Sooner or later you will commission a failure investigation — from us or from someone else — and receive a report that you must judge without being a metallurgist. Here is the checklist. It is the standard we hold our own reports to, so use it on ours too.
1. A summary a decision-maker can act on
One or two pages, in front: what failed, how (the mechanism), why (the cause), how certain the conclusion is, and what to do. If you cannot extract the answer and the action from the summary alone, the report has failed its first duty.
2. A visible evidence chain
What was received, when, in what condition, with photographs; what was done to each item; where material was sectioned and why. You should be able to trace any conclusion back through the examinations to the physical evidence. Reports that jump from "the shaft was received" to "the cause was fatigue" with nothing traceable in between deserve your suspicion.
3. Methods that fit the questions
Each test should have a reason: what hypothesis was it checking? A report listing many tests with no logic connecting them is often a lab package, not an investigation. Where testing was done through partner laboratories, that should be identified honestly — what matters is who designed the programme and who stands behind the interpretation. More →
4. Mechanism with its evidence
The failure mechanism stated plainly, with the specific observations that support it — fractographic features, microstructure, deposit chemistry — and, importantly, the observations that ruled competing mechanisms out. Mechanism vs cause vs root cause →
5. Honest certainty language
Look for the distinction between confirmed and most likely, and for stated bounds where evidence was incomplete. A report that is certain about everything — including things the destroyed evidence cannot support — is telling you about its author, not the failure.
6. Root cause connected to your operation
Not just "fatigue from cyclic loading" but which loading, from what source, present since when, and why it went unnoticed. The cause should be recognisable to the people who run the equipment.
7. Recommendations you can implement
Specific, prioritised actions — material, design detail, operating limit, procedure, inspection — each linked to the cause it addresses. "Improve maintenance practices" is not a recommendation; it is a shrug formatted as one.
8. Appendices that let an expert check the work
Full test results, micrographs with scale markers and locations, and the data behind every figure — so any competent reviewer could audit the reasoning. Reports withstand second opinions when they are built to invite them.
When to contact MTIS
To commission an investigation built to this standard — or for a second opinion on a report you have already received. Reviewing existing reports is normal work for us, and if the original is sound, we will say so. Failure Investigation → · Start a job request →