Equipment Reliability, Setup Reduction and Operational Flexibility

Introduction and context

Manufacturing performance is constrained not only by labour, materials and demand, but also by the dependability and adaptability of equipment. A technically capable machine creates no value when it is unavailable, produces defects, operates below its designed rate, or requires excessive time to change from one product to another. Total Productive Maintenance (TPM), Overall Equipment Effectiveness (OEE), autonomous maintenance, planned maintenance and Single-Minute Exchange of Die (SMED) provide an integrated response to these problems.

The central managerial challenge is to improve equipment performance without pursuing isolated local targets. Maximising utilisation may increase inventory, while excessive preventive maintenance may raise cost without reducing failure risk. Similarly, automation may increase speed but reduce flexibility or create a costly dependence on specialist support. Effective decisions therefore combine reliability engineering, lean thinking, lifecycle economics, workforce development and operational strategy.

Learning objectives

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

1. Explain TPM and distinguish its principal pillars.
2. Calculate and interpret OEE and its components.
3. Design autonomous and planned-maintenance practices.
4. Analyse the causes and cost of equipment losses.
5. Apply SMED to reduce setup time and increase flexibility.
6. Evaluate lifecycle cost and alternative levels of automation.
7. Critically assess implementation risks, governance and contemporary digital technologies.

1. Equipment effectiveness as a system concept

Reliability is the probability that an item performs its required function for a specified time under stated conditions. Maintainability is the probability that a failed item can be restored to service within a specified time. Availability reflects both characteristics and is also affected by logistics, staffing, utilities and operating procedures.

A useful distinction is between:

• Inherent availability, based mainly on equipment reliability and maintainability;
• Achieved availability, which includes preventive maintenance but excludes waiting for external resources; and
• Operational availability, which includes delays such as spare-parts shortages, permits, technicians and production scheduling.

For a repairable item with constant failure and repair rates, a simplified steady-state relationship is:

where MTBF is mean time between failures and MTTR is mean time to repair, both measured in the same time unit. The expression assumes stable operating conditions and that failure and repair behaviour can reasonably be represented by constant rates. It does not include planned maintenance, changeovers or waiting losses; those are included in broader operational measures.

Reliability is not identical to utilisation. A machine can be highly utilised yet unreliable if production is maintained through urgent repairs and overtime. Conversely, a reliable machine may be underutilised because demand is low. Lean improvement therefore begins by defining the relevant loss and the value stream in which it occurs.

The six major equipment losses

TPM traditionally groups equipment losses into six categories:

The categories matter because different losses require different countermeasures. Replacing a component may not solve a recurring minor stop caused by poor material presentation; increasing speed may worsen defects if process capability is inadequate.

2. Total Productive Maintenance

TPM is a company-wide approach to maximising equipment effectiveness throughout its life. It combines preventive and predictive maintenance with operator involvement, improvement teams, standardisation and management support. TPM is not simply a maintenance department programme. Operators are responsible for routine care and detection; maintainers provide technical expertise; engineers improve maintainability and design; managers remove systemic barriers.

Common TPM pillars include:

1. Focused improvement (kaizen): cross-functional teams eliminate chronic and sporadic losses.
2. Autonomous maintenance: operators clean, inspect, lubricate, tighten and identify abnormalities.
3. Planned maintenance: maintenance is scheduled according to risk, condition and equipment history.
4. Quality maintenance: equipment conditions are controlled so that defects are prevented rather than detected late.
5. Early equipment management: maintainability, safety, changeover and life-cycle cost are considered during design and acquisition.
6. Training and development: personnel gain the skills required for operation, diagnosis and improvement.
7. Safety, health and environment: equipment improvement must reduce exposure and environmental risk.
8. Office TPM: administrative processes supporting maintenance, purchasing, planning and data quality are improved.

The pillars should not be treated as independent projects. For example, autonomous inspection may reveal a recurring seal failure; planned maintenance can then revise the inspection interval, while focused improvement redesigns the seal arrangement.

Autonomous maintenance

Autonomous maintenance develops operator ownership of basic equipment conditions. A mature programme often progresses through the following stages:

1. Initial cleaning and identification of abnormalities.
2. Elimination of contamination sources and inaccessible points.
3. Creation of lubrication and inspection standards.
4. General inspection training.
5. Autonomous inspection using visual controls and check sheets.
6. Workplace organisation and standardisation.
7. Continuous improvement and self-management.

Cleaning is diagnostic, not cosmetic. Oil, dust, loose fasteners, unusual vibration, heat and leakage are signals of deterioration. However, autonomous maintenance must not transfer specialist tasks to untrained operators. Lockout/tagout, electrical work, pressure systems and legally controlled activities require authorised personnel.

Planned maintenance

Planned maintenance includes preventive, predictive and corrective activities. Preventive maintenance is performed at predetermined intervals; predictive maintenance uses condition information such as vibration, thermography, oil analysis or motor-current signatures; corrective maintenance restores or improves equipment after a fault or identified deficiency.

Time-based maintenance is appropriate when deterioration is age-related and an interval can be justified. Condition-based maintenance is preferable when measurable condition predicts failure. Run-to-failure may be rational for a low-cost, non-critical component whose failure has little safety, quality or delivery consequence. Maintenance policy should therefore be risk-based rather than ideologically uniform.

A practical maintenance strategy considers:

• consequence of failure to people, environment, quality, delivery and cost;
• failure mode and its predictability;
• redundancy and recovery time;
• spare-part lead time;
• regulatory obligations;
• skill and access requirements; and
• evidence from failure history.

Failure Mode and Effects Analysis (FMEA) supports this assessment by linking failure modes to effects, causes, controls and actions. Risk Priority Numbers can support discussion, but they should not be treated as precise probabilities, especially when ordinal ratings are multiplied. Criticality matrices, fault-tree analysis and actual failure data should supplement FMEA.

3. Overall Equipment Effectiveness

OEE measures how effectively scheduled production time is converted into conforming output at the ideal rate. Its conventional components are availability, performance and quality:

The three ratios are:

Here, planned production time is the time scheduled for production; operating time is planned production time less downtime losses; ideal cycle time is the fastest sustainable cycle for a conforming product under specified conditions; total count includes good and defective units; and good count excludes scrap and rework unless the organisation explicitly defines rework differently. All time quantities must use the same unit, and the ideal cycle time must be expressed in time per unit.

The formula is valid when the definitions are mutually consistent and the same time window and production population are used. OEE is not a universal measure of factory profitability. It can be distorted by the choice of planned time, ideal rate, product mix and treatment of minor stops. It should therefore be used for loss analysis and improvement, not as a simplistic ranking mechanism.

Worked example: OEE

A machine is scheduled for 480 minutes. Planned breaks total 30 minutes, so planned production time is 450 minutes. Breakdowns and adjustments total 45 minutes, giving operating time of 405 minutes. The ideal cycle time is 0.50 minutes per unit. The machine produces 720 units, of which 684 are good.

Therefore:

The machine’s effective ideal production equivalent is approximately 76% of the planned production opportunity. The result does not by itself identify the remedy: the loss tree shows that downtime, speed and quality each require investigation.

OEE and capacity decisions

OEE should be connected to takt time, demand and constraint analysis. If demand can be met after reducing changeovers, buying a faster machine may create excess capacity and capital cost. If a bottleneck has high OEE but still cannot meet demand, the issue may be insufficient scheduled time, excessive product variety or an upstream/downstream constraint. A lower OEE on a non-bottleneck may be economically acceptable if it does not impair customer service.

4. Setup reduction and SMED

A setup is the work required to change a process from producing one product or specification to another. Setup includes preparation, removal, mounting, adjustment, trial production, inspection and return to the previous condition. Long setups encourage large batches, increase inventory and make the system less responsive to demand.

SMED is a structured approach to reducing changeover time, often associated with the goal of a single-digit number of minutes. The numerical target is less important than the method:

1. Observe and document the current changeover.
2. Separate internal work, possible only when the machine is stopped, from external work, possible while it is running.
3. Convert internal work to external work where technically safe and feasible.
4. Simplify, standardise and parallelise the remaining tasks.
5. Eliminate adjustment through locating devices, presetting, poka-yoke and parameter recipes.
6. Sustain the new method through visual standards, training and measurement.

Examples include preparing tools before shutdown, using quick-release clamps, employing standardised die heights, presetting fixtures offline, using colour-coded connections and assigning parallel roles to operators and technicians.

Setup reduction affects more than time. It can improve safety by eliminating awkward lifting, reduce first-off defects, enable smaller economic batches and increase product-mix flexibility. Nevertheless, a shorter setup that creates quality instability or unsafe rushing is not an improvement.

Changeover measurement

Changeover time should have a precise start and finish definition. A useful operational definition is the elapsed time from the last conforming unit of product A to the first conforming unit of product B at the required rate. This definition includes trials and adjustments that are often hidden by reporting only machine downtime.

If a 600-minute shift contains 60 minutes of planned breaks, 40 minutes of changeovers and 20 minutes of breakdowns, the available production time after breaks is 540 minutes. Reducing changeover by 20 minutes increases available time to 560 minutes, a gain of 3.70% relative to 540 minutes, assuming no new losses:

The economic value depends on whether the additional time is used for profitable demand, bottleneck relief or reduced overtime.

5. Operational flexibility

Operational flexibility is the ability to respond economically to variation in volume, mix, routing, specification and timing. Equipment flexibility is only one component. Others include workforce skills, supplier responsiveness, information systems, tooling, material handling and production planning.

Important dimensions include:

• Mix flexibility: ability to produce different products;
• Volume flexibility: ability to vary output;
• Routing flexibility: ability to use alternative machines or paths;
• Expansion flexibility: ability to add capacity;
• Delivery flexibility: ability to alter sequence and timing; and
• Modification flexibility: ability to accommodate engineering or customer changes.

SMED increases mix flexibility, while modular tooling, common interfaces, cross-training and standard work increase resilience. However, flexibility has a cost. Excessive general-purpose capacity may be less productive than specialised equipment, and frequent switching can increase quality risk. The appropriate level depends on demand uncertainty, product life cycle, contribution margin and strategic importance.

6. Lifecycle cost and automation alternatives

Investment decisions should compare alternatives across the full life cycle rather than purchase price alone. Lifecycle cost includes acquisition, installation, training, energy, consumables, maintenance, software, downtime, quality losses, changeover effects, upgrades, decommissioning and disposal.

For an alternative with cash flows over periods 0 to n, net present value is:

where C₀ is the initial investment, CF_t is net cash flow in period t, r is the discount rate per period, and n is the number of periods. Cash flows must be expressed in consistent currency and time units. The analysis assumes the discount rate reflects the organisation’s opportunity cost and risk treatment; tax, inflation and residual value must be handled consistently.

A simpler payback period can be useful for screening but ignores cash flows after payback and the time value of money. NPV is generally more suitable for comparing mutually exclusive investments, provided strategic, capacity and risk factors are also considered.

Automation alternatives may include manual work, mechanised assistance, programmable automation, robotics, machine vision, autonomous material movement and integrated cyber-physical systems. Evaluation criteria should include:

Automation should not be used to conceal unstable processes. Standardise, simplify and stabilise first; then automate where the business case and risk assessment support it. Partial automation, assistive technology or modular cells may outperform a highly integrated system when demand and product design are uncertain.

7. Implementation architecture and governance

A robust programme begins with a baseline: downtime codes, changeover observations, quality data, maintenance history, safety incidents and demand requirements. Data definitions must be governed. For example, if one shift records a minor stop as downtime and another records it as reduced speed, trend comparisons become misleading.

A practical deployment sequence is:

1. Select a critical pilot asset using business and safety criteria.
2. Establish basic conditions and restore equipment to standard.
3. Create a loss tree and verify data at the workplace.
4. Train operators and maintainers in abnormality detection.
5. Address chronic losses through structured problem solving.
6. Introduce condition-based or risk-based maintenance where justified.
7. Apply SMED to high-frequency or strategically important changeovers.
8. Replicate proven standards and audit adherence.

Leadership should avoid using OEE or downtime data as a punitive individual-performance measure. Such use encourages concealment, inflated ideal rates and unsafe workarounds. Metrics should support learning and system improvement. Maintenance records, sensor data and worker observations also raise issues of access, privacy, cybersecurity and accountability. Connected equipment must be protected against unauthorised changes, and automated decisions should remain explainable and subject to human oversight.

8. Contemporary developments and limitations

Industrial Internet of Things systems can collect vibration, temperature, current, pressure and cycle data. Machine-learning models may detect anomalies or estimate remaining useful life. These tools are valuable when they lead to a defined maintenance decision, but prediction is not the same as action. Poor sensor placement, changing product mix, biased historical data and false alarms can reduce trust. A simple inspection standard may be more effective than a complex model for a low-criticality asset.

Digital twins can support commissioning, capacity analysis and maintenance simulation. Augmented-reality instructions may improve training and troubleshooting. Collaborative robots can assist with ergonomically difficult tasks, although risk assessment must address unexpected motion, safeguarding, human dependency and software updates.

Important limitations include:

• OEE can reward local optimisation and ignore customer value.
• Preventive maintenance can introduce maintenance-induced failures.
• Autonomous maintenance may fail if operators lack time, authority or training.
• SMED may shift work rather than eliminate it.
• Automation can reduce flexibility and increase technological lock-in.
• Lifecycle estimates are sensitive to uncertain demand, reliability and labour assumptions.
• Benchmark targets are not universal; equipment type, product mix and operating context matter.

Applied case: illustrative packaging-line improvement

Background

A fictional food-packaging plant operates one filling and sealing line for six product formats. Customer demand has shifted toward smaller, more frequent orders. The line is the production constraint, and supervisors report frequent overtime despite apparently adequate nameplate capacity.

Problem and analysis

A four-week study finds 480 scheduled minutes per shift, 30 minutes of planned breaks, average changeover of 42 minutes, 28 minutes of breakdowns, frequent minor jams and a first-pass quality rate of 96%. The team observes that operators search for tools after shutdown, while technicians perform adjustments that could have been prepared externally. OEE is calculated consistently and separated into availability, performance and quality rather than treated as a single blame metric.

A changeover video shows that 14 of the 42 minutes are externalisable: fixtures, labels and cleaning materials can be prepared before the stop. Quick-release clamps and preset guides can reduce adjustment. A visual inspection standard identifies loose guards and product residue as precursors to jams. A criticality review classifies the sealing-temperature sensor as high consequence and suitable for condition monitoring, while a low-cost non-critical guide is managed by run-to-failure with a spare available.

Approach and solution

The plant forms a cross-functional team of operators, maintenance, quality, engineering and planning. It establishes a standard changeover sequence, pre-stages materials, introduces parallel roles, installs locating pins and records the first conforming unit at the required rate. Operators receive training in cleaning, inspection and abnormality tagging; authorised technicians retain electrical and control-system work. Maintenance intervals are revised using failure evidence rather than calendar habit.

Outcome

After eight weeks, average changeover falls from 42 to 24 minutes, and first-pass quality rises from 96% to 98%. Breakdown time declines, but the team does not claim that every improvement is caused by SMED: changes in product mix and operator learning are recorded as possible confounders. Capacity is reassessed at the line and value-stream levels before any automation purchase. The business case supports a modest sensor investment rather than a fully automated replacement because demand remains uncertain and the existing line has adequate structural capacity.

Learning points

The case demonstrates that flexibility is created by a system of standard work, maintenance, skills and planning. It also shows why causal claims require a defined baseline, stable measurement and attention to external factors. A high OEE target alone would not have revealed that changeover preparation and minor jams were the strategic constraints.

Chapter Summary

TPM integrates equipment care, planned maintenance, focused improvement, quality, training and safety across the organisation. Reliability, maintainability and operational availability are related but distinct concepts. OEE decomposes effectiveness into availability, performance and quality, enabling loss-focused analysis when definitions and data are consistent. Autonomous maintenance builds operator capability without compromising legally controlled specialist work, while planned maintenance should be based on failure consequences, predictability and evidence.

SMED improves mix and response flexibility by separating internal and external setup work, simplifying the remaining tasks and stabilising the first conforming output. Lifecycle decisions should include acquisition, operation, maintenance, quality, energy, risk and disposal costs. Automation is most effective after processes are standardised and its technical, economic, human and cybersecurity implications have been assessed. Ultimately, equipment improvement is a socio-technical and strategic activity, not merely a maintenance project.

Key Terms

Reliability: Probability that equipment performs its required function for a specified time and condition.

Maintainability: Probability that equipment can be restored within a specified time.

Operational availability: Proportion of required operating time during which equipment is ready for productive use, including logistical delays.

MTBF: Mean time between failures of a repairable item.

MTTR: Mean time required to restore a failed item.

TPM: Company-wide system for maximising equipment effectiveness across its life cycle.

Autonomous maintenance: Operator-led routine care, inspection and abnormality detection.

Planned maintenance: Preventive, predictive and corrective maintenance organised in advance.

OEE: Product of availability, performance and quality.

Availability: Operating time divided by planned production time.

Performance: Actual ideal-cycle production equivalent divided by operating time.

Quality: Good output divided by total output.

Six major losses: Breakdown, setup, minor stop, reduced speed, defects and startup losses.

SMED: Method for reducing changeover time through internal/external separation and simplification.

Internal setup: Setup work possible only while equipment is stopped.

External setup: Setup work possible while equipment is operating.

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

Lifecycle cost: Total cost of owning, operating, maintaining and disposing of an asset.

Operational flexibility: Ability to respond economically to variation in volume, mix, routing or timing.

Predictive maintenance: Maintenance triggered or informed by measured equipment condition.

Review Questions

5 Short Answer Questions

1. Distinguish reliability, maintainability and operational availability.
2. Name the six major equipment losses and give one example of each.
3. What is the purpose of autonomous maintenance?
4. Explain the difference between internal and external setup work.
5. Why should OEE not be used as a standalone measure of factory profitability?

5 Long Answer Questions

1. Explain the TPM pillars and show how they reinforce one another.
2. Derive the interpretation of OEE components and discuss the data conditions required for valid calculation.
3. Describe a complete SMED implementation for a high-mix production process.
4. Compare time-based, condition-based and run-to-failure maintenance policies.
5. Discuss the technical, economic, human and governance factors in an automation decision.

3 Analytical / Application Questions

1. A machine has 600 scheduled minutes, 40 minutes of planned breaks, 50 minutes of downtime, an ideal cycle time of 0.4 minutes per unit, 1,150 total units and 1,104 good units. Calculate availability, performance, quality and OEE, and identify the largest numerical loss.
2. A changeover takes 55 minutes, of which 18 minutes can be performed before shutdown. Propose a SMED plan and calculate the percentage reduction if the remaining internal work is reduced by a further 12 minutes.
3. A plant is considering a highly automated replacement for a flexible semi-automatic cell. Develop a decision framework that incorporates NPV, demand uncertainty, maintainability, skills, safety, cybersecurity and product-mix flexibility.

Multiple Choice Questions with Answers

Question 1

Which expression correctly represents conventional OEE?

A. Availability + Performance + Quality

B. Availability × Performance × Quality

C. MTBF ÷ MTTR

D. Good Count ÷ Planned Production Time

Correct Answer: B. Availability × Performance × Quality

Question 2

A machine has 450 minutes of planned production time and operates for 405 minutes. What is its availability?

A. 80%

B. 85%

C. 90%

D. 95%

Correct Answer: C. 90%

Question 3

Which activity is an example of external setup work?

A. Removing a die after the machine stops

B. Adjusting the machine while stopped

C. Preparing and checking tools before shutdown

D. Producing trial parts after changeover

Correct Answer: C. Preparing and checking tools before shutdown

Question 4

What is the primary purpose of autonomous maintenance?

A. To transfer all maintenance work to operators

B. To develop routine care and early abnormality detection by operators

C. To eliminate the maintenance department

D. To maximise machine speed regardless of quality

Correct Answer: B. To develop routine care and early abnormality detection by operators

Question 5

Which maintenance policy is most appropriate when condition measurements reliably predict failure?

A. Condition-based maintenance

B. Uncontrolled run-to-failure

C. Random replacement

D. No maintenance

Correct Answer: A. Condition-based maintenance

Question 6

A process produces 800 units, of which 760 are good. What is its quality rate?

A. 90%

B. 92.5%

C. 95%

D. 97.5%

Correct Answer: C. 95%

Question 7

Which statement about lifecycle cost is correct?

A. It is limited to the purchase price

B. It includes only labour cost

C. It includes acquisition, operation, maintenance and disposal-related costs

D. It excludes energy and downtime

Correct Answer: C. It includes acquisition, operation, maintenance and disposal-related costs

Question 8

Why can a high OEE target produce undesirable behaviour?

A. It always reduces production capacity

B. It may encourage concealment of losses or unsafe workarounds

C. It prevents measurement of quality

D. It eliminates the need for standards

Correct Answer: B. It may encourage concealment of losses or unsafe workarounds

Question 9

What is a central SMED principle?

A. Increase batch size to avoid all changeovers

B. Separate internal and external setup activities

C. Perform all preparation after shutdown

D. Replace every machine with automation

Correct Answer: B. Separate internal and external setup activities

Question 10

Which statement best describes a sound automation decision?

A. Automate before stabilising the process

B. Choose the highest-speed technology regardless of flexibility

C. Compare technical, lifecycle, human, strategic and cybersecurity factors

D. Use purchase price as the sole selection criterion

Correct Answer: C. Compare technical, lifecycle, human, strategic and cybersecurity factors