Skip to main content
en

Cooling an electric motor: put the losses in the right place

A motor's heat does not come from one place. Winding, iron and magnet losses arise in different places and proportions at different operating points.

The thermal analysis of an electric motor does not begin with knowing the total loss — it begins with knowing where the loss arises.

The loss terms behave differently

Copper (winding) loss. Rises with the square of the current. Dominant at high torque and low speed. It arises inside the winding and must pass through the insulation and slot fill to get out.

Iron (core) loss. Rises with frequency. Becomes dominant at high speed. It arises distributed through the stator laminations.

Magnet and rotor losses. Arise on the rotor and are the hardest heat to remove: the rotor is turning, and its contact with the outside is through the air gap and the bearings. In permanent-magnet machines the temperature is a critical criterion, because magnets lose their magnetisation irreversibly when hot.

A correct analysis takes the loss not as a fixed total but as a distribution dependent on the operating point.

Bottlenecks in the heat path

  • Slot-to-tooth insulation. The largest thermal resistance on the path from the winding to the lamination stack is usually here; the conductivity of the fill material and its void fraction set the winding temperature directly.
  • Stack-to-housing contact. A contact resistance that depends on the interference fit and is sensitive to manufacturing tolerance.
  • The air gap. Thin and rotating; it is the path for heat from rotor to stator and depends on speed.
  • End windings. Outside the stack, with different cooling — and often the hottest point.

Modelling decisions

The problem is conjugate: solid regions dominate, the flow is secondary but governing.

Resolving the winding wire by wire is not practical; it is represented as a homogeneous region with equivalent anisotropic conductivity — high along the axis (copper), low across it (insulation).

The quantity sought is usually transient: how far the winding temperature rises under a short overload. That requires the thermal mass to be defined correctly.

References

  1. JEDEC Solid State Technology Association. JESD51 — Methodology for the Thermal Measurement of Component Packages (Single Semiconductor Device).
  2. Shah, R. K. & Sekulić, D. P. (2003). Fundamentals of Heat Exchanger Design. John Wiley & Sons. DOI: 10.1002/9780470172605

Let's discuss your project

Let us pin down together the system to be analysed and the question that needs answering.

Let's discuss your project