Run your first field study

01What you will make

This recipe produces a no-load field result at one rotor position from an unchanged built-in reference design. It is a way to learn the study workflow, not a reproduction of the published benchmark campaign.

The controls below follow the development preview. Check release status for packaged-build availability and activation if local solver access is unavailable.

02Open the reference design

1. Open Design, then find Explore an example.
2. Use Find an example to locate Parviainen · 5 kW AFIR and open it.
3. Check that the design has 36 slots and 12 poles, with two stators and a central rotor. Leave its dimensions, winding and materials unchanged.
4. Save a working copy before starting the study. Keep the original example available for comparison.

The reference's nominal speed is 300 rpm. Its solver magnet-temperature setting is 100 °C; this is an input to the electromagnetic model, not a predicted operating temperature.

03Set up one no-load position

Open Simulate → Field solve and use:

• Analysis method: FEM · standard.
• Native solver: the native 3-D backend, if available in your build and licence.
• Study: Single solve (no-load).
• Speed: keep the example's automatic rated speed, 300 rpm. If overridden, set it back to 300 rpm and record that choice.
• Run label: use a recognisable name such as “Parviainen · first no-load position”.

Leave the remaining solver defaults unchanged for this first exercise. The no-load study applies zero winding current and uses a single mechanical rotor position at 0°. The permanent magnets still create a magnetic field.

04Run and inspect

Check the Field solve review, then choose Run field study. Follow the job status until it completes; a queued job has not yet produced a solved result.

Open Completed study results. Confirm the study type, backend and inputs match your intended run. Inspect the available field view and flux linkage, and read any mesh, geometry or solver warnings before using the values. A missing field artifact or a failed job is a reason to investigate, not a zero-valued result.

Keep the saved project and stored run together. Record the application version, solver backend and settings when sharing an observation. See local files and storage for where those records live.

05Know what this result cannot establish

A single position samples a field; a period sweep samples how quantities change with angle. One position does not establish a back-EMF waveform, mean torque over a period, cogging amplitude or loaded torque ripple.

To study no-load angular variation, choose Cogging sweep (no-load) as a separate study and review its angular sampling. Loaded studies also require a checked current convention and rotor d-axis calibration. See current and drive conventions.

The published Parviainen back-EMF comparison uses a period calculation and specified processing and operating conditions. Its numerical values are not acceptance targets for this single-position exercise. Read the reference evidence and results guide before making a comparison.

06Before refining the model

First make sure you can explain the geometry, materials, current, speed, temperature assumption and study type. Then vary resolution or angular sampling deliberately and retain the earlier run for comparison.

Completion proves that a calculation finished. Mesh convergence, angular convergence and agreement with a measurement are separate checks. The fidelity guide explains the next choices; the glossary defines the quantities.