200 + 8) = 0.9934

Thus, the approximate inherent availability is 99.34%. This result does not prove that the pump is satisfactory: production consequence, standby capacity, quality effects and safety performance must also be assessed.

5. Corrective Work, Defect Elimination and Failure History

Corrective work should be separated into emergency, urgent, routine and deferred categories, with definitions agreed by operations and maintenance. Priority must reflect consequence and risk, not the seniority of the person requesting work.

Failure history should contain more than a short text such as “bearing failed.” Useful records include:

• operating conditions and load;
• failure mode and failed component;
• symptoms and detection method;
• immediate cause and underlying cause;
• repair action and parts used;
• downtime, labour and production impact;
• photographs, measurements and test results;
• recommendations and responsible owner.

Repeated corrective work demands structured analysis. Root-cause analysis may use causal trees, five-whys, fault-tree analysis, change analysis or engineering testing. The objective is to distinguish a physical failure mechanism from a management-system cause such as poor lubrication control, incorrect assembly, weak design, contamination or inadequate training.

A recurring failure should not automatically generate another preventive task. It may require redesign, alignment improvement, operating-limit control, spare-quality improvement or elimination of a harmful maintenance practice.

6. Job Planning and Work Orders

Planning determines what is required; scheduling determines when it will be done. A planner should normally produce a job plan before work is released, except for genuine emergencies.

A high-quality job plan includes:

1. clear scope and functional location;
2. asset history and technical drawings;
3. job sequence and estimated craft hours;
4. skills, tools, lifting equipment and special equipment;
5. parts, consumables and reserved materials;
6. isolation, lockout/tagout, permits and energy-control steps;
7. quality checks, tolerances and test requirements;
8. restoration and commissioning instructions;
9. waste, environmental and housekeeping controls;
10. feedback fields for actual findings and improvements.

The work order is the governance record. It authorizes work, links cost and risk, preserves traceability and provides evidence of completion. Closing a work order without meaningful feedback destroys the learning loop.

Work-order statuses

A controlled status model might include:

Request → Screened → Approved → Planned → Ready → Scheduled → In progress → Technically complete → Verified → Closed.

Each transition should have an owner and a definition. “Ready” should mean that scope, labour, materials, permits and access conditions are sufficiently prepared—not simply that someone has entered a date in a computer system.

7. Scheduling, Shutdowns and Opportunity Work

Weekly scheduling should be based on available craft hours, planned work duration, priority, access and production constraints. A schedule that contains more work than available capacity is a wish list, not a schedule.

A simple schedule-capacity calculation is:

The ratio must be interpreted carefully. Available hours should exclude planned leave, training and known nonproductive allowances. A load factor near 100% may be fragile because urgent work and execution variability remain. Organizations should set a locally validated target rather than treating one universal percentage as correct.

Shutdown planning is best managed through a scope-control process:

• establish objectives and boundaries;
• identify mandatory statutory and safety work;
• develop a work breakdown structure;
• estimate duration and resources;
• identify logic links and the critical path;
• confirm materials, contractors and permits;
• freeze and challenge scope;
• conduct readiness reviews;
• execute with daily control meetings;
• verify start-up and capture post-shutdown learning.

Opportunity work should pass a decision test: Is the asset already safely accessible? Is the additional work risk-reducing or value-adding? Are parts and competence available? Could it delay restart or obscure the original defect? If the answer to the final question is yes, the work should normally be deferred or separately controlled.

8. Backlog Control and Prioritization

Backlog is approved or identified work not yet completed. It is not automatically waste. A healthy backlog contains properly planned work that can be loaded into future schedules. An unhealthy backlog contains obsolete requests, vague scopes, duplicate notifications, overdue compliance tasks or work with no owner.

Backlog governance should include:

• age profile by priority;
• safety and statutory exceptions;
• ready-to-schedule work versus unplanned work;
• estimated hours by craft and week;
• material and access constraints;
• deferred-risk register;
• periodic work-order cleansing.

A practical review asks whether every item is still necessary, whether risk has changed, whether the scope is clear, and whether the item can be bundled with another intervention. Deferral must be an explicit risk decision with an owner and review date.

Useful indicators include schedule compliance, planned-work percentage, emergency-work percentage, overdue statutory work, repeat failures, maintenance-induced defects and backlog age. No single metric should be optimized in isolation. For example, raising schedule compliance by deleting difficult work from the schedule is performance manipulation, not improvement.

9. Maintenance Standards, Compliance and Governance

Maintenance standards translate engineering intent and legal obligations into repeatable work. They may cover lubrication, torque, alignment, welding, inspection, electrical testing, calibration, cleaning, contamination control and post-maintenance testing.

A standard should state its scope, required competence, equipment, method, acceptance criteria, records, revision authority and exceptions process. Standards must be controlled documents. Obsolete procedures, uncontrolled local copies and undocumented deviations create operational and legal exposure.

Compliance has at least three dimensions:

• technical compliance: the asset meets specified functional and engineering requirements;
• legal or regulatory compliance: inspections, tests and records meet applicable law and permits;
• management-system compliance: work follows approved processes, competence requirements and audit controls.

Safety-critical maintenance requires independent verification where appropriate. Examples include proof testing of protective systems, inspection of lifting equipment, pressure-boundary examination and verification of isolation. Records must be trustworthy, retrievable and protected from retrospective alteration.

Ethically, maintenance leaders must not conceal defects, falsify completion, pressure workers to bypass isolation or manipulate indicators. Production urgency does not remove the duty to control hazardous energy or communicate residual risk.

10. Performance Measurement and Continuous Improvement

A balanced maintenance dashboard combines leading and lagging measures.

Definitions should be documented. For example, schedule compliance may be calculated as:

The organization must specify whether partial completion counts, how cancelled work is treated and whether emergency additions are included. Without a stable definition, trend comparisons are unreliable.

Planned-maintenance optimization is a closed-loop activity. Performance data should lead to strategy review, not merely monthly reporting. A rise in MTBF can be misleading if production hours declined or failures were underreported. Conversely, a rise in recorded failures can indicate better detection and reporting rather than deteriorating equipment.

11. Digital Systems, Analytics and AI

A computerized maintenance management system (CMMS) or enterprise asset-management system can integrate asset hierarchy, work orders, materials, labour, permits, meters and cost. Its value depends on data quality and disciplined use. A sophisticated system cannot compensate for unclear ownership, poor failure coding or weak field feedback.

Sensors and analytics can support anomaly detection, remaining-useful-life estimation and risk-based inspection. However, models require representative data, stable sensor quality and human validation. False alarms cause alarm fatigue; missed alarms create misplaced confidence. AI recommendations should therefore be explainable enough for engineering review, protected against data drift and subject to approval for safety-critical actions.

Technology should support—not replace—field inspection, engineering judgment, worker consultation and formal management of change. Cybersecurity is also a maintenance concern because compromised control or asset data can create safety and availability risks.

12. Illustrative Case: Planned Maintenance of a Process-Cooling System

Background

A fictional food-processing plant operates three centrifugal cooling-water pumps. Pump P-201 is the duty pump, P-202 is standby and P-203 is available for maintenance rotation. Production losses occur when cooling-water flow falls below the process requirement.

Problem

During six months, P-201 experienced four seal failures and two bearing-temperature alarms. Emergency repairs caused 31 hours of downtime. The existing program consisted of monthly visual inspection and annual bearing replacement on all pumps.

Analysis

The team reviewed work orders, operating conditions and vibration data. Seal failures were concentrated after pump changeovers. Investigation found that the standby pump was occasionally started against a closed discharge valve, producing hydraulic shock. Bearing alarms occurred after lubrication intervals were shortened by operators using an unsuitable grease. The annual replacement policy also introduced assembly variation and did not address alignment.

The failure modes were ranked as high production consequence but moderate likelihood, because standby capacity usually existed. The team therefore selected a mixed strategy rather than simply increasing replacement frequency.

Approach and solution

The revised program included:

• a verified start-up and valve-position checklist;
• operator training for pump changeover;
• quarterly laser-alignment verification after maintenance;
• monthly vibration and temperature trending;
• lubrication by specified grease, quantity and method;
• condition-based bearing replacement rather than automatic annual replacement;
• a stocked seal kit and defined reorder point;
• a planned quarterly rotation of duty and standby pumps;
• a work-order failure code distinguishing seal, bearing, alignment and operating causes.

The job plan included isolation, drain-down, contamination control, alignment tolerances, test-run requirements and independent verification of valve position.

Outcome

In the following six months, seal failures fell from four to one and unplanned downtime fell from 31 hours to 9 hours. The remaining seal failure was traced to a damaged shaft sleeve and generated a redesign action. Maintenance cost increased initially because of alignment equipment and training, but total operational loss decreased.

Learning points

The case illustrates that planned maintenance is not synonymous with more frequent replacement. The effective program combined operating discipline, condition monitoring, job quality, spare readiness and failure coding. It also shows why outcomes should be evaluated across reliability, risk, production and total cost.

Challenges and Limitations

Planned-maintenance systems face several persistent limitations. Failure data may be incomplete or biased toward dramatic events. Criticality scores can create false confidence when consequence categories are poorly defined. Calendar intervals may be retained because they are familiar rather than effective. Resource constraints can turn a logical program into deferred work. Contractors may perform tasks without returning useful technical feedback. Digital systems can encourage data entry without learning.

Maintenance also operates under uncertainty. Equipment interacts with operating practices, process variation, environment and human decisions. Consequently, strategy should be reviewed after significant process changes, repeated failures, modifications, incidents or changes in consequence—not only at a fixed annual date.

Best-Practice Checklist

An effective planned-maintenance program normally has:

• a current asset hierarchy and criticality register;
• documented failure modes and maintenance strategies;
• clear definitions for priority, backlog and schedule compliance;
• job plans with safety, quality and restoration requirements;
• materials and competence available before scheduling;
• disciplined work-order closure and failure coding;
• formal shutdown and opportunity-work controls;
• risk-based management of deferrals;
• controlled maintenance standards and auditable records;
• periodic optimization using evidence rather than habit;
• ethical escalation of unsafe or noncompliant conditions.

Chapter Summary

Planned maintenance is the managed lifecycle of identifying, assessing, preparing, scheduling, executing, verifying and learning from maintenance work. Its foundation is risk- and failure-mode-based strategy rather than indiscriminate time-based servicing. Preventive, predictive, corrective, shutdown, opportunity and run-to-failure policies are selected according to function, failure behaviour, consequence, detectability and cost.

Asset criticality helps allocate attention, while FMEA and RCM reasoning connect failure modes to technically effective tasks. Job planning converts strategy into executable work orders containing scope, labour, materials, permits, quality criteria and feedback requirements. Scheduling must respect capacity, production access and resource constraints. Shutdowns require scope control and critical-path discipline; opportunity work must not compromise safety or restart.

Backlog control, maintenance standards, compliance records and ethical governance protect the organization from hidden risk. Measures such as MTBF, MTTR, availability, schedule compliance, repeat failures and overdue statutory work are useful only when definitions are stable and metrics are interpreted together. CMMS, sensors and AI can improve visibility and prediction, but they cannot replace engineering judgment, field verification or management of change. The ultimate test of a planned-maintenance program is sustained safe asset function at acceptable life-cycle cost.

Key Terms

Asset criticality: Relative importance of an asset based on risk and operational consequences.

Backlog: Approved or identified maintenance work not yet completed.

Condition-based maintenance: Maintenance initiated by evidence of actual asset condition.

Corrective maintenance: Maintenance performed to restore an identified defect or failure.

Failure mode: The specific manner in which an asset or component fails.

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

Job plan: A prepared description of work method, resources, safety and acceptance requirements.

Maintenance standard: Controlled technical instruction defining how maintenance is to be performed.

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

MTTR: Mean corrective repair time.

Opportunity maintenance: Work performed during an access or production window created for another purpose.

Planned corrective maintenance: Corrective work deliberately scheduled after a defect is identified.

Preventive maintenance: Predetermined maintenance intended to reduce failure probability or degradation.

RCM: Reliability-centred maintenance, a process for selecting policies by function, failure and consequence.

Run-to-failure: Deliberate operation until failure when risk and consequence are acceptable.

Schedule compliance: The proportion of scheduled work completed according to an agreed definition.

Shutdown maintenance: Concentrated maintenance performed during a planned production interruption.

Work order: Authorized record used to control, execute, document and close maintenance work.

Review Questions

Short Answer Questions

1. What distinguishes maintenance strategy from work planning?
2. Define opportunity maintenance and state one condition for its safe use.
3. Why should manufacturer recommendations be adapted to operating context?
4. What information should be recorded when a failure is investigated?
5. State two limitations of using a simple criticality score.

Long Answer Questions

1. Explain how FMEA or RCM reasoning can be used to design a preventive-maintenance program.
2. Discuss the roles of job planning, scheduling and work-order closure in maintenance performance.
3. Compare preventive, condition-based, corrective, shutdown and run-to-failure policies.
4. Explain how backlog, standards and compliance management influence operational risk.
5. Evaluate the benefits and limitations of CMMS, sensors and AI in planned maintenance.

Analytical / Application Questions

1. A critical pump has repeated seal failures despite monthly inspections. Develop an investigation and revised maintenance strategy, including operating, design and job-quality considerations.
2. A maintenance department has 1,600 scheduled craft hours and 1,800 available craft hours in a week, but 250 hours of statutory work are added unexpectedly. Analyze the capacity problem and propose a risk-based scheduling response.
3. Design a criticality-based maintenance policy for a duty/standby production system containing one safety-critical valve, two process pumps and several noncritical ventilation fans. Justify task types, evidence requirements and deferral rules.

Multiple Choice Questions with Answers

Question 1

Which statement best describes planned maintenance?

A. Maintenance performed only after catastrophic failure

B. A managed process of assessing, preparing, scheduling, executing and learning from maintenance work

C. Replacement of every component at the same calendar interval

D. A system concerned only with maintenance department cost

Correct Answer: B. A managed process of assessing, preparing, scheduling, executing and learning from maintenance work

Question 2

Which situation most strongly supports a condition-based maintenance task?

A. The failure is completely random and has no detectable degradation

B. A measurable condition provides adequate warning before functional failure

C. The component is inexpensive and noncritical

D. The task is required only because a supplier brochure recommends it

Correct Answer: B. A measurable condition provides adequate warning before functional failure

Question 3

In the simple risk-prioritization expression R = P × C, what do P and C represent?

A. Production and capacity

B. Probability category and consequence category

C. Planning and compliance

D. Parts and craft hours

Correct Answer: B. Probability category and consequence category

Question 4

An asset operates for 7,200 hours and has six failures. What is its MTBF?

A. 120 hours

B. 600 hours

C. 1,200 hours

D. 7,206 hours

Correct Answer: C. 1,200 hours

Question 5

Which item is primarily a work-planning output?

A. An unreviewed complaint from an operator

B. A job plan containing scope, resources, permits and acceptance criteria

C. A corporate sales forecast

D. A list of all assets without failure modes

Correct Answer: B. A job plan containing scope, resources, permits and acceptance criteria

Question 6

What is the principal purpose of a controlled work-order status model?

A. To ensure every job is classified as an emergency

B. To provide governance and clear ownership from request through closure

C. To eliminate the need for field verification

D. To increase the number of computer records regardless of work quality

Correct Answer: B. To provide governance and clear ownership from request through closure

Question 7

Which statement about backlog is most accurate?

A. All backlog is evidence of poor maintenance

B. Backlog should be reviewed for validity, risk, readiness and age

C. Backlog should contain only emergency work

D. Backlog can be ignored when schedule compliance is high

Correct Answer: B. Backlog should be reviewed for validity, risk, readiness and age

Question 8

A pump has MTBF of 1,200 hours and MTTR of 8 hours. Using A_i = MTBF ÷ (MTBF + MTTR), its approximate inherent availability is:

A. 93.40%

B. 98.34%

C. 99.34%

D. 99.93%

Correct Answer: C. 99.34%

Question 9

Which is an ethical maintenance practice?

A. Closing a work order without performing the work to protect a performance indicator

B. Bypassing hazardous-energy isolation when production is urgent

C. Escalating an unsafe condition and documenting the residual risk

D. Deleting repeated failures from the history to improve MTBF

Correct Answer: C. Escalating an unsafe condition and documenting the residual risk

Question 10

Why should AI recommendations in safety-critical maintenance remain subject to engineering review?

A. AI systems never use data

B. Models can produce false alarms, missed alarms or unreliable results when data and conditions change

C. Engineering review prevents all maintenance optimization

D. AI can only be used for financial accounting

Correct Answer: B. Models can produce false alarms, missed alarms or unreliable results when data and conditions change