Docs / 3-D field validation

3-D field validation (cufem3d)

The 3-D lane. Geometry → mesh → a 3-D magnetostatic edge-element solve → the magnetic field and flux, with a relative-error field verdict and a VTU field artifact. 3-D geometry and field validation are available today; a validated full 3-D motor torque figure is on the roadmap.

Where it fits. cufemlab has two lanes. Lane A (2-D) is the cross-checked, validated no-load cogging/torque workflow. Lane B (3-D, this page) is cufem3d: it evaluates the 3-D magnetostatic field on your geometry and reports how close that field is to a reference (a relative-error verdict). It does not compute a validated 3-D torque.

What it does

Imports or builds a 3-D geometry (STEP / STP / IGES / STL) and meshes it.
Assembles and solves the 3-D magnetostatic problem with edge (Nédélec) elements — the correct family for 3-D magnetic fields.
Returns the magnetic field / flux density as a VTU artifact you can open in ParaView.
Reports a field-validation verdict: the relative error of the computed field against a reference, against a declared tolerance — PASS only inside it.
Runs on CPU (edge-element), is free-trial eligible, costs 1 credit, finishes in a few minutes.

What it does not claim

3-D torque is on the roadmap. The same pre-declared ±20% bar that validates the 2-D moving-band cogging path was applied to a first 3-D (2.5-D) torque attempt — and it did not pass. That result is recorded openly as a FAIL, not hidden. That is why the validated 3-D deliverable today is geometry-and-field. Full production 3-D torque, end-effects and skew are roadmap extensions on the same pipeline.

Run it in the app

Open New analysis, choose 3-D field validation, optionally upload a 3-D geometry (or use the built-in reference case), and submit. Track it in the Queue; the finished result carries the field verdict and the VTU field artifact.

Run it from Python

3-D field validation — client SDKPython

import os
from cufemlab_client import Client

client  = Client(api_key=os.environ["CUFEMLAB_API_KEY"])
project = client.create_project("3-D field validation")

# optional: upload a 3-D geometry; omit to use the built-in reference case
job = client.analyze(
    project_id=project.id,
    analysis_type="cufem3d_field_validation",
    input_params={},
)
job.wait(poll_interval=10)
r = job.result()
print(r.verdict)   # PASS if the field relative error is inside tolerance
print(r.value)     # a field-validation metric, NOT a torque

Reading the result

verdict — PASS / FAIL against the declared field tolerance.
value / metrics — the relative field error and mesh statistics.
artifact — a VTU field/flux file for ParaView.

For the engine-level notebook version of this 3-D pipeline, see cufem in Jupyter. For the full example set, see Examples & recipes.

Two lanes

Lane A · 2-D · validated

2-D cogging / torque

Cross-checked vs FEMM/GetDP with a per-run validation card — the audit-grade path. See the motor workflow.

Lane B · cufem3d · 3-D field

3-D field validation

3-D geometry, mesh and magnetic field/flux with a relative-error verdict. 3-D torque is on the roadmap.