In-Process Inspection

Closed-loop manufacturing is not a marketing phrase—it is a practical strategy for making parts right the first time when tolerances are tight, material is expensive, and delivery schedules are tied to qualification milestones. In aerospace and defense programs, the cost of nonconformance is amplified by regulated workflows (AS9100, NADCAP special processes, ITAR/DFARS controls), long material lead times, and the documentation burden of a complete certification pack.

In process inspection (often implemented through in-process metrology and sensor-based monitoring) is a cornerstone of closed-loop manufacturing. Instead of waiting for final inspection to discover a defect, you measure and correct during the build or during machining—when the part can still be saved. The result is fewer scrapped builds, less rework, more predictable capability, and better evidence for process validation.

What in-process metrology is

In-process metrology is the measurement of a part while it is being manufactured, using sensors or measurement systems integrated into the production step. It sits between purely “open-loop” manufacturing (run the program, then inspect) and fully closed-loop control (measure, then automatically compensate in real time).

In the defense and aerospace context, in-process metrology typically shows up in three places:

1) CNC machining and post-processing: Probing a datum, measuring features, and updating tool offsets or coordinate systems before the finish pass. This includes spindle probes, laser tool setters, in-machine scanning, and integrated temperature compensation strategies.

2) Additive manufacturing (AM): Monitoring powder bed fusion (PBF) builds—DMLS/SLM—using layerwise imaging, melt pool monitoring, recoater force monitoring, and oxygen/moisture controls. The goal is to detect anomalies early and correlate signatures to defect risk.

3) Hybrid and downstream validation: Using metrology gates between operations—e.g., after stress relief, after HIP, and after rough machining—to ensure the part remains in-control before consuming more value-add time.

It is important to separate terms that are often conflated:

In-process inspection means checking during manufacturing (e.g., probing a bore after roughing). In-process monitoring means collecting process signals (e.g., melt pool photodiode data) that may indicate a risk condition. Both support closed-loop manufacturing, but not all monitoring is metrology, and not all metrology is closed-loop control.

How it reduces scrap

Scrap in advanced manufacturing is rarely caused by a single mistake; it is usually the accumulation of small variations that go unnoticed until final inspection. In-process metrology reduces scrap by shortening the feedback cycle and by making defects “visible” when they are still correctable.

Common scrap drivers and how in-process inspection addresses them:

Datum shift and fixturing error: A small misalignment in setup can translate into a full scrap at final CMM. In-machine probing can validate datum features and part orientation before you cut critical geometry. If the part is slightly shifted, the CNC can update the work coordinate system (WCS) rather than cutting the entire part off-location.

Thermal growth: Long 5-axis cycles, heavy roughing, and thin-wall features are sensitive to temperature. In-process measurement (probe or scan) can detect drift and allow controlled compensation. Even simple controls—stable coolant temperature, warm-up routines, machine tool thermal compensation enabled and validated—reduce variation dramatically.

Tool wear and breakage: Tool life variability is a real cost in nickel alloys, titanium, and cobalt-chrome. A laser tool setter or spindle probe can detect tool length changes, broken tools, and wear trends. That prevents late-stage failures like undersized bosses or scalloped surfaces that miss minimum wall requirements.

Additive anomalies carried into downstream operations: In PBF, lack-of-fusion, recoater strikes, spatter accumulation, and local overheating can create defects that may survive HIP or distort during stress relief. Layerwise monitoring flags anomalies early so you can stop the build, adjust parameters on a controlled basis, or quarantine parts for targeted NDE (e.g., CT scanning) rather than discovering defects after machining.

Excess material removal and rework loops: Without metrology gates, shops often “machine to print and hope.” With in-process checkpoints (after roughing, after semi-finishing), you can confirm you have adequate stock for finish features, avoiding the classic failure mode of machining a feature undersize and having no recovery path.

For procurement teams, the practical takeaway is that in-process inspection lowers total cost not only by reducing scrapped hardware, but by reducing schedule disruption—fewer MRBs, fewer concessions, fewer expedite fees, and fewer repeat builds because a problem was found too late.

CNC closed-loop feedback

CNC machining is where closed-loop manufacturing is most mature. Done correctly, it looks less like “checking a part” and more like running a controlled plan that ties datums, tool offsets, and feature measurement into a repeatable workflow.

A practical closed-loop CNC workflow (typical aerospace/defense):

Step 1: Verify setup and datums
Use a spindle probe to locate primary datums on the in-process workpiece (or fixture datum artifacts). Confirm part orientation, verify clamp integrity, and set/confirm the WCS. Record probe results if your quality system requires objective evidence.

Step 2: Control tools before cutting
Use a laser tool setter (or touch probe) to set tool length and diameter offsets. For critical tools, add a quick “pre-flight” measurement at the beginning of the cycle to catch gross offset errors or worn tools.

Step 3: Rough machining with stock allowance strategy
Rough with explicit stock targets and validate key stock conditions before committing to finishing. For thin-wall or distortion-prone parts (common after AM + HIP), plan for intermediate stress relief if required by material/geometry.

Step 4: In-process feature measurement
Probe bores, bosses, or planar surfaces after semi-finish. Compare to in-process control limits (which may be tighter than print tolerances to protect finishing). If out-of-family, correct tool wear offsets, adjust WCS, or route the part for rework before finishing.

Step 5: Finish machining with compensation
Apply controlled offset updates based on measured data. In high-volume work, this can be automated; in high-mix aerospace work, it is often semi-automated with documented rules and sign-offs.

Step 6: Close with verification and traceability
Capture final results in the inspection record, linking measurement data to part serial/lot, machine, program revision, and operator. This supports AS9100 traceability and strengthens the certification pack.

Where shops get tripped up: “Closed-loop” does not mean changing offsets without governance. In regulated environments, you need defined decision rules: who is authorized to apply offsets, what limits trigger rework vs. scrap, and how changes are recorded. The best implementations treat offset updates as controlled process adjustments—not informal tribal knowledge.

Integration with CMM and NDE: In-process metrology doesn’t replace final inspection. It reduces the probability of a late discovery. Many teams use a hybrid approach: in-machine probing for process control, and CMM for final acceptance of critical characteristics. For internal passages or complex AM lattices, CT scanning may be specified for acceptance or for periodic validation, especially during initial qualification.

Additive monitoring basics

Additive manufacturing—especially metal PBF (DMLS/SLM)—adds unique sources of variability: powder condition, atmosphere control, recoater behavior, thermal history, and scan strategy all influence porosity, distortion, and microstructure. In-process monitoring is the bridge between “build and pray” and a controlled, repeatable AM process.

What you can monitor during PBF (and why it matters):

Powder and atmosphere: Oxygen and moisture levels are leading indicators for oxidation risk, spatter behavior, and defect propensity. Controlled powder handling, sieving strategy, and lot traceability should tie back to material certifications and internal acceptance criteria.

Layerwise imaging: Optical images of each layer can detect streaking, incomplete spread, and obvious recoater disturbances. It’s especially useful for catching recoater blade damage or build plate issues early.

Melt pool monitoring: Photodiodes and cameras measure emission intensity and size/shape proxies. These signatures can correlate to lack-of-fusion risk, keyholing, or parameter drift. The value increases when you establish correlation to destructive testing, density measurements, and CT scan outcomes for your specific machine/material/parameter set.

Recoater force/torque: Abnormal force profiles can indicate part growth, warping, or spatter accumulation—conditions that can culminate in a recoater crash and a scrapped build. Force monitoring can trigger a controlled pause, inspection, or termination before the damage propagates.

Thermal monitoring: Thermal cameras or pyrometry can highlight hotspots and uneven cooling that drive distortion, residual stress, and cracking (notably in some nickel alloys). Even when not used for real-time control, these signals help refine support strategies and scan paths.

How this ties into AM + HIP (PM-HIP) and machining:

Many aerospace programs use AM to create near-net shapes, then apply stress relief and/or Hot Isostatic Pressing (HIP) to reduce internal porosity, followed by 5-axis machining to final tolerances. Monitoring helps you decide what should proceed through HIP (high value-add) versus what should be quarantined.

A realistic step-by-step gated workflow for metal PBF parts destined for flight or mission-critical use:

Gate 0: Powder and machine readiness
Verify powder lot, reuse ratio, and handling records; confirm machine calibration, filter condition, and atmosphere sensors. Document in the traveler for traceability.

Gate 1: Build monitoring review
Review layerwise anomalies and key process signals. Flag suspect regions or parts (when building multiple components). Define criteria for “continue,” “hold for NDE,” or “terminate.”

Gate 2: Post-build inspection
Perform visual inspection, dimensional checks for gross distortion, and density coupons (when required). Remove from plate per controlled method; maintain part identification.

Gate 3: Heat treat / stress relief
Execute controlled thermal cycle per material spec and internal procedures. Record furnace charts and calibration status (often critical in NADCAP-managed heat treat operations).

Gate 4: HIP (as required)
HIP parameters must be controlled and recorded. HIP can close internal pores but will not “fix” all defect types (e.g., severe lack-of-fusion, contamination, or geometric distortion beyond machining allowance). Use monitoring and pre-HIP checks to avoid sending bad parts into an expensive step.

Gate 5: Pre-machining metrology
After HIP, verify critical datums and stock condition. If distortion is present, adjust fixture strategy and machining plan before you chase tolerances.

Gate 6: Final machining + inspection
Use in-process probing to protect critical features, then final CMM and required NDE (e.g., FPI, CT scanning) per drawing/spec. Compile objective evidence for the certification pack.

This approach makes additive monitoring “procurement-relevant” because it converts sensor data into decision points that protect cost, lead time, and compliance.

Reporting

In regulated manufacturing, measurement without documentation is not a control—it’s a missed opportunity. Effective reporting turns in-process inspection into proof of capability and supports faster source approval, cleaner PPAP-like submissions (where applicable), and fewer back-and-forths during customer audits.

What good in-process inspection reporting includes:

Part identity and traceability: Serial number/UID, build ID (for AM), traveler/route card reference, and links to material heat/lot. For defense programs, confirm ITAR handling requirements and ensure data access controls are consistent with your compliance program.

Process context: Machine ID, program revision, tooling revision, fixture ID, operator, shift/date/time, and environmental conditions when relevant (temperature can matter for tight-tolerance machining and CMM correlation).

Measurement method: Probe/scanner type, calibration status, measurement routine revision, and stated uncertainty (especially when using in-machine probing for acceptance decisions). Many teams treat in-machine probing as process control evidence and reserve acceptance for CMM unless they have validated equivalency.

Results tied to requirements: Clearly indicate nominal, tolerance, actual, and disposition criteria (pass, rework, hold). Avoid “data dumps” that require interpretation; procurement and quality teams need clear accept/reject logic.

Nonconformance workflow linkage: If an out-of-tolerance result occurs, reporting should connect to the NCR/MRB process—what was contained, what was corrected, and what preventive action was taken.

Certification pack alignment: Many aerospace/defense customers expect a coherent pack that may include certificates of conformance (CoC), material certs, heat treat/HIP records, NDE reports, calibration certifications, first article inspection (FAI) per AS9102 when applicable, and dimensional reports. In-process data doesn’t always go to the customer, but it strengthens internal control and can be selectively shared when needed for issue resolution.

Data retention and cybersecurity considerations: High-resolution AM monitoring data can be large. Define what is retained (raw vs. summarized), for how long, and how it is protected. For DFARS-sensitive programs, ensure your handling of controlled technical information aligns with contractual requirements.

When it’s worth it

Not every part needs advanced in-process metrology. The best ROI comes when the cost of a late discovery is high, or when process variability is inherently difficult to control without frequent feedback.

In-process inspection is usually worth the investment when:

The part is high value-add: Expensive powder (Ti-6Al-4V, Inconel), long machine cycles, complex 5-axis setups, or multiple special processes (HIP, heat treat, NADCAP NDE). Early detection prevents throwing away weeks of lead time.

The tolerance stack is tight or datums are critical: If a single datum shift causes cascading nonconformances, probing and closed-loop datum management pays back quickly.

The geometry is distortion-prone: Thin walls, large flatness requirements, or AM near-net shapes that move after stress relief/HIP. Intermediate metrology gates prevent chasing distortion at final inspection.

You are in qualification or scaling: During first articles, process prove-out, or when adding a new machine/material parameter set, monitoring plus in-process metrology accelerates learning and establishes a defensible baseline for capability.

You have recurring MRB drivers: If your Pareto chart shows repeated failures (undersized bores, positional errors, wall thickness issues), in-process checks can be targeted to those characteristics rather than applied everywhere.

How to scope it pragmatically (so it doesn’t become “inspection theater”):

1) Start with critical characteristics
Identify features that historically drive scrap or customer escapes. Add in-process checks only where a correction is possible (offset update, re-cut, re-fixture).

2) Define decision rules
Write simple, auditable rules: allowable offset changes, when to stop the machine, when to route for rework, and when to escalate to quality engineering.

3) Validate correlation
If you intend to use in-machine measurements for acceptance, demonstrate correlation to CMM and account for uncertainty. Many organizations keep acceptance on the CMM and use in-process results as process control until correlation is proven.

4) Build the reporting into the traveler
Make data capture automatic where possible. If reporting is optional or manual, it will drift and lose value during schedule pressure.

5) Use the data to improve the process
The end goal is not “more data.” It is fewer defects. Feed results back into fixture design, toolpath strategy, AM parameter tuning, and supplier development.

Closed-loop manufacturing becomes real when measurement informs action. For aerospace and defense suppliers, in-process metrology and monitoring are not just quality upgrades—they are risk-reduction tools that protect schedule, certification, and mission performance.