Transformer and Inductor Design

Introduction

Transformers transfer electrical energy between circuits through mutual magnetic flux, while inductors store and release magnetic energy. Their design requires simultaneous satisfaction of electromagnetic, thermal, dielectric, mechanical, manufacturing, efficiency and cost constraints. A design that meets the voltage ratio may still fail because of excessive temperature rise, insulation stress, acoustic noise, harmonic heating or short-circuit force.

At power frequency, laminated electrical steel is normally used to restrict eddy currents. At higher frequencies, ferrites, powdered alloys, amorphous metals and nanocrystalline materials become attractive. The design process is iterative: select a tentative core, calculate windings and losses, assess regulation and temperature, verify insulation and fault strength, and then revise dimensions or materials.

Learning Objectives

After studying this chapter, a student should be able to:

• derive transformer volts-per-turn and area-product equations;
• select preliminary core and window dimensions from rating;
• design single- and three-phase windings;
• estimate losses, leakage reactance, regulation and efficiency;
• construct convection–radiation and multi-node thermal models;
• coordinate insulation with operating and impulse stresses;
• evaluate short-circuit forces using correct field directions;
• design gapped inductors from inductance, energy and saturation requirements;
• account for harmonics and high-frequency winding effects; and
• integrate electromagnetic, thermal and mechanical requirements in a practical design.

1. Magnetic-Circuit Foundations

For a magnetic path divided into sections, the magnetomotive-force balance is

where N is turns, I is current in amperes, H_k is magnetic-field intensity in A/m and l_k is path length in metres. For an approximately linear section,

Here ℜ is reluctance in A/Wb, μ is permeability in H/m, A is cross-sectional area in m², Φ is flux in webers and B is flux density in teslas. Ferromagnetic permeability is nonlinear; therefore constant-permeability calculations are preliminary approximations rather than exact saturation models.

Faraday’s law gives the induced winding voltage:

For sinusoidal flux with peak density B_m, rms induced voltage is

where f is frequency in hertz and A_c is effective core area in m². The equation assumes sinusoidal flux, negligible flux harmonics and a uniform effective cross-section. Consequently,

The selected integer number of turns must be checked by recalculating the actual flux density.

2. Core Materials and Geometries

Core loss consists mainly of hysteresis and eddy-current components. Manufacturer loss data should be used at the intended flux waveform, frequency and temperature. A commonly fitted model is

where P_c is core loss in watts, m_c is core mass in kilograms, and the empirical coefficients apply only to the tested material and stated waveform range. Extrapolation outside that range is unreliable.

Common geometries include core-type and shell-type laminated structures, toroids, E–I cores, U cores and planar cores. Core-type construction facilitates winding and insulation; shell-type construction can reduce mean turn length and provide mechanical support. Toroids have low external field but can be difficult to wind and insulate.

3. Rating-to-Core Selection: The Area-Product Method

The area product connects transformer rating to magnetic and electric loading. Let A_w be the geometrical window area available to both windings and k_u the window utilization factor, including conductor insulation, interlayer insulation, ducts, margins and unused space. With equal allowable current density J in both windings,

For an approximately ideal single-phase transformer of apparent-power rating S,

Using E = 4.44fNBA_c gives

Therefore,

and the required area product is

The equation is dimensionally consistent: fB_mJA_p has units of VA. It is a preliminary single-phase output equation under sinusoidal excitation, approximately ideal power transfer, common current density and the stated window-area convention. Three-phase cores and manufacturer catalogues may define window area differently; their topology-specific output coefficient must therefore be used rather than applying 2.22 blindly.

Practical core-selection procedure

1. Specify rating, frequency, voltage ratio, ambient temperature, cooling method and permitted temperature rise.
2. Choose provisional B_m from material loss and saturation data.
3. Choose J from thermal conditions; natural-cooled power-frequency windings often use lower current density than forced-cooled or high-frequency designs.
4. Estimate k_u, commonly reduced by high-voltage insulation, ducts or round-wire packing.
5. Calculate A_p.
6. Select a standard core with adequate A_cA_w, allowing manufacturing margin.
7. calculate integer turns and conductor areas;
8. verify actual fill, losses, temperature, regulation and fault forces;
9. iterate dimensions if any constraint is violated.

Worked Example 1: Preliminary single-phase design

A 10 kVA, 230/115 V, 50 Hz transformer is to use B_m = 1.50 T, J = 2.5 A/mm² = 2.5×10⁶ A/m² and k_u = 0.30.

Suppose a standard core provides A_c = 60 cm² = 0.0060 m² and A_w = 140 cm² = 0.0140 m². Its area product is 840 cm⁴ and is therefore provisionally adequate.

Choose 116 primary turns. The actual peak density is

For approximately 115 V, choose 58 secondary turns. Rated currents are 43.48 A and 86.96 A. Required bare conductor areas at the selected current density are 17.39 mm² and 34.78 mm². The idealized occupied window area is

Available copper area is k_uA_w = 0.30×14000 = 4200 mm². The fit is close, so detailed insulation and duct layout may require a slightly larger window.

4. Three-Phase Transformer Design

For a balanced three-phase transformer,

For a star connection, V_φ = V_L/√3 and I_φ = I_L. For a delta connection, V_φ = V_L and I_φ = I_L/√3. Turns must always be calculated from phase voltage, not automatically from line voltage.

A three-limb core carries three phase fluxes whose instantaneous sum is approximately zero under balanced sinusoidal operation. Core cross-section is selected per limb, while total window allocation must accommodate all coil sides, insulation and cooling ducts. Detailed designs use a topology-specific output coefficient or finite-element/window layout analysis.

Connections influence harmonics and system behavior. A delta winding provides a closed path for triplen harmonic currents and can stabilize phase voltages under some unbalanced conditions. A star neutral permits zero-sequence current only if the external system and core magnetic return path support it. Vector group, phase displacement, grounding and parallel-operation requirements must be established before winding design.

5. Windings, Resistance and Losses

The preliminary conductor area is

Copper resistance at temperature T is

where l is conductor length. For copper over a moderate temperature range,

with α₂₀ ≈ 0.00393/°C. Load loss includes dc copper loss and stray loss:

Here R_ac includes skin and proximity effects. Transposed strands, continuously transposed conductors, foil, litz wire and sectionalized windings may reduce circulating and proximity losses when correctly applied.

The skin depth in a nonmagnetic conductor is

where ω = 2πf. Conductor thickness comparable with or greater than several skin depths can cause substantial ac resistance, especially in multilayer windings exposed to strong leakage fields.

6. Leakage Flux, Reactance and Regulation

Not all flux links both windings. Leakage flux closes through winding gaps, ducts, air and nearby structural parts. A physically complete leakage-flux path is closed, as required by ∇·B = 0.

Figure: Both main and leakage magnetic flux are represented by closed, directed paths.

Leakage inductance can be obtained generally from magnetic-field energy:

Thus,

This energy method is suitable for analytical field models and finite-element calculations. Leakage generally increases with radial build and separation between primary and secondary windings. Interleaving reduces leakage but may increase capacitance and complicate high-voltage insulation.

Referred to one side, the approximate per-unit voltage regulation at load power-factor angle φ is

for lagging current. For leading power factor, the reactance term changes sign. The approximation assumes a relatively small voltage drop. A more accurate result is obtained from the phasor magnitude of V₂ + I₂(R_eq+jX_eq).

7. Efficiency

Transformer efficiency is

At a fractional load x, approximately constant voltage and sinusoidal excitation,

Maximum efficiency occurs approximately when variable load loss equals constant core loss. Distribution transformers may be optimized for all-day energy efficiency rather than rated-load efficiency because they remain energized continuously.

8. Thermal and Cooling Design

Permissible loading is often limited by insulation temperature rather than magnetic capability. Cooling may be natural-air, forced-air, oil-natural, oil-forced, water-assisted or based on heat plates and cold plates.

For a surface at temperature T_s exposed to ambient air and large surroundings at T_a, heat rejection is approximately

where h_c is the convection coefficient in W/(m²·K), ε is emissivity, and σ = 5.670×10^-8 W/(m²·K⁴). Absolute temperatures must be used in the radiation term.

Radiation may be linearized about the operating temperatures:

The surface-to-ambient resistance is then approximately

Conduction resistance through a uniform layer is R_cond=l/(kA). Contact resistance, ducts and anisotropic winding conductivity should be represented separately where significant.

Multi-node thermal network

A two-node model distinguishes winding temperature T_w from core temperature T_c. If R_wa, R_ca and R_wc are winding-to-ambient, core-to-ambient and winding-to-core resistances, steady-state balances are

The resistances may combine conduction, convection and linearized radiation. Because convection and radiation depend on temperature, practical calculations iterate until the assumed and calculated temperatures agree.

Worked Example 2: Two-node temperature calculation

A dry transformer has winding loss 180 W and core loss 70 W. At the estimated operating temperature, R_wa=0.30 K/W, R_ca=0.55 K/W, R_wc=0.12 K/W and T_a=40°C. Let x=T_w-T_a and y=T_c-T_a. Then

Solving gives x=49.2 K and y=47.3 K. Therefore,

The thermal designer must compare hot-spot—not merely average—temperature with the insulation-system limit. Transient overload analysis additionally requires thermal capacitances and differential equations of the form C_thdT/dt = P-∑(T-T_j)/R_th.

9. Insulation Coordination

Insulation design addresses continuous voltage, temporary overvoltage, switching impulses, lightning impulses, partial discharge, pollution and thermal ageing. Relevant quantities include:

• turn-to-turn, layer-to-layer and winding-to-winding voltage;
• clearance through air and creepage along surfaces;
• insulation thickness and dielectric strength;
• impulse distribution caused by winding capacitance;
• partial-discharge inception voltage;
• thermal class and expected life; and
• altitude, humidity, contamination and manufacturing tolerances.

A basic insulation level or impulse withstand specification must be coordinated with system arresters and grounding. Clearances and creepage distances must follow the applicable transformer and insulation-coordination standards; a single universal distance is unsafe because requirements depend on voltage, environment, material group and altitude.

Electrostatic shields can reduce capacitive coupling but must be grounded correctly and must not form a short-circuited turn. Quality control includes drying, impregnation, hipot testing, induced-voltage testing, partial-discharge testing where required, and traceability of insulation materials.

10. Short-Circuit Mechanical Strength

Fault current can reach many times rated current before protection clears it. Electromagnetic body-force density is

where f is in N/m³. The direction must be obtained from the vector cross-product:

• circumferential winding current interacting with axial leakage flux produces a radial force, creating inward or outward pressure and hoop stress;
• circumferential winding current interacting with radial leakage flux, particularly near winding ends, produces an axial force.

For an idealized magnetic field, magnetic pressure is approximately

This is not a universal substitute for field integration, but it is useful where the field is locally tangential to a boundary. For a thin cylindrical winding under uniform radial pressure p, mean radius r and effective radial structural thickness t, the membrane estimate is

Axial forces are obtained by integrating the appropriate J×B force density or Maxwell stress over the winding. Clamping systems must withstand axial compression, conductor buckling, spacer crushing and dynamic motion. Since force is approximately proportional to current squared, modest increases in prospective fault current can cause large mechanical demands.

11. Inductor and Reactor Design

An inductor stores magnetic energy:

Its inductance is approximately

When a discrete air gap dominates reluctance and fringing is initially neglected,

where g is total effective gap length. Flux density follows from flux linkage:

Thus a saturation constraint at peak current requires

After choosing integer turns, the approximate gap is

These equations assume the gap dominates core reluctance. Fringing increases effective gap area and can heat nearby conductors; detailed designs use corrected reluctance or field simulation.

Worked Example 3: Gapped inductor

Design a 200 μH inductor for 10 A peak using A_c=2.5 cm²=2.5×10^-4m² and B_max=0.25T.

With 32 turns,

Stored energy at peak current is

The design still requires checks of fringing, copper loss at rms current, core loss from ripple flux, temperature and tolerance of gap length.

12. Harmonics and High-Frequency Requirements

Nonsinusoidal voltage must be treated using volt-seconds rather than the sinusoidal 4.44 factor:

Any dc volt-second imbalance causes flux walking and possible saturation. For an ideal symmetrical square-wave voltage of magnitude V applied for half a period,

High-frequency design must consider:

• frequency- and waveform-dependent core loss;
• skin and proximity effects;
• interwinding and common-mode capacitance;
• leakage-inductance voltage overshoot;
• dielectric loss and partial discharge;
• gap fringing and local conductor heating;
• electromagnetic interference and shielding; and
• thermal interfaces in compact structures.

Planar magnetics provide repeatable low-profile windings but may have high capacitance and severe ac copper loss. Litz wire is effective only when strand diameter and bundle construction suit the operating frequency. Wide-bandgap semiconductor converters increase switching speed, making parasitic inductance, insulation stress and common-mode current increasingly important.

13. Illustrative Design Case: Dry-Type Transformer Redesign

Background

This is an illustrative industrial case. A manufacturer develops a 25 kVA dry-type isolation transformer for a power-electronics test station. The load contains harmonic current, ambient temperature can reach 40°C, and the installation has limited airflow.

Problem

The first prototype meets voltage ratio but exceeds its winding hot-spot target. It also has excessive secondary voltage drop and insufficient margin against short-circuit winding movement.

Analysis

Loss separation shows that harmonic current raises winding I²R_ac loss beyond the sinusoidal estimate. A large primary–secondary insulation gap increases leakage reactance. Mechanical analysis identifies axial compression near winding ends due to radial leakage flux, while axial leakage flux produces significant radial hoop loading in the winding bodies.

A thermal network distinguishes inner winding, outer winding, core and enclosure nodes. It shows that increasing exposed enclosure area alone gives little benefit because winding-to-enclosure conduction is the dominant bottleneck.

Approach

The revised design:

1. increases window area and reduces current density;
2. divides the secondary conductor into parallel transposed strands to reduce eddy loss;
3. adds radial cooling ducts and improves the winding-to-core thermal path;
4. uses controlled sectional interleaving to reduce leakage without violating impulse-clearance requirements;
5. reinforces end blocks and axial clamping;
6. verifies hoop stress under peak asymmetrical fault current;
7. evaluates loss using the specified harmonic spectrum rather than rated fundamental current alone; and
8. confirms dielectric clearances and partial-discharge performance by test.

Outcome

The redesign reduces winding loss and leakage reactance, keeps the predicted hot spot within the selected insulation-system limit, and increases mechanical fault margin. The transformer becomes slightly larger and more expensive, but reliability and compliance improve.

Learning Points

Electromagnetic, thermal, dielectric and mechanical changes are coupled. Interleaving can improve regulation while worsening capacitance; larger insulation spacing can improve dielectric strength while increasing leakage; higher current density reduces material but raises loss and temperature. Successful design therefore requires controlled iteration rather than isolated formula use.

14. Design Governance, Standards and Best Practice

Designers should apply the relevant national or international standards for transformer testing, insulation coordination, temperature rise, efficiency, short-circuit withstand, fire safety and electromagnetic compatibility. Requirements differ among power, distribution, traction, medical and converter transformers.

Responsible design also includes material traceability, realistic loss declarations, recyclability, fire and oil-spill risk, worker safety, and supply-chain control. Simulation and AI-assisted optimization can search large design spaces, but generated geometries must be checked against physical laws, manufacturability, tolerances and independent tests. Finite-element results are particularly sensitive to material data, boundary conditions, mesh quality and representation of end regions.

Best practice is to maintain a verified design record containing assumptions, material curves, calculations, drawings, tolerances, test results and revision history.

Chapter Summary

Transformer and inductor design begins with Faraday’s law, magnetic reluctance and material limits but must integrate many disciplines. The area-product equation provides a defensible route from transformer rating to preliminary core and window dimensions. Turns, conductor areas and connection details then establish flux density and electric loading. Leakage energy determines reactance and influences voltage regulation, parasitic fields and short-circuit forces.

Thermal design requires more than a qualitative cooling choice: convection, radiation, conduction and multi-node thermal networks determine component and hot-spot temperatures. Insulation coordination must address steady-state, impulse and environmental stresses. Under short circuit, axial leakage flux acting on circumferential current produces radial force, whereas radial leakage flux produces axial force. Inductor design additionally requires stored-energy, air-gap and saturation calculations. Harmonics and high switching frequencies increase core, copper, dielectric and electromagnetic-interference challenges, making iterative analysis and testing essential.

Key Terms

Area product: Product of effective core area and winding-window area used for preliminary magnetic-component sizing.

Window utilization factor: Fraction of geometrical window area occupied by electrical conductors.

Flux density: Magnetic flux per unit area, measured in teslas.

Reluctance: Opposition of a magnetic path to magnetic flux.

Core loss: Hysteresis- and eddy-related loss in magnetic material.

Copper loss: Resistive winding loss, including relevant ac effects.

Leakage flux: Flux that does not link all coupled windings.

Leakage reactance: Reactance associated with leakage inductance.

Voltage regulation: Change in terminal voltage between specified loading conditions.

Current density: Current divided by conductor cross-sectional area.

Skin depth: Characteristic depth of current penetration in an ac conductor.

Proximity effect: Redistribution of conductor current caused by nearby alternating magnetic fields.

Thermal resistance: Temperature difference per unit heat flow under stated conditions.

Hot spot: Highest local operating temperature in a component or insulation system.

Insulation coordination: Selection of dielectric strength relative to expected electrical stresses and protection.

Partial discharge: Localized dielectric discharge that only partially bridges insulation.

Magnetic pressure: Field-associated pressure approximated by B²/(2μ₀) under suitable field conditions.

Air gap: Low-permeability region introduced to control inductance and energy storage.

Flux walking: Progressive dc displacement of magnetic flux caused by volt-second imbalance.

Triplen harmonics: Odd multiples of the third harmonic that are co-phasal in a balanced three-phase system.

Review Questions

Short Answer Questions

1. State the assumptions behind the sinusoidal transformer emf equation.
2. What does the window utilization factor represent?
3. Why can interleaving reduce leakage inductance?
4. Distinguish core loss from load loss.
5. Why must absolute temperature be used in radiation calculations?

Long Answer Questions

1. Derive the single-phase transformer area-product equation and discuss its limitations.
2. Explain the complete iterative procedure for designing a power-frequency transformer.
3. Discuss convection, radiation, conduction and multi-node thermal-network modelling.
4. Explain insulation coordination for a high-voltage transformer.
5. Discuss the effects of harmonics on transformer core, windings, neutral current and temperature.

Analytical / Application Questions

1. Design the turns and conductor areas of a 5 kVA, 400/200 V, 50 Hz transformer for specified core area, flux density and current density, and verify window fit.
2. Construct and solve a three-node thermal network for a core, winding and enclosure using specified losses and thermal resistances.
3. A gapped reactor must provide 500 μH at 15 A peak. Select turns and gap for a given core area and maximum flux density, then identify the checks omitted by the ideal gap model.

Multiple Choice Questions with Answers

Question 1

For sinusoidal core flux, which expression gives the rms induced voltage of a winding?

A. E = 2πfNB_mA_c

B. E = 4.44fNB_mA_c

C. E = fNB_m/A_c

D. E = ½LI²

Correct Answer: B. E = 4.44fNB_mA_c

Question 2

In the single-phase area-product derivation used in the chapter, the required area product is:

A. A_p=S/(2.22fB_mJk_u)

B. A_p=2.22S fB_mJk_u

C. A_p=S J/(4.44fB_mk_u)

D. A_p=4.44fB_m/(SJk_u)

Correct Answer: A. A_p=S/(2.22fB_mJk_u)

Question 3

A transformer core has A_c=0.005m², N=100, f=50Hz and B_m=1.2T. What is the approximate sinusoidal rms winding voltage?

A. 133.2 V

B. 266.4 V

C. 444 V

D. 600 V

Correct Answer: A. 133.2 V

Question 4

Circumferential winding current interacting with axial leakage flux produces primarily which force direction?

A. Radial

B. Axial

C. Circumferential only

D. No electromagnetic force

Correct Answer: A. Radial

Question 5

Circumferential winding current interacting with radial leakage flux near a winding end produces primarily:

A. Radial force

B. Axial force

C. Torsional force only

D. Zero force

Correct Answer: B. Axial force

Question 6

For a star-connected balanced three-phase winding, phase voltage is:

A. V_L

B. √3V_L

C. V_L/√3

D. V_L/3

Correct Answer: C. V_L/√3

Question 7

Which equation correctly represents combined convection and radiation from a surface?

A. Q=h_cA(T_s-T_a)+εσA(T_s⁴-T_a⁴)

B. Q=h_cA(T_s⁴-T_a⁴)+εσA(T_s-T_a)

C. Q=(h_c+εσ)A/(T_s-T_a)

D. Q=h_cεσA(T_s-T_a)

Correct Answer: A. Q=h_cA(T_s-T_a)+εσA(T_s⁴-T_a⁴)

Question 8

An ideal gap-dominated inductor has L=200μH and carries 10 A peak. What is its peak stored energy?

A. 1 mJ

B. 5 mJ

C. 10 mJ

D. 20 mJ

Correct Answer: C. 10 mJ

Question 9

What is the main consequence of dc volt-second imbalance in a high-frequency transformer?

A. Flux walking toward saturation

B. Elimination of leakage inductance

C. Zero core loss

D. Automatic cancellation of winding capacitance

Correct Answer: A. Flux walking toward saturation

Question 10

Which design change generally reduces transformer leakage inductance but can increase interwinding capacitance?

A. Increasing winding separation

B. Interleaving winding sections

C. Reducing core permeability

D. Increasing air-gap length

Correct Answer: B. Interleaving winding sections