ECSS Additive Manufacturing — NDT of Built Parts (space-systems/ecss/q7080-ndt-of-am-parts)
Use when the task is the inspection step of ECSS-Q-ST-70-80C on an additively manufactured part -- choosing the non-destructive methods, in the method set of ECSS-Q-ST-70-15C, that a criticality category demands, and showing that the setup actually named in the procedure resolves a flaw smaller than the one fracture control calls critical.
Domain quick reference
- A built part carries defect populations a wrought part does not: lack-of-fusion between tracks or layers, gas porosity carried in from the feedstock, unfused powder trapped in an internal channel, and cracking driven by the thermal gradient of the build. The volumetric requirement exists because of those populations, and they also decide which method can see them.
- Tomography is the only method that reaches the wall of an enclosed channel. Its resolution is geometric: the voxel is the detector pixel divided by the magnification, and the magnification is the detector distance over the object distance. A bulky part cannot be brought close to the source, so size costs resolution before any setting is touched, and a long radiographic path starves the reconstruction in the core whatever the voxel says.
- Ultrasound is the sharpest instrument against a planar lack-of-fusion flaw, and the least tolerant of an as-built surface. A shot-blasted or unmachined skin does not couple, and an enclosed channel gives no clean back-wall path, so the method is credited only on a machined, externally swept feature.
- Radiographic contrast comes from the material the flaw removes along the beam. A planar flaw lying square to the beam is nearly invisible, and the smallest flaw still visible grows with the reciprocal cosine of the tilt between beam and flaw plane. The build orientation therefore fixes the radiographic sensitivity, because it fixes where the lack-of-fusion planes lie.
- Detectability is only half the verdict. The other half is coverage: the criticality category says whether every part is volumetrically inspected, a sampled fraction is, or the surface alone is enough.
- Where the smallest detectable flaw is larger than the critical flaw, the answer is never a looser acceptance limit. It is a better setup, a machined scan surface, a different build orientation, or a proof test that substitutes for the inspection.
Workflow
- Declare the part criticality category and read the coverage it demands. Reject an uncategorized part rather than defaulting it, because coverage and sampling both hang off this one value.
- Take the critical flaw size from the fracture-control assessment. This is the requirement the inspection is measured against, and it is an input to the inspection, never an output of it.
- Size each candidate method from its declared setup: the voxel from the cone-beam geometry, the ultrasonic limit from the wavelength at the probe frequency, the radiographic limit from the section thickness and the flaw tilt.
- Strike out the methods this part defeats -- a radiographic path beyond what the source penetrates, a surface too rough to couple, an enclosed channel with no back-wall path -- and record the reason, not just the exclusion.
- Credit only a method that is both applicable and resolving, and compare the best credited detectable flaw with the critical flaw.
- Close with a verdict: coverage demonstrated, surface inspection sufficient for the category, or coverage not demonstrated with the named shortfall and the setup change that would close it.
Pitfalls
- Reading a tomography scan as if resolution were a machine property. It is a geometry: the same machine gives a coarse voxel on a large part because the part cannot be brought near the source, and quoting the machine's best voxel on a bulky bracket overstates detectability by the whole magnification ratio.
- Crediting ultrasound on an as-built surface. The coupling is lost before the pulse enters, so the procedure passes with no sensitivity at all, and the failure is silent because a clean scan and a blind scan look alike.
- Taking a radiographic sensitivity figure as if flaw orientation did not matter. Lack-of-fusion is planar and lies with the layers, so a build orientation chosen for support reasons can render the dominant defect population invisible to the very method chosen to find it.
- Treating an enclosed channel as inspected because the part was scanned. Powder left unfused inside a channel reads as parent material on a projection, so the evidence that the channel is clear comes from tomography or from flow and endoscopic checks, not from a radiograph.
- Comparing the detectable flaw with the critical flaw by bare arithmetic. Both are built from divisions and a cosine, so a setup meant to sit exactly on the limit can land a few units in the last place above it; the comparison absorbs that representation error while the critical flaw size stays untouched.
Behavior contract (gate 3)
The voxel geometry, ultrasonic and radiographic detectability, method applicability, coverage selection and the inspection verdict are exercised by the gate 3 contract test: scripts/test_q7080_ndt_of_am_parts.py against scripts/q7080_ndt_of_am_parts_logic.py (stdlib unittest, offline). Run: python3 scripts/test_q7080_ndt_of_am_parts.py
Compliance
- ECSS standards are freely downloadable (ESA); cite the source and paraphrase per standards-map.yaml.
- compliance: STANDARDS-REF, gated: false.
