Capstone: Defend the Technique and the Result

Purpose

The capstone tests whether the learner can join the entire crossover workflow under realistic constraints. The assignment is not a contest to produce the sharpest image. It requires a defined inspection problem, controlled technique proposal, performance evidence, valid acquisition, disciplined interpretation, auditable report and technical defence.

Successful completion demonstrates supervised crossover competence within the assessment conditions. It does not independently confer certification, employer authorisation, Level 3 authority or approval to issue new production procedures. Those decisions remain with the employer’s written practice, certification system and governing documents.

Learning outcomes

After completing this chapter, the learner will be able to:

define an examination problem and its approval boundary;

design and justify an FMC/TFM/PCI technique under supervision;

verify system health and demonstrate relevant technique performance;

acquire and quality-check encoded matrix data;

interpret, corroborate and size within qualified limits;

produce an auditable report and data package;

defend choices, limitations and uncertainty during technical review; and

revise the Chapter 1 concept map to show accurate knowledge transfer.

Crossover map: the complete evidence chain

Capstone brief and constraints

The learner receives a representative component or a validated simulation/data package, an examination objective, controlled drawings, available equipment, a procedure framework and an acceptance-source extract. Some information may be deliberately incomplete to test whether the learner identifies a hold point rather than inventing a value.

The brief should include realistic constraints such as limited access, surface geometry, scan-time limit, material attenuation, data-storage allowance or a required difficult reflector orientation. Constraints are not invitations to bypass proof. They require documented trade-offs and technical approval where essential variables are affected.

Before touching the instrument, submit:

problem definition and required examination volume;

discontinuity/orientation model;

component coordinate convention;

proposed probe/wedge and matrix acquisition;

path/view set with ray diagrams;

ROI, grid and image-domain controls;

scan rate/increment and storage estimates;

essential-variable table and approval gates;

verification/performance target plan; and

data/reporting plan.

The instructor or authorised reviewer may approve the training trial, return it for revision or identify items requiring a Level 3 decision. Approval to begin the capstone is not approval for production use.

Stage 1 — Design defence

The learner gives a short technical defence before setup. The defence should answer:

Why is each path physically plausible and relevant?

Which expected reflector behaviour does each view address?

What region remains weak, inaccessible or uncertain?

Which variables can change within the plan, and which require review?

How will amplitude and phase/coherence evidence be kept separate?

Which targets demonstrate the most difficult performance claims?

The reviewer should challenge at least one path, one essential variable and one target choice. A competent response can be to withdraw an unsupported view or add a hold point. Defending every initial choice at all costs is not technical confidence; it is just stubbornness wearing a lanyard.

Stage 2 — System and technique proof

Complete the three proof gates from Chapter 12. Demonstrate channel/matrix health, timing and coordinate registration, encoder accuracy, known-target location and the application-specific task. Use target types that challenge both specular and diffracting behaviour when PCI is part of the technique.

If a check fails, stop, investigate and record the corrective loop. Assessment credit comes from disciplined diagnosis, not from a flawless first attempt. A learner who conceals a failure or modifies tolerances fails the governance requirement even if the final image looks correct.

Freeze the released acquisition and processing revisions after successful demonstration. Record target results and limitations before scanning the unknown specimen.

Stage 3 — Encoded acquisition

Conduct the field workflow from Chapter 15:

verify component identity, surface and datum;

complete pre-use checks;

use unique source-data identifiers;

monitor coupling, channel health, clipping, position and storage;

verify scan extent and encoded continuity;

record deviations and repeated segments; and

complete post-use checks before releasing the setup.

The instructor may introduce one controlled event, such as coupling interruption, storage warning, encoder sign mismatch or a weak receiver channel. The learner must recognise the evidence and take the authorised action. The correct response may be to pause and repeat a segment, not to continue for the sake of schedule.

Stage 4 — Reconstruction and interpretation

Work on a controlled copy. Apply the released path, ROI, grid and processing recipes. Any diagnostic changes must be saved separately and marked non-reportable until reviewed.

For each reportable response:

assign an indication ID and component coordinate;

identify the path and image domain;

inspect supporting A-scans and matrix/aperture behaviour;

compare complementary amplitude/coherence/path evidence;

test known geometry and artefact explanations;

characterise only to the demonstrated level;

apply the authorised sizing method;

repeat the measurement and state uncertainty; and

identify whether technical or acceptance escalation is needed.

The learner should include at least one rejected response and explain why it was judged non-relevant or artefactual. Absence of reportable indications does not remove the requirement to demonstrate complete review.

Stage 5 — Report and data package

The final package contains:

released technique proposal and review status;

calculations, path diagrams and essential-variable table;

equipment/setup/performance check records;

raw source data, setup and auxiliary channels;

transfer/integrity record;

processing recipes and versioned derived images;

indication/evidence table and sizing worksheets;

uncertainty and limitation statement;

report with acceptance authority clearly identified;

deviation, rework and amendment history; and

revised Chapter 1 concept map.

Another competent person should be able to reproduce a selected reported image from the package. If they cannot, the learner must investigate and correct the provenance gap before technical defence.

Stage 6 — Oral technical defence

The defence should test judgement rather than memorised definitions. The reviewer may ask the learner to:

trace one pixel response to selected pair A-scans;

redraw a reported path and identify every velocity;

explain why a silent view does or does not weaken a conclusion;

distinguish amplitude from coherence evidence;

predict the effect of one setup error;

justify target choice and qualified range;

explain a rejected alternative cause;

reproduce a reported image from source data;

defend significant figures and uncertainty; and

state exactly which decisions require Level 3 or other authorised approval.

“The instrument selected it automatically” is not an adequate technical defence. Automation can apply a controlled recipe; responsibility still includes knowing what the recipe assumes and where it is valid.

Assessment principles

Assessment should separately score:

technical model: coordinates, paths, modes, timing and physics;

technique design: coverage, acquisition, reconstruction and targets;

verification: proof gates, tolerances and corrective action;

field execution: coupling, position, data quality and check discipline;

interpretation: raw evidence, alternatives, sizing and uncertainty;

governance: revisions, data integrity, reporting and authority; and

defence: clear reasoning, honest limitations and response to challenge.

A critical safety/governance failure overrides the numerical score. Examples include fabricating acceptance criteria, concealing a failed check, overwriting source data, knowingly reporting a physically impossible path, or acting outside authorisation.

Practical 16 — Design, execute, analyse, report and defend

Aim

Complete the full crossover workflow on a representative examination problem and defend the result under technical questioning.

Method

Review the brief and list missing information, assumptions and approval gates.

Prepare and submit the technique proposal.

Revise the plan after technical challenge and obtain training release.

Set up the identified system and complete equipment health checks.

Verify timing, geometry, encoder, paths, ROIs and output domains.

Demonstrate the required application performance on representative targets.

Freeze released configurations and acquire encoded data.

Complete data-quality and post-use checks.

Transfer and integrity-check the dataset.

Reconstruct, interpret and size on a controlled copy.

Prepare the report and complete evidence package.

Support a peer reproduction check.

Attend the oral defence and respond to one changed-condition scenario.

Revise the Chapter 1 concept map and identify three practices that changed during the course.

Evidence to retain

all design, approval, check, acquisition and reporting records listed in the data package;

peer reproduction and audit result;

oral-defence assessment record;

corrected items with revision history; and

final concept map and personal scope statement.

Stop rules

Stop if the brief lacks information needed for a safe/valid plan; the proposed path or coverage cannot be defended; equipment/setup/application proof fails; acquisition quality is invalid; source data or configuration identity is lost; an unqualified sizing or acceptance rule is requested; a critical variable changes outside the released range; or the learner is pressured to act beyond authorisation. Record the stop, preserve evidence and seek the designated decision.

Common traps

Treating the capstone as a hunt for hidden flaw locations.

Selecting many views without a threat-based rationale.

Optimising the display after seeing the answer.

Hiding a failed check because the final image appears acceptable.

Using a coherence hot spot as automatic crack-tip proof.

Reporting more precision than the demonstration supports.

Forgetting that diagnostic settings are non-reportable until approved.

Omitting the rejected-response reasoning.

Treating a peer reproduction failure as an IT nuisance.

Claiming that course completion grants production authority.

Chapter summary

The capstone integrates problem definition, technique design, proof, acquisition, reconstruction, interpretation, reporting and defence. Competence is demonstrated through a complete evidence chain and controlled judgement—not one correct-looking image. The learner must be able to explain paths, trace responses to source data, recognise limitations, preserve revisions and state the approval boundary.

Completion supports supervised crossover development. Employer authorisation, certification, procedure approval and production acceptance remain external decisions. The most credible technician is not the one who never stops; it is the one who knows exactly when the evidence has reached its limit.