Failure Testing, Damage Mapping & Post-Test Inspection

When and Why to Test to Failure

Testing to failure is the deliberate application of load beyond the design capacity until the structure fails. It is performed when the engineering question requires knowledge of the actual failure load, the failure mode and the failure location — questions that a proof test at a sub-failure load cannot answer. The decision to test to failure depends on the programme objectives, the consequence of failure, the availability of test articles and the value of the failure evidence relative to the cost of the article. Where the structure is critical, where the analysis prediction has significant uncertainty, or where the failure mode cannot be confidently predicted from analysis alone, a test to failure provides evidence that no sub-failure test can. Testing to failure destroys the article, which means it can only be done once on that article, and the evidence must be captured in real time. There is no opportunity to go back and re-measure. The instrumentation, the imaging and the post-test inspection must be planned before the test, because once the article has failed, the transient evidence — the crack propagation sequence, the acoustic events, the load drop profile — is gone. The engineer must decide what evidence to capture, plan the capture method and verify the capture system before the failure occurs.

FINAL FRACTURE LOCATION IS NOT AUTOMATICALLY THE LOCATION WHERE FAILURE INITIATED.

The Failure Event Sequence

A structural failure is not a single event; it is a sequence. In many structures, failure begins with crack initiation at a stress concentration, a defect or a critical location predicted by analysis. The crack propagates under increasing load (or under continued cycling in fatigue), redistributing load to adjacent material. At some point, the remaining cross-section can no longer sustain the load, and final fracture occurs — often rapidly and catastrophically. The final fracture location may be the same as the initiation location, or it may be elsewhere — load redistribution during crack propagation can shift the critical location, and the final fracture may occur at a secondary site that became critical after the primary crack reduced the load-carrying capacity. This distinction is crucial for failure mode identification. If the engineer examines only the final fracture surface, they see the result of the failure, not its origin. The initiation site — which carries the evidence of why the failure started — may be on a different part of the fracture surface, or may be obscured by the final fracture. The engineer must locate the initiation site, examine it and preserve it, not simply photograph the most obvious fracture surface. This is a recurring source of incorrect failure mode identification.

High-Speed Imaging

High-speed imaging captures the failure event at frame rates sufficient to resolve the progression of damage — crack propagation, buckling, fastener failure, delamination growth. The required frame rate depends on the failure speed: a ductile metal fracture may propagate at metres per second, a brittle fracture or a composite delamination may propagate at hundreds of metres per second, and a shock-induced failure may occur in microseconds. The engineer must estimate the failure speed and select a frame rate that captures the event with adequate temporal resolution. High-speed imaging requires planning. The camera must be aimed at the expected failure location (informed by the analysis prediction), with sufficient field of view to capture the region of interest and sufficient resolution to see crack propagation. Lighting must be adequate for the frame rate — high frame rates require intense illumination. The camera must be triggered at the right moment: too early and the buffer fills before the failure; too late and the failure is missed. Pre-trigger buffering — the camera continuously records and retains the most recent frames when triggered — captures the initiation even if the trigger occurs slightly after. Multiple cameras at different angles provide a three-dimensional view of the failure sequence.

Acoustic Events and Acoustic Emission

Structural failure generates acoustic events — transient stress waves produced by crack initiation, crack propagation, delamination, fibre breakage or fastener failure. Acoustic emission (AE) monitoring uses piezoelectric sensors bonded to the structure to detect these events in real time. The timing, location and character of the events provide information about the failure sequence: the first event indicates when and where damage began, the rate of events indicates propagation activity, and the event character (amplitude, frequency content, duration) can distinguish between different damage mechanisms. AE is particularly valuable for composite structures, where internal damage (delamination, fibre breakage, matrix cracking) is not visible on the surface until late in the failure process. It is also useful for detecting the onset of crack propagation in metallic structures before it is visually apparent. The limitation of AE is that it detects the events but does not directly image the damage — the engineer must correlate the AE data with other measurements (strain, load, imaging) to interpret the failure sequence. AE sensor location and coupling are critical: a poorly coupled sensor may miss events, and the sensor layout determines the location accuracy.

Load Drop, Strain Deviation and Displacement Jump

During a test to failure, several measured quantities show characteristic signatures at the moment of failure. A sudden load drop — the load cell reading decreases even as the actuator continues to displace — indicates that the structure has lost load-carrying capacity. A strain deviation — a gauge reading that suddenly changes, drops to near zero, or reverses sign — indicates that the local strain field has changed, typically because a crack has passed through or near the gauge. A displacement jump — a sudden increase in displacement at constant or decreasing load — indicates a loss of stiffness, as from a crack opening, a fastener failing or a panel buckling. These signatures, correlated in time, tell the engineer when failure occurred, approximately where it occurred (from which gauges responded) and what type of failure it was (sudden vs. progressive, local vs. global). The time correlation between channels is essential — it requires synchronised acquisition (see Article 12). The engineer should review the high-rate data from the failure event to reconstruct the sequence: which channel changed first, what the load was at that moment, and how the failure progressed across the structure. This reconstruction is the foundation of the failure mode identification.

Damage Mapping

After the test, the engineer should systematically map all damage on the article. This includes visible cracks (with location, length, orientation and width), permanent deformation (measured and recorded with location and magnitude), buckled regions, fastener failures (sheared, pulled, loosened), disbonds or delaminations (located by tap test, ultrasonic or other NDT), and any other damage mode observed. The damage map should be referenced to a coordinate system on the article so that each damage feature can be located precisely. The damage map serves multiple purposes. It documents the final state of the article for the test record. It allows correlation between the damage and the measured response (the strain gauges near a crack should show the deviation that corresponds to the crack). It supports the failure mode identification by showing which damage features are primary (the initiation and propagation) and which are secondary (consequential damage from the final fracture or from the post-failure load path). It provides the basis for comparison with the analysis prediction — did the structure fail where the analysis predicted, in the mode predicted, at the load predicted?

Crack Initiation and Propagation Analysis

Identifying the crack initiation site is a primary objective of failure testing. The initiation site carries the evidence of why the failure started: a stress concentration, a manufacturing defect, a material inclusion, a surface flaw or a fretting damage site. On a fatigue fracture surface, the initiation site is typically characterised by a smooth, featureless region (the crack nucleation and slow-growth region), transitioning into a region with beach marks or striations (the stable propagation region), and finally a rough, fibrous region (the fast final fracture). The initiation site is at the origin of the beach marks or at the edge of the smooth region. For static fracture, the initiation site may be less visually obvious, but the fracture surface features — chevron marks pointing toward the origin, shear lips at the edges, a flat region in the centre — provide clues. The engineer should examine the fracture surface under magnification, identify the initiation region and document its features. If the initiation site is at a feature (a hole, a fillet, a weld, a defect), that feature is a candidate for the root cause. The analysis prediction of the critical location should be compared with the actual initiation site — agreement confirms the analysis; disagreement requires investigation.

Fracture Surface Preservation

The fracture surface is a record of the failure history. It contains the initiation site, the propagation features, the final fracture morphology and any evidence of the failure mechanism (fatigue striations, cleavage facets, ductile dimples, delamination planes). Once the surface is damaged, contaminated or oxidised, that evidence is lost. The engineer must preserve the fracture surface immediately after the test: photograph it before any handling, cut the fracture region from the article if necessary (without damaging the fracture surface), store it in a desiccant or inert environment to prevent oxidation, and avoid touching, cleaning or coating the surface unless following a documented preservation procedure. If the surface must be cleaned (to remove oil, debris or test fluid), the cleaning should be minimal and should follow a procedure that does not damage the fracture features. Aggressive cleaning, ultrasonic cleaning or wire brushing can destroy the features that identify the failure mode. The preserved fracture surface should be made available for fractographic examination — optical microscopy, scanning electron microscopy (SEM) — as required by the failure investigation. The fractographic findings are part of the failure evidence and should be documented in the test report.

INTERPRETING THE FRACTURE SURFACE WITHOUT PRESERVING IT OR WITHOUT EXAMINING THE INITIATION REGION CAN LEAD TO AN INCORRECT FAILURE-MODE IDENTIFICATION. The fracture surface contains the failure history — it must be preserved and examined.

Post-Test NDT

After the visible damage has been mapped, non-destructive testing can reveal internal damage that is not visible on the surface. Ultrasonic inspection detects internal cracks, delaminations and disbonds. Eddy current testing detects surface and near-surface cracks in conductive materials. Radiography reveals internal features and foreign objects. Dye penetrant and magnetic particle inspection detect surface-breaking cracks. The choice of NDT method depends on the material, the expected damage type and the access. Post-test NDT serves two purposes. First, it completes the damage map — internal damage that was not visible but may be relevant to the failure mode or to the load path is identified and located. Second, for articles that did not fail but were tested to a high load, it reveals whether any non-visible damage occurred that would affect the article's continued use or its interpretation as evidence. The NDT results should be documented with the damage map, and any features relevant to the failure mode or the structural behaviour should be highlighted.

Failure Test Evidence Types and What They Reveal

The table below summarises the principal evidence types captured during and after a failure test, what each captures, when it occurs and how to interpret it. The combination of evidence types — not any single one — produces a defensible failure mode identification.

The Failure Event Sequence Diagram

The diagram below illustrates the sequence of events in a typical test-to-failure, from load application through crack initiation, propagation, final fracture and post-test inspection.

FAILURE TEST EVENT SEQUENCE

  Load
  │                        ╭── final fracture (sudden load drop)
  │                      ╱
  │                    ╱  crack propagation (load may still rise)
  │                  ╱
  │                ╱  ← crack initiation (first AE event, strain deviation)
  │              ╱
  │            ╱    linear elastic region
  │          ╱
  │        ╱
  │      ╱
  │    ╱
  │  ╱  preload
  └─────────────────────────────────────────► time

  Evidence captured:
  │                │         │              │          │
  │                │         │              │          └── post-test: NDT, damage map,
  │                │         │              │              fracture surface preservation
  │                │         │              └── high-speed imaging, load drop,
  │                │         │                  strain deviation, displacement jump
  │                │         └── AE events, strain gauges near crack,
  │                │             imaging (if aimed at location)
  │                └── first AE event, first strain deviation,
  │                    possible imaging capture
  └── load, strain, displacement recorded continuously

  Critical: the INITIATION site and time are not the same as
  the FINAL FRACTURE location and time. Locate and preserve both.

Reconstructing the Failure and Reporting

After the test, the engineer assembles the evidence — load history, strain and displacement records, high-speed images, AE data, damage map, NDT results and fracture surface examination — into a coherent reconstruction of the failure. The reconstruction answers: what was the failure load? Where did failure initiate? What was the failure mode? How did it propagate? Did it agree with the analysis prediction? What was the actual margin? What does the failure tell us about the design, the manufacturing or the analysis? The report should present the evidence, the reconstruction and the conclusions, with the evidence traceable to the raw data and the conclusions supported by the evidence. Any uncertainties in the reconstruction — ambiguity in the initiation site, disagreement between evidence types, inability to capture an event — should be stated. A failure test report that presents a confident conclusion without acknowledging the evidence limitations is weaker than one that presents a reasoned conclusion with its uncertainties. The purpose is not to appear certain; it is to provide defensible evidence for an engineering decision.