Skip to content

What's new in 4.3

Version 4.3 of the NDE Open File Format introduces native support for Eddy Current Testing (ECT) and Eddy Current Array (ECA) modalities. This is an additive release, all files conforming to versions 4.0, 4.1, or 4.2 remain fully valid.

Motivation

The .nde format was originally designed around ultrasonic testing. Version 4.3 extends the format to cover the eddy current modality, which operates on fundamentally different physical principles: instead of time-of-flight waveforms, EC instruments record complex impedance signals (real and imaginary components) from inductive probes interacting with conductive materials.

Supporting ET/ECA required additions in four areas:

  1. A new probe model reflecting the sensor-group abstraction of ECA probes
  2. A new acquisition process capturing EC-specific hardware parameters
  3. New data classes and axes for impedance datasets
  4. New data mapping and post-processing processes matching EC scanning workflows

New probe type: eddyCurrentProbe

The eddyCurrentProbe organizes sensors into sensor groups. Each group corresponds to an independently driven set of coils sharing a common frequency and drive configuration (analogous to a beam group in PAUT). Individual sensors within a group are positioned in a 2D coordinate system co-planar with the specimen surface using primaryOffset and secondaryOffset.

probe
 └── sensorGroups[]
      └── sensors[]
           ├── id
           ├── primaryOffset
           └── secondaryOffset

New wedge type: EC wedge

A new wedge variant for eddy current probes is added to the wedges array. EC wedges use the same positioning object as UT wedges to describe their position and orientation relative to the specimen surface. The wedgeAssociation object now supports four orientation values: Normal, Reverse, FlipPrimary, and FlipSecondary.

New processes

eddyCurrent: hardware acquisition

The eddyCurrent process captures the hardware configuration of an EC acquisition: test frequency, drive amplitude, hardware gain, and the mapping of sensors to acquisition channels (including multiplexing timeslots).

impedanceTransformation: phase and gain normalization

The impedanceTransformation process applies phase rotation and gain to raw impedance data. It operates at two levels:

  • Per-channel: individual gain and rotation for inter-channel normalization (e.g. balancing ECA probe channels against a reference notch)
  • Group-level: shared real gain, imaginary gain, and rotation applied after per-channel normalization

eddyCurrentMix: mixed-frequency processing

The eddyCurrentMix process combines two impedance data streams into a mixed output, typically to suppress an unwanted signal (e.g. geometry or support plate response) while retaining defect indications. The two streams (\(Z_1\) and \(Z_2\)) are referenced through the process inputs (z1InputId, z2InputId), and the four mixing coefficients (k1–k4) are specified independently per acquisition channel through a channels array.

conductivity: material conductivity measurement

The conductivity process converts complex impedance measurements into material electrical conductivity and probe lift-off using a set of calibration reference points. Each reference point pairs a known conductivity and lift-off value with the corresponding recorded impedance.

Filter processes

Four filter processes are added for spatial or temporal filtering of ET signals, all configurable with cutoff frequency in Hz or PerMeter:

Process Description
highPassFilter High-pass IIR or FIR filter
lowPassFilter Low-pass IIR or FIR filter
figure6Filter Figure-6 band-pass filter
figure8Filter Figure-8 band-pass filter

layerDetection

The layerDetection process identifies material layer boundaries from impedance data.

mapToDescrete: spatial remapping

The mapToDescrete process converts data acquired with an allCycle data mapping into a discreteGrid data mapping, using recorded encoder positions to assign each acquisition cycle to its corresponding spatial position on the grid.

New data mapping strategy: allCycle

The existing discreteGrid mapping assumes a regular grid of known positions. This does not fit EC acquisitions using a rotary encoder or acquisitions where the spatial grid must be reconstructed in post-processing.

The new allCycle mapping stores every acquisition cycle individually, together with its encoder position. A companion encoder-position dataset is created for each motion device. Spatial reconstruction is then performed by the mapToDescrete process.

allCycle data mapping
 └── datasets[]     ← one encoder-position dataset per motion device
      ├── dataClass: Encoder
      └── dimensions: [AcquisitionCycle]

New Polar dimension axis

A new Polar axis is available in discreteGrid dimensions for rotational scanning configurations (e.g. ECT tube inspection with a rotating probe). It behaves like UCoordinate or VCoordinate but uses angular resolution in degrees.

New dataset data classes and axes

ET data classes

Two new dataClass values are defined for ET datasets (see supported datasets):

dataClass Description
Impedance Complex impedance samples (strip chart / line scan)
ImpedanceStatus Validity/saturation flags per sample

New axes

Axis Context Description
Channel ET One entry per EC acquisition channel
AcquisitionCycle ET Sequential acquisition cycle index, used in allCycle datasets

New rotary encoder modes

Two new mode values are added to the encoder object to support rotary encoders:

Mode Description
Tach Each step corresponds to a partial rotation; stepResolution gives steps per revolution
Index Each step corresponds to one full revolution; stepResolution is typically 1.0

New specimen type: unspecifiedGeometry

A new unspecifiedGeometry specimen type is added for acquisitions where the specimen geometry is unknown or not relevant to the recording. Only the inspectable surfaces are declared (by id); no dimensions or material are required. This is useful for any context where positioning data is not tied to a fully-described part in a scan plan.