Induction Machine Design

Introduction

Induction-machine design translates electrical ratings into a manufacturable electromagnetic, mechanical, and thermal structure. The designer must choose the stator bore, core length, air gap, slots, winding, rotor construction, conductor sizes, magnetic loading, electric loading, insulation, cooling system, and materials. These choices interact: reducing the air gap lowers magnetizing current but increases sensitivity to eccentricity, acoustic noise, cogging, and mechanical tolerances; increasing electric loading reduces machine size but raises copper loss and temperature.

Squirrel-cage machines dominate fixed-speed and inverter-fed industrial drives because they are robust and economical. Wound-rotor machines are used where external rotor resistance, high starting torque, or slip-power recovery is required.

Learning Objectives

After studying this chapter, students should be able to:

• determine the principal dimensions of an induction machine;
• select stator and rotor slot combinations and winding arrangements;
• size conductors, slots, teeth, yokes, and the air gap;
• estimate magnetizing, leakage, and resistance parameters from geometry;
• design squirrel-cage and wound rotors;
• predict torque–speed behavior, starting performance, and losses;
• explain crawling, cogging, harmonic torque, and slot-related noise;
• evaluate inverter-fed operation, thermal limits, and applicable standards.

1. Design Requirements and Constraints

A design specification normally includes rated output power, voltage, frequency, phase number, pole number or speed, duty cycle, efficiency class, enclosure, cooling method, insulation class, starting requirements, supply type, ambient temperature, altitude, and permissible noise and vibration.

For a machine with P poles supplied at frequency f, synchronous speed is

Slip is

where N_r is rotor speed. Rotor electrical frequency is f₂=sf.

Major design constraints include:

2. Main Dimensions and Output Equation

Let D be stator bore diameter, L effective core length, B_av average air-gap flux density, and A specific electric loading in ampere-conductors per metre of air-gap circumference.

The approximate flux per pole is

For a balanced m-phase winding, the phase emf is

where k_w=k_pk_d is the fundamental winding factor and T_ph is series turns per phase. Specific electric loading is

for one effective series path per phase; parallel paths must be treated using the current and turns of each path. Combining these expressions gives the apparent-power output equation

with

when S is in volt-amperes, dimensions are in metres, n_s is in revolutions per second, B_av is in teslas, and A is in A/m. For kilovolt-amperes, multiply C₀ by 10^-3. Numerically, 1.11\pi²\approx10.96.

This is a preliminary sizing relation, not a guarantee of rated mechanical output. Efficiency, power factor, saturation, temperature rise, and manufacturability must subsequently be verified.

Selecting the diameter-to-length ratio

Pole pitch is

The ratio L/\tau_p is selected by balancing cooling, end-winding length, rotor dynamics, power factor, and manufacturing constraints. A long machine can reduce diameter and end-winding copper but may require radial ventilation ducts and a stiffer shaft.

Worked Example 1: Preliminary dimensions

A 15 kW, 400 V, three-phase, 50 Hz, four-pole motor is expected to have efficiency 0.90 and power factor 0.84. Select k_w=0.955, B_av=0.48\ T, and A=30,000\ A/m.

The input apparent power is

Synchronous speed is 25 r/s. The output coefficient is

in kVA/(m³·r/s). Therefore,

A feasible initial choice is D=0.192\ m and L=0.143\ m, giving D²L=0.00527\ m³. These dimensions remain provisional until the winding, magnetic circuit, losses, and temperature rise are checked.

3. Stator Slots and Windings

3.1 Slot and winding selection

For an integral-slot winding,

where S_s is the number of stator slots and q is slots per pole per phase. Distributed windings reduce harmonic emf and improve the air-gap field. Fractional-slot windings can reduce cogging and end-winding length but require detailed harmonic analysis.

For a harmonic of order \nu, the winding factor is

with

where \alpha_c is the coil span in electrical radians and \beta is the electrical slot angle. Chording can suppress a selected harmonic while slightly reducing the fundamental winding factor.

3.2 Conductor and slot-dimension procedure

A practical stator-slot design follows these steps:

1. Calculate rated phase current from the specified connection and apparent power.
2. Determine turns per phase from the emf equation, allowing for winding voltage drop.
3. Select slot number and verify that conductors per slot are manufacturable.
4. Choose conductor area from permissible current density J.
5. Divide total copper area by slot fill factor to obtain gross slot area.
6. Select slot opening, wedge, insulation, top width, bottom width, and depth.
7. Verify tooth flux density, yoke flux density, leakage reactance, thermal path, and mechanical tooth strength.

For z_s conductors per slot and conductor area a_c, required gross slot area is approximately

where k_f is the copper fill factor, including the effect of conductor insulation, slot liner, wedge clearance, and manufacturing voids. Typical values depend strongly on wire shape and winding process; they must not be treated as universal constants.

For a trapezoidal active slot region,

where h_s is active depth and b_t and b_b are top and bottom widths. Wedge and opening regions are calculated separately when they do not contain active copper.

An approximate tooth flux-density check at a selected radius is

where \tau_s is slot pitch at that radius, b_t is tooth width, and k_fe is lamination stacking factor. Local finite-element analysis is preferable near tooth tips because flux is not uniformly distributed.

The stator-yoke depth may be estimated from

because each half of the yoke carries approximately half the pole flux. This relation assumes reasonably uniform yoke flux away from ventilation ducts and joints.

Figure: Principal radial regions and dimensions used in induction-machine design.

4. Air-Gap Design and Carter’s Coefficient

A small air gap reduces magnetizing current but increases sensitivity to manufacturing tolerances, bearing wear, rotor deflection, thermal expansion, eccentricity, and unbalanced magnetic pull. The selected gap must exceed the mechanical minimum under all operating conditions.

For conventional industrial motors of moderate size, a preliminary empirical estimate sometimes used is

where D and L are in metres and g_mm is in millimetres. This is only a starting estimate for ordinary radial-flux machines; high-speed, large-frame, unusual-enclosure, or precision-drive designs require mechanical and electromagnetic optimization.

Slot openings make the effective magnetic air gap greater than the physical gap. For one slotted surface, a commonly used approximate Carter coefficient is

where \tau_s is slot pitch, b₁ is slot-opening width, and all lengths use the same units. Thus the slot-opening correction is Kb₁=b₁²/(b₁+5g). The expression is an approximation for regular slot openings and a nearly uniform gap; detailed tooth-tip geometries should be analyzed by conformal mapping or finite elements.

If both surfaces are slotted,

where k_cs and k_cr are stator and rotor Carter coefficients.

For the preliminary motor, g\approx0.2+2\sqrt{0.192\times0.143}=0.531\ mm. Selecting 0.55 mm, with a stator slot pitch of 12.57 mm and opening of 3 mm, gives

A nearly closed rotor with k_cr=1.012 gives g_e\approx0.635\ mm.

5. Magnetic Circuit and Magnetizing Reactance

For a sinusoidal gap field, the peak flux density is approximately

The peak air-gap magnetomotive force for one effective gap is

where \mu₀=4\pi\times10^-7\ H/m. Tooth and yoke mmfs must be added using their nonlinear material magnetization curves. Leakage, joints, ducts, and local saturation also require correction.

For a three-phase sinusoidally distributed winding, the approximate rms magnetizing current is

where \hat F_m is the required peak resultant mmf per pole, including iron and effective-gap mmf. The corresponding unsaturated magnetizing reactance is approximately

when core-loss current and stator impedance drop are neglected. In practice, X_m is nonlinear and voltage dependent.

6. Leakage and Resistance Parameters from Geometry

6.1 Stator resistance

At operating temperature \theta, phase resistance is

for a_p identical parallel paths, provided T_ph and current distribution are defined consistently for each path. Here l_turn is mean turn length and a_c is conductor cross-sectional area. For copper,

with temperatures in degrees Celsius, using the customary 235 °C copper constant.

6.2 Slot leakage

For a simplified slot having an opening or neck of height h_o and width b_o, followed by a rectangular conductor region of height h_c and width b_c, a useful permeance coefficient is

The factor one-third follows from the approximately linear variation of enclosed slot current through a uniformly filled conductor region. For n_sp same-phase slots connected in series, each containing z_s conductors carrying phase current, the local slot-leakage inductance is approximately

This simplified expression excludes end, zigzag, skew, belt, and differential leakage. Double-layer slots containing different phase belts require a winding-specific energy calculation.

For L=0.143\ m, n_sp=16, z_s=13, h_o=1\ mm, b_o=3\ mm, h_c=20\ mm, and b_c=9\ mm, \lambda_s=1.074. Therefore,

End-winding leakage is estimated from end-turn geometry or three-dimensional field analysis. Differential leakage can be obtained by summing space-harmonic winding contributions. Production designs are normally calibrated against finite-element results and blocked-rotor tests.

7. Rotor Design

7.1 Squirrel-cage rotor

Rotor slot number should avoid combinations that reinforce cogging, synchronous harmonic torques, vibration, and noise. The stator and rotor slot counts should generally not be equal, and candidate combinations must be evaluated using harmonic permeance analysis rather than a single universal selection rule.

A rotor bar resistance is

where l_b includes effective bar length, A_b is bar area, and \rho_r is rotor-conductor resistivity at operating temperature.

For N_b equally spaced bars and p=P/2 pole pairs, the rms end-ring segment current for a sinusoidal fundamental bar-current distribution is approximately

where I_b is rms bar current. If R_e is the resistance of one segment between adjacent bars, total cage copper loss is

The factor 2N_b represents N_b segments in each of two end rings. Contact resistance and nonuniform current distribution may require additional allowance.

With 38 bars, two pole pairs, I_b=120\ A, R_b=35\ \mu\Omega, and R_e=2\ \mu\Omega,

The bar and ring losses are 19.15 W and 20.20 W, respectively, giving approximately 39.35 W under the stated current distribution.

Deep-bar and double-cage rotors exploit skin effect. At starting, rotor frequency is high, forcing current toward the upper bar region and increasing effective resistance. At normal low slip, current penetrates more deeply, lowering resistance and improving efficiency.

7.2 Wound rotor

A wound rotor normally carries a balanced three-phase insulated winding connected to slip rings, often in star. Rotor turns are selected from the required standstill rotor voltage, while conductor area is based on rotor current, starting duty, and thermal limits. Adequate insulation against switching and open-circuit voltages is essential.

External resistance increases starting torque and limits current. Referred to the stator, maximum torque occurs at

Thus, maximum torque occurs at standstill when total referred rotor resistance is set equal to the right-hand side. Resistance must then be progressively removed as speed rises if efficient continuous operation is required.

8. Equivalent Circuit and Torque–Speed Performance

Using the stator-side Thevenin equivalent, electromagnetic torque is

where V_th, R_th, and X_th are the Thevenin voltage and impedance seen from the air gap, R'₂ and X'₂ are stator-referred rotor parameters, and \omega_s is mechanical synchronous angular speed. The expression assumes balanced sinusoidal steady-state operation and constant parameters.

Maximum torque occurs at

Worked Example 2: Correct torque calculation

A four-pole, 50 Hz motor has V_th=236\ V, R_th=0.40\ \Omega, R'₂=0.25\ \Omega, and X_th+X'₂=1.50\ \Omega. Since \omega_s=2\pi(25)=157.08\ rad/s, at s=0.05,

At starting, s=1,

Thus, for these parameters, torque at 5% slip is substantially greater than starting torque. The maximum-torque slip is

so the motor develops its maximum torque at about 16.1% slip.

9. Losses, Efficiency, and Thermal Design

The major losses are:

• stator copper loss 3I₁²R₁;
• rotor copper loss, equal to sP_g under the fundamental equivalent-circuit model;
• core loss from hysteresis and eddy currents;
• mechanical loss from bearings and ventilation;
• stray-load loss caused by leakage fields, harmonics, circulating currents, and surface effects;
• inverter-related harmonic and common-mode losses where applicable.

Mechanical power developed is

where P_g is air-gap power. This relation applies to the fundamental balanced induction-machine model.

Thermal design uses a loss network or computational fluid dynamics model to predict winding hot-spot, rotor, bearing, and frame temperatures. Current density alone is not a sufficient thermal criterion. Cooling area, impregnation, contact resistance, ventilation, altitude, duty cycle, and inverter harmonics all matter.

10. Harmonics, Crawling, Cogging, and Noise

Three-phase windings commonly produce space harmonics of order 6k\pm1. The fifth harmonic rotates backward relative to the fundamental, while the seventh rotates forward at one-seventh of fundamental synchronous speed. A strong seventh-harmonic torque can create a stable low-speed operating point called crawling.

Cogging is magnetic locking caused by unfavorable alignment of stator and rotor teeth. It is most likely when slot combinations create strong coincident permeance patterns. Remedies include appropriate slot-number selection, rotor skew, fractional-slot windings, and optimized tooth tips.

If rotor bars are skewed by electrical angle \gamma, the skew factor for harmonic order \nu is

Skew suppresses slot harmonics and noise but slightly reduces fundamental torque and increases effective leakage. Electromagnetic force harmonics should also be checked against structural natural frequencies to avoid resonant vibration.

11. Inverter-Fed Machine Design

Variable-frequency drives change the machine’s electrical and thermal environment. Below base speed, approximately constant air-gap flux is maintained by controlling voltage-to-frequency ratio, with low-speed voltage compensation for stator resistance. Above base speed, voltage limitation produces field weakening and approximately constant-power operation.

Important design effects include:

• PWM harmonic copper and iron losses;
• reduced self-ventilation at low speed;
• high dv/dt stress on turn and ground-wall insulation;
• common-mode voltage and bearing currents;
• reflected-wave overvoltage in long motor cables;
• skin and proximity effects in conductors and rotor bars;
• torque ripple and acoustic excitation;
• possible overspeed during field weakening.

Mitigations include inverter-duty insulation, phase or sine-wave filters, insulated bearings, shaft-grounding brushes, optimized PWM, forced ventilation, temperature sensing, and derating. Electromagnetic finite-element models are increasingly coupled with thermal, structural, and drive simulations. Surrogate and machine-learning models may accelerate optimization, but training data, constraints, and final physics-based verification must be documented.

12. Applied Design Case: Preliminary 15 kW Motor

Background. This illustrative case develops the earlier 15 kW, 400 V, 50 Hz, four-pole industrial motor into a preliminary electromagnetic design.

Problem. The machine must provide reliable direct-on-line starting while remaining suitable for later inverter operation.

Analysis and approach. The output equation produced D=192\ mm and L=143\ mm. Forty-eight stator slots give q=4. The phase current is

For 104 turns per phase, conductors per slot are

The required flux per pole from the emf equation is 0.01047 Wb, corresponding to B_av=0.485\ T. At a current density of 4.5 A/mm², conductor area is 6.36 mm². With fill factor 0.42,

A preliminary 22 mm-deep trapezoidal copper region with widths 8 mm and 10 mm provides 198 mm². The bore slot pitch is 12.57 mm. A 4.57 mm tooth at the bore gives approximately

A yoke designed for 1.4 T requires approximately 27.5 mm radial depth. A 0.55 mm physical gap and Carter correction give an effective gap near 0.635 mm. Thirty-eight rotor bars avoid equality with the 48 stator slots; the final selection must still pass harmonic-force, starting, and noise analyses. Deep rotor bars may be adopted if direct-on-line starting torque is inadequate.

Outcome. The design is geometrically credible, but not yet released for manufacture. Required next steps are nonlinear finite-element analysis, equivalent-circuit calibration, starting-current verification, thermal analysis, rotor-stress assessment, inverter-insulation review, and prototype testing.

Learning points. Output equations provide only initial dimensions. Slot area, tooth saturation, air-gap tolerance, rotor harmonics, thermal limits, and starting behavior must be solved iteratively rather than independently.

13. Professional Practice, Standards, and Responsible Design

Designers should apply current editions of relevant standards, including IEC 60034 or applicable NEMA requirements for ratings, efficiency, temperature rise, testing, vibration, and enclosure classification. Hazardous-area motors require specific explosion-protection standards and certified construction.

Responsible design includes honest efficiency reporting, traceable loss measurements, safe overspeed margins, protection against excessive temperature and bearing failure, repairability, and responsible material selection. Efficiency improvements must be evaluated over the operating duty rather than only at one rating point. Optimization software should retain auditable assumptions, material data, tolerance limits, and verification records.

Chapter Summary

Induction-machine design begins with the output equation, which relates apparent power to bore diameter, core length, speed, magnetic loading, electric loading, and winding factor. Detailed design then establishes slots, turns, conductor area, teeth, yokes, air gap, and rotor geometry. Carter’s coefficient corrects the magnetic gap for slot openings, while magnetic-circuit and energy methods provide preliminary magnetizing and leakage parameters. The Thevenin torque equation predicts starting, breakdown, and operating torque. Rotor-bar and end-ring losses must include every bar and both sets of ring segments. Harmonics influence torque, crawling, cogging, vibration, and noise. Inverter-fed machines additionally require attention to PWM losses, insulation stress, bearing currents, cooling, and overspeed. A successful design is an iterative, multidisciplinary process validated by field analysis, thermal and mechanical calculations, standards-based tests, and prototype evidence.

Key Terms

Specific electric loading: Total rms ampere-conductors per metre of air-gap circumference.

Average air-gap flux density: Pole flux divided by pole area at the air gap.

Output coefficient: Constant relating apparent power to D²Ln_s.

Pole pitch: Circumferential distance corresponding to one magnetic pole.

Winding factor: Product of pitch and distribution factors for a specified harmonic.

Slot fill factor: Ratio of active conductor area to gross slot area.

Carter coefficient: Factor correcting physical air gap for slot-opening permeance effects.

Effective air gap: Magnetically equivalent gap after slotting corrections.

Magnetizing reactance: Equivalent-circuit reactance representing establishment of the main flux.

Slot leakage: Leakage flux confined principally to a slot and its opening.

Deep-bar effect: Frequency-dependent increase of effective rotor resistance caused by skin effect.

Slip: Per-unit difference between synchronous and rotor speeds.

Breakdown torque: Maximum steady electromagnetic torque.

Crawling: Stable operation near a subharmonic speed due to harmonic torque.

Cogging: Tooth-locking tendency caused by unfavorable slot alignment.

Rotor skew: Axial displacement of rotor slots used to reduce slot harmonics.

Stray-load loss: Additional load-dependent loss not represented by basic copper and core losses.

Field weakening: Operation above base speed with reduced air-gap flux because voltage is limited.

Review Questions

Short Answer Questions

1. Define specific electric loading and state its unit.
2. Why is the effective air gap larger than the physical air gap?
3. What causes the deep-bar effect?
4. Distinguish crawling from cogging.
5. Why can inverter operation increase bearing stress?

Long Answer Questions

1. Derive the induction-machine output equation from the emf and electric-loading equations.
2. Explain a complete procedure for selecting stator slots, turns, conductors, teeth, and yoke dimensions.
3. Discuss squirrel-cage rotor design, including bar resistance, end-ring current, skew, and slot selection.
4. Derive the Thevenin torque equation and explain the effects of rotor resistance on starting and breakdown torque.
5. Evaluate the electromagnetic, thermal, insulation, and mechanical implications of inverter-fed operation.

Analytical / Application Questions

1. Develop preliminary main dimensions for a 30 kW, 415 V, 50 Hz, six-pole motor using explicitly justified magnetic and electric loadings.
2. For a proposed stator and rotor slot combination, calculate relevant winding and skew factors and assess the risks of harmonic torque, crawling, and cogging.
3. Construct a loss and thermal budget for an inverter-fed motor operating continuously at 25% base speed and rated torque, and recommend cooling and insulation measures.

Multiple Choice Questions with Answers

Question 1

In the main output equation S=C₀D²Ln_s, with S in volt-amperes, which expression gives C₀?

A. 1.11\pi²k_wB_avA

B. 4.44\pi k_wB_av/A

C. 2\pi fB_avA

D. \pi²B_av/(k_wA)

Correct Answer: A. 1.11\pi²k_wB_avA

Question 2

For V_th=236\ V, R_th=0.40\ \Omega, R'₂=0.25\ \Omega, X_th+X'₂=1.50\ \Omega, \omega_s=157.08\ rad/s, and s=0.05, what is the electromagnetic torque?

A. 28.5 N·m

B. 99.5 N·m

C. 169.3 N·m

D. 272.5 N·m

Correct Answer: C. 169.3 N·m

Question 3

Using the stated Carter approximation, what slot-opening correction is subtracted from slot pitch in the denominator of k_c?

A. b₁g/(b₁+5g)

B. b₁²/(b₁+5g)

C. 5b₁g/(b₁+g)

D. g²/(b₁+5g)

Correct Answer: B. b₁²/(b₁+5g)

Question 4

A cage has 38 bars and two end rings. If each ring segment carries 364.5 A rms and has resistance 2 μΩ, approximately what is total end-ring loss?

A. 5.05 W

B. 10.10 W

C. 20.20 W

D. 40.40 W

Correct Answer: C. 20.20 W

Question 5

Which space harmonic is most directly associated with forward crawling near one-seventh of fundamental synchronous speed?

A. Third harmonic

B. Fifth harmonic

C. Seventh harmonic

D. Eleventh harmonic

Correct Answer: C. Seventh harmonic

Question 6

For R'₂=0.25\ \Omega, R_th=0.40\ \Omega, and X_th+X'₂=1.50\ \Omega, what is the maximum-torque slip?

A. 0.050

B. 0.161

C. 0.250

D. 0.400

Correct Answer: B. 0.161

Question 7

A slot contains 13 conductors of area 6.36 mm² and has a fill factor of 0.42. What gross slot area is required?

A. 82.7 mm²

B. 127 mm²

C. 197 mm²

D. 303 mm²

Correct Answer: C. 197 mm²

Question 8

Why may a self-ventilated inverter-fed motor overheat at low speed while producing rated torque?

A. Synchronous speed becomes independent of frequency

B. Fan cooling decreases while copper loss remains substantial

C. Rotor copper loss becomes identically zero

D. The air gap contracts electrically

Correct Answer: B. Fan cooling decreases while copper loss remains substantial

Question 9

For a wound-rotor motor, maximum torque occurs at standstill when total referred rotor resistance equals which quantity?

A. R_th

B. X'₂

C. \sqrt{R_th²+(X_th+X'₂)²}

D. V_th/I_m

Correct Answer: C. \sqrt{R_th²+(X_th+X'₂)²}

Question 10

If k_cs=1.142, k_cr=1.012, and the physical gap is 0.55 mm, what is the approximate effective gap?

A. 0.476 mm

B. 0.550 mm

C. 0.635 mm

D. 1.154 mm

Correct Answer: C. 0.635 mm