How to Plan a Failure Investigation for Critical Assets

A failed component rarely arrives with its history intact. It may have been cut from service, cleaned, moved, or exposed to weather before the investigation begins. Knowing how to plan failure investigation work before handling the evidence is therefore critical. A disciplined plan protects the failure scene, focuses laboratory effort on the right questions, and produces findings that can withstand technical, commercial, and regulatory scrutiny.

For asset owners, manufacturers, contractors, and insurers, the objective is not simply to identify a visible fracture or corroded area. The objective is to establish a defensible failure mechanism, determine contributing conditions, assess the extent of risk, and define practical corrective actions.

Start With the Decision the Investigation Must Support

A failure investigation should begin with the decision that needs to be made. That decision may concern return to service, repair versus replacement, warranty responsibility, process changes, inspection intervals, or the integrity of similar assets across a site. Without this context, an investigation can become technically interesting but operationally unfocused.

Define the primary investigation question in clear terms. For example: Did a pressure vessel crack because of fatigue, material nonconformance, welding defects, corrosion, or operation outside its design envelope? Was a concrete spall caused by reinforcement corrosion, inadequate cover, chloride ingress, impact, or an installation defect? A focused question helps distinguish essential evidence from information that is merely available.

It is also useful to identify what the investigation will not address. A laboratory examination of a fractured shaft may establish the fracture origin and mechanism, but it may not by itself confirm the full operating load spectrum. That may require maintenance records, operating data, vibration monitoring, or a mechanical design review. Establishing these boundaries early prevents unsupported conclusions.

Preserve Evidence Before It Is Altered

Evidence preservation is often the most time-sensitive part of a failure investigation plan. Cleaning, grinding, cutting, repainting, or even repeated handling can remove fracture features, corrosion products, coating layers, deposits, or marks that are central to determining cause.

The plan should nominate a person responsible for scene control and evidence release. Photograph the failed asset in situ where safe to do so, using overall, medium-range, and close-up images. Record orientation, adjacent components, environmental exposure, operating position, identification markings, and any signs of leakage, deformation, overheating, impact, or prior repair.

Each item should receive a unique identification number and a documented chain of custody. Record who collected it, when it was collected, its original location, and every subsequent transfer. This level of control is particularly important where findings may support contractual, insurance, or legal decisions.

Avoid destructive preparation until a suitably qualified investigator has reviewed the evidence. If sectioning is necessary for transport or examination, identify proposed cut locations that preserve suspected crack origins, welds, corrosion sites, and transition zones. A poorly placed saw cut can eliminate the most valuable part of the specimen.

Build a Working Failure Hypothesis

A good plan does not assume the cause. It develops several plausible hypotheses and then selects methods that can confirm or eliminate them. This is more efficient than ordering broad testing without a clear rationale.

For a fractured metallic component, preliminary hypotheses may include overload, fatigue, hydrogen-assisted cracking, stress corrosion cracking, brittle fracture, manufacturing defects, or improper heat treatment. For a coating failure, possibilities may include poor surface preparation, incorrect film thickness, incompatible coating systems, underfilm corrosion, moisture exposure, or mechanical damage.

Each hypothesis should be linked to observable evidence. Fatigue, for example, may be supported by a crack origin at a stress concentrator and progressive fracture markings. Corrosion-related failure may require evidence of localized attack, deposits, environmental contaminants, or loss of section. Material substitution may require positive material identification, chemical analysis, and comparison against the specified grade.

This hypothesis-led approach preserves impartiality. It also makes the final report easier to follow because every test result can be connected to a defined engineering question.

Define Scope, Deliverables, and Technical Authority

A written scope should identify the assets involved, available specimens, records to be reviewed, site activities, laboratory examinations, expected deliverables, and required turnaround time. The scope should distinguish between initial screening and the full investigation. In many cases, a rapid first-stage assessment is appropriate to support an urgent operational decision, followed by detailed metallurgical or chemical analysis.

Appoint a technical lead with authority to adjust the investigation as evidence emerges. Failure mechanisms do not always follow the initial theory. A component expected to have failed in fatigue may reveal an incorrect alloy composition, or a suspected weld defect may prove to be cracking initiated by an external restraint condition. The plan should allow for controlled changes in scope, with reasons documented.

Agree on the reporting standard before work begins. A defensible report should state the evidence reviewed, methods used, observations, limitations, engineering interpretation, root cause where supported, contributing factors, and recommended actions. It should clearly separate verified facts from assumptions and professional opinions.

Select Examination Methods That Answer the Questions

The testing program should be proportionate to the asset criticality, failure consequences, and available evidence. Not every failure requires every technique. The strongest investigations use complementary methods in a logical sequence, beginning with non-destructive observations and progressing to targeted destructive examination where justified.

A typical metal failure investigation may include visual examination, dimensional checks, photographic documentation, non-destructive testing, and fracture surface assessment. Optical microscopy can reveal microstructural features, weld condition, heat-affected zones, and corrosion morphology. Scanning electron microscopy with energy-dispersive spectroscopy can provide high-magnification fracture characterization and localized elemental information. Chemical analysis and positive material identification may confirm whether the material matches the specified alloy.

Where phase identification or deposits are relevant, X-ray diffraction can assist in identifying corrosion products, scale, or crystalline contaminants. Fourier-transform infrared spectroscopy may be appropriate for polymers, elastomers, oils, residues, and organic coatings. Mechanical testing, hardness testing, metallography, and coating thickness measurement may be required where degradation, manufacturing quality, or compliance with a specification is in question.

The sequence matters. Fracture surfaces should generally be examined before cleaning or sectioning. Deposits may need to be sampled before washing. Non-destructive examination of adjacent areas may reveal whether the observed failure is isolated or part of a wider damage pattern.

Gather the Operating History, Not Just the Failed Part

Laboratory results are only one part of root cause determination. A sound plan gathers the service history that explains why a susceptible component failed when it did. Request drawings, material certificates, weld procedure records, inspection reports, repair history, maintenance records, operating pressures and temperatures, load data, process chemistry, vibration trends, and photographs taken before the event.

Interview operators, maintainers, and inspectors early. Their observations can identify changes that are absent from formal records, such as unusual noise, intermittent leakage, altered duty cycles, cleaning chemical changes, delayed maintenance, or a recent impact event. These accounts should be documented and tested against physical evidence rather than accepted uncritically.

For repeated failures, examine the system rather than treating each broken item as an isolated event. Alignment, restraint, drainage, electrical isolation, cyclic loading, water chemistry, and fabrication practices can all create recurring conditions. Correcting only the visible break may return the asset to service without removing the underlying driver.

Plan for Safety, Access, and Representative Sampling

Investigation activities can introduce their own hazards. Failed equipment may retain pressure, stored energy, hazardous substances, sharp edges, unstable supports, or contaminated residues. The plan should include site access requirements, isolation and permit arrangements, personal protective equipment, lifting needs, and controls for transport and storage.

Sampling must also be representative. If a corroded pipeline section is removed, retain both the most severely affected area and sufficient unaffected material for comparison. If a weld is under review, preserve the weld metal, heat-affected zone, parent material, and any relevant repair region. When bulk material is unavailable, document the limitation rather than overstating certainty.

Set Decision Gates and Communicate Early Findings

A failure investigation is most valuable when it supports decisions at the right time. Establish decision gates such as initial evidence review, preliminary findings, test completion, and final engineering assessment. At each stage, identify what can be stated with confidence, what remains uncertain, and whether immediate controls are required for similar equipment.

Preliminary findings should be carefully framed. It may be appropriate to recommend inspection of comparable assets, temporary load restrictions, corrosion monitoring, or removal from service before the final report is issued. However, avoid presenting an early indication as a confirmed root cause without sufficient evidence.

AECTL plans failure investigations around evidence integrity, accredited technical methods, and clear engineering decisions. The most effective plan is not the one with the longest test list. It is the one that preserves the right evidence, asks the right questions, and gives decision-makers a reliable basis for preventing recurrence.