Equipment Reliability and Rapid Changeover

Introduction

Lean manufacturing depends on the dependable flow of value. A line cannot deliver just-in-time production when equipment stops unexpectedly, produces defects, or requires lengthy changeovers. Equipment reliability and rapid changeover therefore connect two traditionally separate concerns: maintaining asset capability and increasing production flexibility. Total productive maintenance (TPM) develops the organizational system for preserving equipment effectiveness, while Single-Minute Exchange of Die (SMED) reduces the time required to switch equipment from one product or process condition to another.

The managerial challenge is not simply to maximize machine utilization. Excessive utilization can accelerate wear, encourage deferred maintenance, and reduce the capacity to respond to customer demand. The objective is lifecycle asset performance: safe, capable, available, maintainable, flexible equipment operated by competent people at an economically justified cost.

Learning Objectives

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

1. Explain the relationship among reliability, maintainability, availability, flexibility, and quality.
2. Design a TPM system involving autonomous, planned, improvement, and quality maintenance.
3. Calculate and interpret OEE, MTBF, MTTR, availability, and related measures.
4. Distinguish losses caused by breakdowns, speed reduction, minor stops, defects, and changeovers.
5. Apply SMED to separate internal and external setup work and redesign changeover activities.
6. Evaluate maintenance and changeover decisions using operational, financial, safety, and lifecycle criteria.

1. Reliability as a Lean Operations Capability

Reliability is the probability that an item performs its required function for a specified time under stated conditions. Reliability is therefore contextual: a pump may be reliable at its design flow and temperature but unreliable under overload, contamination, or inadequate lubrication.

Maintainability is the ease and speed with which an item can be restored to a specified operating condition. Design access, diagnostic information, standard tools, modular components, and technician competence all affect maintainability.

Availability is the proportion of required time during which equipment is capable of performing its intended function. It reflects both failures and restoration time. Flexibility is the ability of a process to produce different products, volumes, or sequences without disproportionate time, cost, or quality loss.

These concepts should not be confused:

A lean system treats equipment losses as losses of customer-serving capacity. However, maintenance decisions must preserve safety and product integrity rather than merely increase output.

2. Total Productive Maintenance

TPM is a company-wide approach to maximizing equipment effectiveness over its complete life cycle through the participation of operators, maintenance personnel, engineers, managers, and suppliers. Its distinctive feature is shared ownership: maintenance is not left exclusively to a specialist department, although specialist expertise remains essential.

A practical TPM architecture includes the following pillars:

• Autonomous maintenance: Operators perform inspection, cleaning, lubrication, tightening, basic adjustments, abnormality detection, and escalation.
• Planned maintenance: Maintenance work is scheduled using preventive, predictive, corrective, and reliability-centered principles.
• Focused improvement: Cross-functional teams remove chronic losses and recurring failure mechanisms.
• Quality maintenance: Equipment conditions are controlled so that defects are prevented rather than detected late.
• Early equipment management: Maintainability, cleanability, safety, capability, and changeover requirements are designed into new or modified assets.
• Training and skill development: People acquire the technical, diagnostic, and problem-solving competence required for stable operation.
• Safety, health, and environment: Maintenance and production systems prevent injury, exposure, contamination, and environmental harm.
• Administrative TPM: Planning, spare-parts control, work-order management, and information flows are improved because office processes can constrain technical work.

TPM is not equivalent to cleaning campaigns or a checklist program. Its maturity is shown by the quality of standards, the speed and accuracy of abnormality response, the elimination of root causes, and the evidence that equipment performance improves without transferring hidden work to operators.

Autonomous Maintenance

Autonomous maintenance is developed progressively. A typical sequence is: restore basic conditions; eliminate sources of contamination and difficult access; establish cleaning, lubrication, and inspection standards; conduct general inspection; standardize autonomous activities; and pursue autonomous improvement. The operator is not expected to repair complex electrical, software, or safety-critical systems without qualification. Instead, the operator becomes the first line of condition control and abnormality detection.

A good standard specifies what is checked, how it is checked, acceptable limits, frequency, tools, hazards, and escalation criteria. Visual controls should make abnormality visible without encouraging unsafe bypasses.

Planned and Predictive Maintenance

Planned maintenance includes:

• Corrective maintenance: Repair after a fault or failure. It may be rational for noncritical, inexpensive, or easily replaceable items.
• Preventive maintenance: Scheduled inspection, replacement, servicing, or restoration based on time, cycles, usage, or known deterioration.
• Predictive or condition-based maintenance: Intervention based on measured condition, such as vibration, temperature, oil debris, ultrasound, motor current, or performance trends.
• Proactive maintenance: Modification of design, operating practice, or environmental conditions to remove the cause of recurring failure.

The choice should follow risk and evidence. Replacing every component at a fixed interval is not automatically more reliable; premature replacement can introduce installation errors and waste. Conversely, run-to-failure is inappropriate for hazards, regulatory devices, critical bottlenecks, or failures that cause major consequential damage.

Reliability-centered maintenance (RCM) asks what functions an asset must perform, how it can fail functionally, what causes each failure, what consequences result, and which technically feasible task manages the risk. RCM encourages a mixture of condition-based, scheduled restoration, scheduled discard, failure-finding, redesign, and deliberate run-to-failure decisions.

3. Measuring Equipment Effectiveness

Availability, performance, and quality

A useful operational decomposition of OEE is:

where each factor is expressed as a decimal or percentage.

For a defined measurement period:

Here, planned production time is the scheduled time during which production is expected; operating time is planned production time minus downtime; ideal cycle time is the theoretically shortest cycle time per unit under specified conditions; total count includes good and defective units; and good count excludes units requiring rejection or rework. All time quantities must use the same unit, and the measurement boundary must be defined. Planned breaks may be excluded, whereas changeovers are normally included as planned production losses when the equipment is scheduled to produce.

An equivalent time-based form is:

This identity follows by substitution, provided the same counts and time definitions are used. OEE is a diagnostic measure, not a complete business-performance measure. It does not directly capture labor cost, energy, safety, inventory, customer value, or whether the ideal cycle time is realistic.

Worked example: OEE

A machine is scheduled for 480 minutes. A planned meal break of 30 minutes is excluded, giving 450 minutes of planned production time. Changeovers and downtime total 60 minutes, so operating time is 390 minutes. The ideal cycle time is 0.5 minutes per unit. The machine produces 700 units, of which 665 are good.

The loss pattern matters more than the single percentage. In this example, quality is the weakest factor, but the organization should examine the actual defect mechanisms rather than simply setting a higher OEE target.

Six Big Losses

The traditional six-loss model links TPM to OEE:

Losses may interact. A worn guide can cause both slow running and defects; a poorly designed changeover can cause prolonged adjustment and start-up scrap. Therefore, analysis should retain event-level detail instead of assigning every loss to only one category.

4. Reliability Measures and Failure Analysis

For repairable equipment, the mean time between failures (MTBF) is commonly calculated as:

MTBF has units of operating time per failure. It is meaningful only when “failure” and operating time are consistently defined. It should not be interpreted automatically as the guaranteed life of every item.

The mean time to repair (MTTR) is:

Repair time should be defined explicitly. It may include diagnosis, waiting for parts, access, repair, testing, and return to service, or only hands-on wrench time. For availability management, elapsed restoration time is usually more useful than hands-on time because queues and parts delays also reduce production capacity.

A simple inherent-availability approximation is:

This assumes a stable repairable system, statistically representative averages, and no other downtime such as logistics, preventive maintenance, or administrative delay. Operational or achieved availability may be lower when those factors are included.

Failure distributions and hazard

If failures occur at a constant rate and operating conditions remain stable, the exponential reliability model is:

where R(t) is the probability of survival to time t, λ is a constant failure rate with reciprocal-time units, and e is the base of natural logarithms. Under this special assumption, MTBF = 1÷λ. The model is often unsuitable for components experiencing wear-out, infant mortality, or changing operating conditions.

The Weibull model is more flexible:

where β is the shape parameter and η is the characteristic life in the same time units as t. A value of β < 1 suggests decreasing failure rate, β = 1 corresponds to the exponential case, and β > 1 suggests increasing failure rate. Parameter estimates require appropriate failure-time data and careful treatment of censored observations; a few failures do not justify confident distributional conclusions.

From symptoms to causes

Reliability improvement requires disciplined problem solving. A useful sequence is:

1. Define the functional failure and its operational consequence.
2. Establish the event timeline and verify data quality.
3. Contain immediate safety, quality, and delivery risk.
4. Identify physical, human, organizational, and latent causes.
5. Test hypotheses using evidence rather than opinion.
6. Implement a countermeasure that prevents recurrence.
7. Verify performance and update standards, training, and maintenance plans.

Tools include Pareto analysis, five-whys questioning, fishbone diagrams, fault-tree analysis, failure mode and effects analysis (FMEA), vibration or oil analysis, and controlled trials. FMEA risk-priority scoring can support discussion, but multiplying ordinal rankings does not produce a precise probability or economic risk estimate. Severity and legal or safety consequences must not be traded away because an occurrence score is low.

5. SMED and Rapid Changeover

A changeover is the complete transition from producing one acceptable product to producing the next acceptable product at the required rate. It includes preparation, tool and material movement, disassembly, replacement, adjustment, trial production, inspection, and stabilization.

SMED, developed from industrial setup-reduction practice, aims to reduce changeover time through a structured conversion of internal work to external work and the simplification of both.

• Internal setup activities can be performed only while the equipment is stopped.
• External setup activities can be performed while the equipment is still producing the previous product or after the next setup is prepared away from the machine.

The core SMED method is:

1. Observe and record the complete changeover, preferably with video and a time line.
2. Separate internal and external activities.
3. Convert internal activities to external wherever technically and safely possible.
4. Streamline remaining internal work using quick clamps, locating pins, standardized dimensions, parallel work, preset tooling, and error-proofing.
5. Eliminate adjustment through reference settings, digital recipes, poka-yoke, and first-piece verification.
6. Standardize, train, audit, and continue improvement.

The objective is not to make people move faster unsafely. It is to remove waiting, searching, transport, repeated adjustment, and unnecessary work. A rapid changeover that creates defects, ergonomic risk, or unreliable settings is not a lean improvement.

Changeover design principles

Preparation: Stage verified tools, materials, documents, and safety equipment before stopping the machine.

Modularity: Use interchangeable fixtures, common bases, standardized fasteners, and quick-release connections.

Reference settings: Record validated positions, torque values, recipes, and parameter limits. Digital instructions should identify revision status and prevent unauthorized changes.

Parallel work: Assign trained personnel to simultaneous tasks when the risk assessment and communication method support it.

Elimination of adjustment: Use hard stops, locating features, poka-yoke, automatic parameter loading, and design changes that make correct assembly repeatable.

First-off quality: Define the acceptance criteria and measurement method before the changeover. Quality approval must not become an unexplained queue.

Changeover time should be defined consistently, for example from the last good unit of Product A to the first good unit of Product B at standard rate. If a business uses a different boundary, it must document it. Otherwise, teams can appear to improve performance by excluding adjustment or stabilization losses.

Worked example: changeover economics

A line performs 20 changeovers per month. The current average changeover is 90 minutes. A SMED project reduces it to 45 minutes. The monthly capacity released is:

If the line produces one acceptable unit every 0.75 minutes during the released time, the theoretical additional capacity is:

This is capacity, not guaranteed sales or profit. The actual benefit depends on demand, material availability, labor, downstream constraints, quality yield, and whether the line can sustain its standard cycle time.

6. Integrating TPM, OEE, and SMED

TPM stabilizes the equipment; SMED increases its responsiveness. Their integration can be represented as a management loop:

Figure: An integrated improvement cycle linking OEE measurement, TPM, SMED, verification, and standardization.

A useful deployment sequence is to select a strategically important pilot asset, establish a credible loss baseline, restore basic conditions, implement operator and planned-maintenance standards, attack chronic losses, and then reduce changeover time. Results should be verified over enough production cycles to distinguish genuine improvement from short-term attention effects.

Maintenance and production metrics should be aligned. If production is rewarded only for output, it may defer maintenance or conceal minor stops. If maintenance is rewarded only for completed work orders, it may create unnecessary preventive tasks. Balanced indicators can include OEE by loss category, schedule compliance, emergency-work proportion, repeat failures, planned-versus-unplanned downtime, changeover distribution, first-off approval time, defect rate, safety observations, energy use, and cost per good unit.

7. Applied Case: Illustrative Packaging-Line Improvement

Background

A fictional food-packaging plant produces four package sizes on one filling and sealing line. Customer demand has become more variable, requiring smaller batches. The line reports 78% OEE, but delivery performance is deteriorating. Operators report frequent film jams and long adjustments after format changes.

Problem and analysis

A six-week loss study finds that the reported OEE uses inconsistent definitions: some teams exclude changeovers and others include them. After a common definition is adopted, the baseline is 71%. The principal losses are:

• 13% of planned production time in changeovers and first-off adjustment;
• 8% in minor stops caused by a misaligned film sensor;
• 5% in speed loss caused by worn guide rollers;
• 3% in start-up defects and seal failures;
• the remainder in ordinary breakdown and waiting losses.

A Pareto chart identifies the sensor bracket, guide-roller wear, and recipe-setting errors as recurring contributors. A review also finds that operators clean around the sensor but have no standard for checking alignment, while maintenance replaces rollers only after visible damage.

Approach

The cross-functional team takes the following actions:

1. Restores the sensor bracket and installs a poka-yoke locating feature.
2. Adds an alignment check to autonomous-maintenance standards.
3. Uses condition-based inspection of roller diameter and surface condition.
4. Pre-stages verified format parts on a labeled changeover cart.
5. Converts recipe preparation and tool cleaning to external work.
6. Replaces multiple fasteners with quick clamps and standard locating pins.
7. Creates validated digital recipes with controlled access and revision history.
8. Defines changeover completion as the first good unit at standard speed.
9. Trains operators and technicians together and audits the first ten changeovers.

Outcome

After eight weeks, average changeover time falls from 68 to 38 minutes, minor stops decrease, and start-up defects decline. The line’s measured OEE rises to 80%. The improvement is not attributed solely to SMED: basic equipment restoration, recipe governance, operator training, and measurement discipline contributed substantially. The plant also discovers that the extra capacity is useful only because production planning combines smaller orders and reduces finished-goods inventory.

Learning points

The case demonstrates five principles. First, a metric is useful only when its definitions are stable. Second, changeover improvement depends on equipment condition and process capability. Third, operator involvement is a technical control, not merely a cultural initiative. Fourth, capacity released by SMED must be connected to demand and scheduling decisions. Fifth, a short-term OEE increase is insufficient evidence of lifecycle improvement; repeat failures, safety, quality, energy, and maintenance burden must also be monitored.

8. Governance, Safety, and Contemporary Technology

Maintenance creates significant safety risks: stored energy, unexpected start-up, hot surfaces, chemicals, pressure, electrical exposure, and working at height. Lockout/tagout or the applicable legal isolation system must govern hazardous work. No OEE or changeover target justifies bypassing an interlock, disabling a guard, falsifying a maintenance record, or pressuring workers to work beyond competence.

Asset records should preserve traceability of inspections, parts, calibrations, software changes, and safety-device tests. Critical equipment may require statutory inspection, validated maintenance, or documented competence. Data governance matters because automated alarms and predictive models can produce false positives, miss rare failures, or encode poor historical maintenance practices.

Industrial Internet of Things (IIoT) sensors, edge computing, digital work instructions, computerised maintenance-management systems, and machine-learning models can improve visibility. Their value depends on sensor validity, reliable failure labels, cybersecurity, interoperability, and actionability. A predictive model that identifies a likely failure but provides no part, skill, access window, or decision rule does not create operational value. Human review remains necessary for safety-critical decisions and unusual conditions.

Sustainability should be included in lifecycle decisions. Longer equipment life may reduce embodied emissions, but inefficient or obsolete assets may consume more energy and generate more waste. Planned replacement, remanufacturing, lubrication control, leak elimination, and right-sized spare-parts inventories can improve both environmental and economic performance.

9. Limitations and Critical Evaluation

Several common practices require caution.

1. OEE targets can distort behavior. Teams may avoid short stops, run unnecessary production, or exclude difficult products. Use loss transparency and customer-relevant outcomes rather than a universal target.
2. MTBF can conceal failure definitions. Increasing the interval between recorded failures may reflect under-reporting or replacing a failed module without recording the underlying event.
3. Preventive maintenance can become ritualistic. Tasks should be evidence-based, risk-based, and periodically reviewed.
4. SMED can shift rather than remove work. External preparation may increase inventory, handling, or labor unless the whole system is considered.
5. Automation is not automatically lean. Automated changeover or predictive analytics may add complexity and maintenance burden. The total lifecycle case must be demonstrated.
6. Average values hide variation. Two lines may have the same average changeover time while one is stable and the other is highly unpredictable. Use distributions, percentiles, and event-level data where appropriate.

The strongest improvement systems combine engineering judgment, statistical evidence, worker knowledge, financial analysis, and ethical governance.

Chapter Summary

Equipment reliability and rapid changeover are complementary foundations of lean manufacturing. Reliability reduces unplanned interruption and protects quality; maintainability reduces restoration time; availability expresses readiness; and flexibility enables economical response to product variety. TPM creates a company-wide system involving autonomous maintenance, planned and condition-based maintenance, focused improvement, quality maintenance, early equipment management, training, safety, and administrative support.

OEE decomposes effectiveness into availability, performance, and quality, while the six-loss model identifies breakdowns, setup, minor stops, speed loss, defects, and reduced yield. MTBF, MTTR, availability models, and reliability distributions support diagnosis but require clear definitions and valid data. SMED separates internal from external setup work, then simplifies, parallelizes, standardizes, and error-proofs the remaining activities. Successful integration requires stable measurement, safe standards, cross-functional problem solving, and verification of quality, capacity, cost, and lifecycle effects.

The managerial lesson is that equipment improvement is a system decision. A faster changeover, higher OEE, or longer MTBF is valuable only when it improves safe customer-serving flow without creating hidden cost, quality risk, environmental harm, or unsustainable work.

Key Terms

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

Maintainability: Ease and speed with which equipment can be restored to service.

Availability: Proportion of required time that equipment is capable of operating.

Total Productive Maintenance: Company-wide system for maximizing equipment effectiveness across its lifecycle.

Autonomous maintenance: Operator-performed basic care, inspection, lubrication, and abnormality detection.

Planned maintenance: Maintenance scheduled or selected using time, usage, condition, risk, or failure knowledge.

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

Reliability-centered maintenance: Risk-based method for selecting technically appropriate maintenance policies.

OEE: Product of availability, performance, and quality.

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

MTTR: Mean time required to restore equipment through corrective repair.

Six Big Losses: Breakdown, setup, minor stop, speed, defect, and yield losses.

SMED: Method for reducing setup time by separating and simplifying internal and external work.

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

External setup: Setup activity possible while equipment is running or prepared away from the machine.

Poka-yoke: Error-proofing feature that prevents or immediately detects incorrect conditions.

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

Failure rate: Frequency or instantaneous tendency of failure under specified conditions.

Lifecycle asset performance: Integrated performance of an asset from design and acquisition through operation, maintenance, and retirement.

Review Questions

Short Answer Questions

1. Distinguish reliability, maintainability, and availability.
2. What are the internal and external elements of a changeover?
3. Name the three factors in OEE and define each one.
4. Why is operator involvement important in autonomous maintenance?
5. State two limitations of using MTBF as a performance measure.

Long Answer Questions

1. Explain the principal TPM pillars and show how they contribute to lifecycle asset performance.
2. Describe the SMED method and discuss how it can reduce adjustment and start-up defects as well as elapsed setup time.
3. Explain the six big losses and their relationship to OEE.
4. Compare corrective, preventive, predictive, and proactive maintenance, including circumstances in which each may be appropriate.
5. Critically evaluate the use of OEE targets in a multi-product manufacturing plant.

Analytical / Application Questions

1. A machine has 500 minutes of planned production time, 425 minutes of operating time, an ideal cycle time of 0.4 minutes, 900 total units, and 855 good units. Calculate availability, performance, quality, and OEE. Identify the weakest factor.
2. A line performs 16 monthly changeovers. A SMED project reduces average changeover time from 75 to 42 minutes. Calculate monthly minutes released and discuss three reasons why the theoretical capacity may not become actual sales.
3. A critical pump experiences repeated seal failures. Develop an investigation and maintenance strategy using functional failure definition, condition data, root-cause analysis, safety controls, and a decision about preventive versus condition-based intervention.

Multiple Choice Questions with Answers

Question 1

A machine has planned production time of 400 minutes and operating time of 360 minutes. Its availability is:

A. 80%

B. 85%

C. 90%

D. 95%

Correct Answer: C. 90%

Question 2

Which activity is normally classified as external setup work?

A. Removing a die while the machine is stopped

B. Adjusting the machine after the new tool is installed

C. Inspecting the first product at operating speed

D. Cleaning and verifying the next tool before the machine stops

Correct Answer: D. Cleaning and verifying the next tool before the machine stops

Question 3

OEE is calculated as:

A. Availability + Performance + Quality

B. Availability × Performance × Quality

C. MTBF ÷ MTTR

D. Good count ÷ Planned production time

Correct Answer: B. Availability × Performance × Quality

Question 4

Which TPM activity is most appropriate for a trained operator?

A. Rewriting safety interlock software

B. Performing high-voltage repairs without authorization

C. Cleaning, lubricating, and reporting abnormalities according to a standard

D. Changing a legally controlled pressure vessel

Correct Answer: C. Cleaning, lubricating, and reporting abnormalities according to a standard

Question 5

A repairable asset operates for 2,000 hours and experiences 5 functional failures. Its MTBF is:

A. 200 hours

B. 400 hours

C. 1,000 hours

D. 10,000 hours

Correct Answer: B. 400 hours

Question 6

Under the exponential reliability model, a constant failure rate implies:

A. A decreasing failure rate caused by burn-in

B. An increasing failure rate caused by wear-out

C. A failure rate that remains constant over time

D. No possibility of failure before a fixed life

Correct Answer: C. A failure rate that remains constant over time

Question 7

Which change most directly reduces internal setup time through error-proofing?

A. Staging tools in a distant storeroom

B. Installing locating pins and a poka-yoke fixture

C. Increasing the inspection queue after setup

D. Scheduling more preventive maintenance during production

Correct Answer: B. Installing locating pins and a poka-yoke fixture

Question 8

A line has MTBF of 90 hours and MTTR of 10 hours. Using the stated inherent-availability approximation, availability is:

A. 10%

B. 81%

C. 90%

D. 100%

Correct Answer: C. 90%

Question 9

Why can an OEE target create dysfunctional behavior?

A. OEE cannot be decomposed into losses

B. Teams may hide losses or run unnecessary production to improve the number

C. OEE includes safety and energy automatically

D. OEE measures only maintenance labor hours

Correct Answer: B. Teams may hide losses or run unnecessary production to improve the number

Question 10

Which maintenance approach uses vibration, temperature, or oil analysis to determine intervention timing?

A. Condition-based maintenance

B. Run-to-failure maintenance

C. Unplanned corrective maintenance

D. Administrative maintenance

Correct Answer: A. Condition-based maintenance