Manufacturing, Testing, Reliability and Asset Management

Introduction and context

Electrical-machine design is realized not on paper but through a controlled chain of material selection, manufacturing, inspection, testing, operation and maintenance. A theoretically satisfactory motor or generator may fail in service if laminations are damaged during stamping, winding insulation is mechanically stressed, a rotor is poorly balanced, a permanent magnet is incorrectly assembled, or impregnation leaves voids. Manufacturing therefore forms part of the design problem. Tolerances must be selected so that the machine can be produced economically while preserving electromagnetic, thermal, mechanical and insulation performance.

Reliability and asset management extend this responsibility over the machine's entire life. The designer must anticipate failure modes, define acceptance tests, identify measurable degradation, and support maintenance and replacement decisions. This chapter connects production processes with design quality, explains prototype and type testing, and introduces FMEA, condition monitoring, diagnostics and life-cycle asset management.

Learning objectives

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

1. Relate stamping, winding, casting, magnet assembly, impregnation and machining to machine tolerances and performance.
2. Distinguish routine, type, prototype and special tests.
3. Apply FMEA and basic reliability measures to electrical machines.
4. Select condition-monitoring methods for common failure mechanisms.
5. Interpret diagnostic evidence and formulate maintenance decisions.
6. Explain how quality assurance, safety, documentation and life-cycle cost influence asset management.

1. Manufacturing as an extension of design

A design drawing defines nominal dimensions, materials, fits, surface requirements and allowable variation. Manufacturing converts these nominal values into a population of real machines. The resulting variation affects air-gap length, magnetic reluctance, copper resistance, vibration, insulation life, noise and efficiency.

A tolerance is the permitted deviation from a specified value. A bilateral tolerance, such as 25.00 ± 0.05 mm, permits deviations in both directions. A unilateral tolerance permits deviation mainly on one side. Dimensional tolerances, geometric tolerances (such as concentricity and run-out), material tolerances and process tolerances should be specified according to function. Overly tight tolerances increase cost without necessarily improving reliability; loose tolerances can cause excessive air-gap variation, rotor-stator contact or unbalanced magnetic pull.

Quality control should combine:

• incoming inspection of steel, copper, insulation, magnets, castings and bearings;
• process control using approved tooling, settings and work instructions;
• in-process inspection of dimensions, burrs, winding placement and impregnation;
• final inspection and electrical, mechanical and functional tests;
• traceability of materials, operators, machines, measurements and rework.

Measurement systems themselves require calibration. A measurement result without known uncertainty is weak evidence for accepting a critical feature.

2. Principal manufacturing processes

2.1 Stator and rotor stamping

Magnetic cores are normally built from thin electrical-steel laminations. Stamping produces slots, teeth, back iron, bolt holes and the outside or inside diameter. Lamination thickness and insulation coating reduce eddy-current loss. The cutting operation must preserve the coating and avoid excessive burrs.

Important defects include burrs, distortion, incorrect slot geometry, damaged interlaminar coating and variation in stacking height. Burrs can short adjacent laminations, increase local flux concentration and reduce breakdown strength. Excessive punching clearance produces rough edges; inadequate clearance increases tool load and work hardening. Tool wear changes dimensions progressively, so first-piece and periodic checks are required.

Stacking factor is the ratio of the volume of magnetic steel to the gross core volume. If the gross stack length is L and the effective steel length is L_s, then:

Here k_s is dimensionless. It is less than unity because of insulation coatings, burrs and small gaps. The effective magnetic length is often approximated by L_s = k_sL, subject to the machine's magnetic-circuit model. A lower stacking factor increases the required core cross-sectional area or raises flux density for the same dimensions.

The design must also account for stamping-induced residual stress. Stress can increase core loss and reduce permeability, especially near teeth. Stress-relief annealing may improve magnetic properties but can affect coating and dimensional stability; its use must therefore be validated for the selected steel.

2.2 Winding manufacture

Windings may be formed from round or rectangular copper conductors and installed as distributed, concentrated, bar or form windings. Manufacturing concerns include conductor dimensions, insulation build, slot fill, coil pitch, phase separation, end-winding support, connection polarity and lead identification.

Slot fill factor is commonly represented as:

where A_Cu is the total bare-copper cross-sectional area in a slot and A_slot is the available slot area. The practical value is less than one because slot liners, conductor insulation, wedges, separators and voids occupy space. A high calculated fill factor may be impossible to manufacture or may prevent resin penetration and adequate heat transfer.

Winding tolerances influence resistance balance, leakage reactance, induced voltage, temperature rise and vibration. Poorly supported end turns may move under electromagnetic forces. In random windings, damaged enamel, inadequate interturn insulation and excessive insertion force are common risks. In form-wound coils, dimensional accuracy and bending radius are critical.

Electrical checks include winding resistance, phase balance, insulation resistance, polarization index where applicable, high-potential testing and surge comparison testing. A high-potential test is a controlled dielectric test and must follow the applicable standard and insulation system limits; it is not a substitute for process control.

2.3 Die casting and rotor manufacture

Aluminium or copper rotor cages may be die-cast into rotor slots. The process must fill bars and end rings completely while limiting porosity, cracks, cold shuts and distortion. Casting temperature, die temperature, injection pressure, fill time and die condition affect quality. Voids increase bar resistance and local heating; incomplete end-ring filling can produce severe current imbalance.

Rotor machining establishes shaft journals, end faces, fits and reference surfaces. Bearing seats require suitable dimensional and surface tolerances. Excessive interference can damage bearings; insufficient interference can permit fretting or creep. Rotor run-out and concentricity directly affect air-gap uniformity and vibration.

Dynamic balancing compensates for mass eccentricity. The residual unbalance permitted depends on rotor speed, mass, bearing arrangement and applicable balancing requirements. Balancing should be performed with the rotor in a configuration representative of service, including fan, coupling or retaining components where relevant.

2.4 Permanent-magnet assembly

Permanent magnets require correct grade, magnetization direction, polarity, temperature capability and mechanical retention. NdFeB magnets provide high energy density but may corrode and can suffer irreversible demagnetization at excessive temperature or opposing magnetic field. Ferrite magnets have lower energy density but generally better corrosion resistance and lower cost.

Assembly hazards include attraction forces, chipping, incorrect polarity, adhesive contamination and inadequate retention at high speed. Adhesive bonding requires controlled surface preparation, mix ratio, cure temperature and cure time. Mechanical retention should be designed so that adhesive is not the sole barrier against centrifugal forces unless its validated strength is sufficient. Rotor containment and overspeed testing are particularly important for high-speed machines.

2.5 Impregnation and insulation processing

Impregnation fills voids around conductors and binds the winding, improving heat transfer, mechanical rigidity and resistance to moisture and vibration. Common processes include varnish dip-and-bake, vacuum-pressure impregnation (VPI), resin-rich systems and resin transfer methods.

The critical variables are cleanliness, dryness, resin viscosity, vacuum level, pressure, temperature, impregnation time and cure schedule. Inadequate impregnation leaves voids that promote partial discharge, moisture accumulation and thermal ageing. Excessive cure temperature may damage insulation or magnets. The insulation system must be treated as a coordinated combination of conductor enamel, slot insulation, phase insulation, varnish or resin, wedges and mica or equivalent materials.

2.6 Machining, assembly and alignment

Machining determines shaft geometry, fits, keyways, shoulders, fan seats and mounting faces. Assembly controls bearing installation, axial positioning, air gap, seals, terminal connections and torque. Incorrect bearing mounting can introduce brinelling or preload. Contamination during assembly is a major source of premature bearing and insulation failure.

Air-gap eccentricity may be static, dynamic or a combination. It can increase unbalanced magnetic pull, noise, vibration and bearing load. Alignment of a coupled machine must consider offset, angular misalignment, thermal growth and foundation stiffness. Laser alignment, dial indicators and run-out measurements are useful, but the selected method must match the machine's accuracy requirement.

3. Prototype, routine and type testing

Testing has different purposes. Prototype tests establish whether a new design meets its intended requirements and reveal weaknesses. Type tests demonstrate that a design family satisfies specified performance and safety requirements. Routine tests screen every production unit for manufacturing defects. Special tests address particular customer, application or regulatory needs.

A test plan should state the objective, specimen configuration, instrumentation, calibration status, environmental conditions, sequence, acceptance criteria, safety controls, data-recording method and disposition of failures. Applicable standards may include IEC, IEEE, NEMA, ISO, ASTM or national regulations; the exact edition and machine category must be identified.

Typical tests include:

Insulation resistance varies with temperature and humidity. Comparisons should therefore be made at controlled or corrected conditions, using the manufacturer's and standard's procedures. A falling trend is often more informative than a single value.

A test uncertainty budget should include instrument accuracy, repeatability, environmental influence and measurement method. Acceptance limits should not be treated as infinitely precise boundaries. Safety is essential during high-voltage, overspeed and high-power tests: barriers, interlocks, grounding, emergency isolation and competent supervision are mandatory.

4. Reliability engineering for electrical machines

Reliability is the probability that an item performs its required function for a stated time under stated conditions. It is not identical to quality: quality concerns conformance at manufacture, whereas reliability concerns continued performance in service. Maintainability is the ability to restore function, and availability combines reliability with repair performance.

For a constant failure rate λ, the exponential reliability model is:

where R(t) is reliability at time t and λ is the constant failure rate in failures per unit time. The model assumes statistically independent failures with a constant hazard rate; it is not suitable for every life stage. The mean time to failure for this idealized non-repairable model is:

For a repairable asset with mean time between failures MTBF and mean time to repair MTTR, steady-state inherent availability is often approximated by:

This excludes many logistical delays unless they are included in the repair-time definition.

Electrical machines often exhibit a bathtub-shaped hazard pattern: early failures from manufacturing defects, a relatively stable useful-life period, and increasing failures from ageing and wear. Thermal ageing, insulation contamination, bearing fatigue, rotor-bar cracking, loosened connections, corrosion and cooling blockage are common mechanisms.

Failure Modes and Effects Analysis

FMEA is a structured, bottom-up method. For each function, the team identifies failure mode, local effect, system effect, cause, existing controls, detection method and recommended action. Traditional risk priority numbers may be calculated as:

where S is severity, O is occurrence and D is detectability, each assigned an agreed ordinal scale. Because ordinal scores are not true ratio measurements, RPN must not be interpreted as a precise probability. High-severity items should receive attention even if their numerical RPN is moderate. Modern practice supplements or replaces RPN with action priority, criticality analysis and explicit consequence assessment.

Example failure modes include:

5. Condition monitoring and diagnostics

Condition-based maintenance uses evidence of condition rather than fixed time alone. Monitoring should be selected from the failure mechanisms and consequences, not from the availability of a fashionable sensor.

• Temperature: RTDs, thermistors, thermocouples and infrared measurements identify overload, cooling loss and bearing problems. Surface temperature may not equal winding hot-spot temperature.
• Vibration: Overall velocity, acceleration, phase and spectral components help identify unbalance, misalignment, looseness, bearing defects and electromagnetic forces. Sensor mounting and speed information are essential.
• Electrical signatures: Current, voltage, negative-sequence components, power and torque estimates can reveal supply imbalance, eccentricity and rotor-bar problems. Load variation can obscure signatures.
• Insulation diagnostics: Insulation resistance, polarization index, dielectric loss, capacitance and partial-discharge measurements assess different aspects of insulation. Results require expert interpretation and consistent test conditions.
• Lubricant and debris analysis: Particle count, viscosity, contamination and wear metals can indicate bearing or gearbox deterioration.
• Acoustic and optical methods: Ultrasound can detect lubrication or discharge phenomena; thermal cameras locate abnormal heating; optical inspection can identify contamination and loosened parts.

A diagnostic system should establish a baseline after commissioning, trend values over time, define alert and alarm levels, and link alarms to an action plan. A single alarm is not automatically proof of a particular failure. Diagnosis should combine operating context, multiple measurements and physical inspection.

Data analytics and machine learning can detect anomalies in high-dimensional sensor data, but they require representative training data, stable sensor behaviour, explainable thresholds and protection against false alarms. AI should support—not replace—engineering judgement, safety procedures and validated protection systems. Cybersecurity, access control, data integrity and privacy are relevant when machines are connected to industrial networks.

6. Maintenance and asset-management decisions

Maintenance strategies include:

1. Corrective maintenance: repair after failure; appropriate where consequence and downtime cost are low.
2. Preventive maintenance: service at time or usage intervals; useful when degradation is age-related and predictable.
3. Condition-based maintenance: intervene when monitored condition indicates a defined risk.
4. Predictive maintenance: use trends and models to estimate remaining useful life; valuable when data quality and failure models are adequate.
5. Reliability-centred maintenance: select tasks according to functions, failure consequences, detectable conditions and operating context.

A rational decision compares total life-cycle cost rather than purchase price alone. A simplified life-cycle cost expression is:

where C₀ is acquisition and installation cost, C_E is energy cost, C_M is planned maintenance cost, C_F is expected failure and outage cost, C_D is disposal or decommissioning cost, and C_R is residual value. Costs must be evaluated over a stated period and discounted if financial comparison is required. Energy efficiency, spare-parts availability, safety and environmental effects may dominate the initial price.

Asset management requires an equipment register, criticality ranking, design and test records, operating history, work orders, spare-parts policy, obsolescence plan and disposal controls. Criticality can consider safety, environmental impact, production loss, repair time, redundancy and replacement lead time. A low-cost motor may be highly critical if its failure stops a continuous process.

7. Applied case: illustrative motor reliability improvement

Background

A 450 kW, 6.6 kV induction motor drives a process pump. It operates continuously in a humid environment. During two years, it experienced repeated bearing failures and one insulation trip. Production loss from an unplanned stop is substantial.

Problem and analysis

Maintenance records showed rising horizontal vibration before each bearing failure, but inspections were performed only every six months. The motor was aligned cold, although the coupling and foundation experienced significant thermal growth. An inspection also found blocked cooling passages and inconsistent grease quantities. The insulation trip followed a period of high humidity and contamination.

An interdisciplinary team performed FMEA. Bearing failure was rated high in consequence because it could damage the rotor and stop production. Occurrence was elevated because of alignment and lubrication evidence. Detectability was poor because no continuous vibration trend existed. Insulation contamination was treated separately, with moisture and cleaning controls.

Approach and solution

The team installed permanently mounted vibration and winding-temperature sensors, established a commissioning baseline, and recorded speed, load and ambient conditions. Alignment was corrected using thermal-growth targets. A controlled lubrication procedure specified grease type, quantity, interval and purge practice. Cooling passages were included in monthly inspection. Insulation resistance testing was performed during planned outages at comparable temperature and humidity, and the terminal box was improved against moisture ingress.

Alarm levels were selected from baseline trends, manufacturer guidance and engineering review rather than copied from an unrelated machine. A work instruction required confirmation of the physical cause before bearing replacement. Spare bearings and a qualified repair route were added to the asset plan.

Outcome

In this illustrative case, bearing vibration remained stable over the following operating period, and no insulation trip occurred after contamination and moisture controls were implemented. The principal improvement was not a single sensor; it was the integration of design information, manufacturing quality, alignment, lubrication, monitoring and documented response.

Learning points

1. Monitoring is valuable only when it leads to a defined decision.
2. Root causes may cross organizational boundaries: design, installation and maintenance can all contribute.
3. Trend data are more useful when operating conditions and baseline measurements are recorded.
4. High-consequence failures deserve attention even when their frequency is low.

8. Challenges, standards and professional responsibilities

Manufacturers must balance performance, cost, environmental impact and supply-chain risk. Replacing materials or suppliers can change magnetic properties, insulation compatibility, casting behaviour or magnet strength; a change-control process and validation testing are essential. Rework must be documented because concealed repairs can alter reliability.

Standards provide common methods but do not remove engineering responsibility. The applicable standard, machine rating, duty, installation category and customer specification must be reconciled. Test personnel must not bypass interlocks or reduce test severity without authorization. Records should be truthful, traceable and retained for the asset's required life.

Environmental considerations include energy efficiency, refrigerants or coolants, solvent and resin emissions, magnet and copper recovery, noise, end-of-life disposal and worker exposure. Ethical practice requires reporting nonconforming results, avoiding selective data presentation and communicating uncertainty. Safety-related decisions must prioritize people and process integrity over schedule or production pressure.

Best-practice workflow

A robust workflow is:

1. Translate functional requirements into critical-to-quality characteristics.
2. Design for manufacture, assembly, inspection, repair and safe disassembly.
3. Validate materials and processes with prototype and process-capability evidence.
4. Apply controlled drawings, specifications, calibration and traceability.
5. Use type tests to validate the design and routine tests to screen production.
6. Build an FMEA linked to inspection, monitoring and maintenance tasks.
7. Establish commissioning baselines and condition-based alarm actions.
8. Review field data, failures and near misses in a continuous-improvement loop.

Chapter Summary

Electrical-machine performance depends on the interaction of design, materials, manufacturing processes and service conditions. Stamping affects magnetic loss and geometry; winding manufacture affects insulation, resistance and thermal performance; casting and machining affect rotor strength, balance and air-gap uniformity; magnet assembly affects polarity, demagnetization resistance and high-speed retention; impregnation affects heat transfer, mechanical rigidity and insulation life.

Prototype and type tests validate design performance, while routine tests detect production defects. Reliability analysis defines the probability of successful operation, and FMEA organizes failure mechanisms, effects, controls and actions. Condition monitoring is most effective when it uses suitable sensors, a commissioning baseline, trend analysis and predetermined responses. Asset management integrates maintenance, safety, energy, spares, data, life-cycle cost and environmental responsibility. The central professional principle is that quality and reliability must be designed into the entire life cycle rather than inspected only at the end.

Key Terms

Tolerance: Permitted variation from a specified requirement.

Stacking factor: Ratio of effective magnetic steel length or volume to gross core length or volume.

Slot fill factor: Ratio of bare-conductor area to available slot area.

Run-out: Variation of a rotating surface from its ideal geometric reference during rotation.

Dynamic balancing: Correction of rotating mass distribution to limit vibration during rotation.

Impregnation: Process of filling and binding winding insulation with varnish or resin.

Prototype test: Test of a new or substantially changed design for validation.

Type test: Test demonstrating compliance of a design type with specified requirements.

Routine test: Test applied to each manufactured unit or defined production sample.

Reliability: Probability of performing a required function for a stated time and condition.

Failure rate: Frequency of failure expressed per unit time under stated assumptions.

Maintainability: Ability to restore an item to service effectively.

Availability: Proportion of time an asset is capable of performing its function.

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

Condition monitoring: Measurement and interpretation of asset condition during service.

Partial discharge: Localized electrical discharge that incompletely bridges insulation.

Criticality: Relative importance of an asset based on consequences of failure.

Life-cycle cost: Total cost of acquisition, operation, maintenance, failure and disposal.

Traceability: Ability to connect a product or result to its materials, processes and records.

Review Questions

Short Answer Questions

1. Why can stamping burrs increase both core loss and insulation risk?
2. Define prototype, type and routine tests.
3. State the assumptions of the constant-failure-rate reliability model.
4. Why is a commissioning baseline important in condition monitoring?
5. List four factors that influence impregnation quality.

Long Answer Questions

1. Explain how rotor machining, balancing and alignment influence vibration and bearing life.
2. Describe a complete test plan for a newly designed medium-voltage induction motor.
3. Explain how FMEA can connect manufacturing controls with maintenance actions.
4. Compare corrective, preventive, condition-based and predictive maintenance.
5. Discuss the technical, ethical and environmental responsibilities involved in asset management.

Analytical / Application Questions

1. A motor has a constant estimated failure rate of 0.00025 failures per operating hour. Using the exponential model, calculate its reliability after 2,000 hours and state whether the model's assumptions are realistic for the entire life of the motor.
2. A high-speed permanent-magnet rotor shows increased vibration after magnet bonding. Develop a diagnostic sequence that distinguishes imbalance, incorrect magnet placement, shaft run-out and adhesive failure.
3. A continuously operating pump motor has low purchase cost but high energy consumption and frequent unplanned outages. Propose an asset-management decision framework using life-cycle cost, criticality, monitoring and safety evidence.

Multiple Choice Questions with Answers

Question 1

Which stamping defect can electrically short adjacent laminations and increase local losses?

A. A clean shear edge

B. An excessive burr

C. A controlled stacking factor

D. A cured winding varnish

Correct Answer: B. An excessive burr

Question 2

The slot fill factor is best represented by which expression?

A. A_slot / A_Cu

B. A_Cu / A_slot

C. L_s / L

D. MTTR / MTBF

Correct Answer: B. A_Cu / A_slot

Question 3

Which test is normally applied to every production unit to detect manufacturing defects?

A. Routine test

B. Prototype test

C. Research test

D. Endurance type test

Correct Answer: A. Routine test

Question 4

For a constant failure rate λ, reliability at time t is:

A. R(t) = λt

B. R(t) = 1 − e^(−λt)

C. R(t) = e^(−λt)

D. R(t) = t/λ

Correct Answer: C. R(t) = e^(−λt)

Question 5

Which condition is most directly associated with partial-discharge testing?

A. Rotor balance

B. Insulation defects and voids

C. Shaft key dimensions

D. Grease quantity

Correct Answer: B. Insulation defects and voids

Question 6

In traditional FMEA scoring, RPN is calculated as:

A. Severity + occurrence + detectability

B. Severity × occurrence × detectability

C. Occurrence ÷ severity

D. Detectability − severity

Correct Answer: B. Severity × occurrence × detectability

Question 7

What is the principal purpose of a commissioning baseline?

A. To replace all acceptance criteria

B. To provide a reference for future trend changes

C. To eliminate calibration requirements

D. To determine the purchase price

Correct Answer: B. To provide a reference for future trend changes

Question 8

Which maintenance strategy uses measured asset condition to determine intervention?

A. Corrective maintenance

B. Calendar-only maintenance

C. Condition-based maintenance

D. Run-to-failure exclusively

Correct Answer: C. Condition-based maintenance

Question 9

Which factor is especially important when assembling high-speed permanent magnets?

A. Polarity and mechanical retention

B. Only the paint colour

C. Eliminating all dimensional inspection

D. Increasing air-gap contamination

Correct Answer: A. Polarity and mechanical retention

Question 10

In the simplified availability expression A = MTBF/(MTBF + MTTR), increasing MTTR while MTBF is fixed will generally:

A. Increase availability

B. Decrease availability

C. Have no effect

D. Make reliability equal to zero

Correct Answer: B. Decrease availability