Engineering designation of critical parts

[Ng2020]

Context - I am in a growing startup design/manufacturing environment. No Production Cert.
Note: we do fixed wing airplanes, not rotorcraft - an important distinction because of
27.602/29.602.

Currently our engineering drawings & process specs do not explicitly identify critical parts, nor do they provide direction for manufacturing to conduct specific inspections on such details.
One of our Manufacturing Eng's is from a rotorcraft background and is interested in identifying each critical part in the BOM so that affected work instruction cards get 2x QA signatures.

In your collective experience:
1. is this standard practice in your org.? If so how are critical parts typically identified on the engineering drawings?
This seems like a trivial question, but I don't recall seeing this practice in other organisations. If it was happeneing, it wasn't commonplace or immediately apparent.

2. What are the Manufacturing/ QA implications of such a designation in your organisation ? does this designation purely relate to ensuring serialisation/marking of critical parts under Part 21, or does your organisation impose additional inspection criteria?

Apologies if these questions seem a bit half baked. hopefully this will provoke some good discussion which will yield answers to questions I haven't even thought about yet.

 

[SWComposites]

Critical parts are rare, just those needing extra inspections/QA/etc over and above your normal production QA system and engineering requirements. Engine mounts for instance.

From one company I have experience with, drawings state "Critical Part per Process Spec XYZ", where the referenced process spec has the requirements for a critical part.

Implications? the extra paperwork and inspections for the part.

Tell your ME you don't have any critical parts.

 

[WKTaylor]

Every company that I'm aware-of... that NEEDS to designate a part as 'manufacturing/flight critical'... has written/numbered policy document(s) or drawings or Flag Notes, that explicitly state fabrication criteria... A-to-Z... for the raw material, drawing quality, manufacture-inspection-certification [to spec], installation, tracking [Part/Serial Number]... etc. AND these parts are identified in the AMM for mechanics and in company proprietary/unique documents for internal use.

BUT be aware, once a part is 'labeled' with this 'critical' status costs and documentation skyrocket. Usually only specific qualified raw-stock, manufacturers-people-equipment- etc have been vetted... changes to any element may require a MAJOR/INDEPTH 're-do of everything for qualification'.

On my old jet there were few such parts for the last 60-years. New generation designers, analysts, M&P, etc tend to default to greater/critical controls... especially when modern materials and processes change the nature of part fabrication... such-as... 'was' a die forging: now machined from forged block or plate, alternate materials, etc... even if margins are high. Sighhh.

Have fun.

 

[rb1957]

Are you making your own parts, or re-making other people's parts ?

 

[Ng2020]

Thanks WK, Sw.
Your description fits my understanding - I want to avoid a proliferation of critical parts designations, the overhead would just about kill our nascent project.

We are a hopeful oem-to-be, manufacuring parts of our own design.

 

[MintJulep]

1. Pick definition of "Safe", "Safety", "Critical" and "Safety Critical". Make sure that everyone on your team understands what they really mean. Use them consistently everywhere.
2. Review all the other definitions of "Critical", etc. and make sure that everyone on your team understands what they mean.
3. Define a real safety process that includes engineering design review and analysis able to correctly identify if a part, in its application is critical or not.
   1. If you work with external requirements, make sure that the review includes understanding your customer's definition of "Critical", etc.
4. If you identify a "Critical" part during the review, redesign and repeat the review to ensure that it's no longer critical.

There are so many frameworks for "safety" in hardware, software and systems engineering that use and define words like "critical", "safety critical", "vital", "safety related" etc. with different definitions and implications. People, especially project managers (and especially ones with MBAs) love to toss these very important sounding words around in meetings, but have no understanding of what the words really mean.

I've seen specifications where different chapters define "safety critical" in ways that are directly contradicting.

I've seen specifications that actually require certain things to be design as "safety critical", which is of course a direct result of the specification writers not understanding the meaning of the words they use.

 

[rb1957]

ok, "we hope to make our own parts" ... well this (designation of critical parts) is only one tiny part of your manufacturing plan, which needs to be approved.

So I'd start by talking to the guys who'll be doing this approval of your operation. What do they want to see ? What guidelines can they give you ?

Have you made other OEM's parts ? If so, you know what to expect.

I think it makes sense to allow yourself a "critical" part designation, and to define the extra work that this entails. You anticipate a lot more expense, so add the VP of Engineering to approving this designation (to keep the numbers down). A major design advantage (to the engineers) is that this helps to leverage multi-loadpath damage tolerant designs; a minimalist statically determinate design is not multiloadpath and therefore everything is "critical", a multi loadpath redundant design allows for loadpaths to fail and so is inherently less critical (though it is heavier and more difficult and expensive to build). Yes, there's a mountain of other considerations, this is just the broad brush strokes.

You can also use the word narratively, rather than in a Quality Control sense; and I'm sure that several will react strongly to that. I worked on a strutted wing plane. The strut is clearly single loadpath, flight critical, fatigue critical, but it was never AFAIK designated that way (ok, the design was over 50 years old). It had all the usual analysis and testing, maybe Operations put some extra checks into it. The company is a well known OEM with an excellent track record of good, safe designs (if unfortunately limited market success).

 

[GregLocock]

Presumably the criticality of a sub system is assessed using an FMEA.

 

[rb1957]

FMEA would be one way, SSA (System Safety Analysis) another, basic structural concepts (like design the lift strut to have two parallel loadpaths, each capable of limit load ... heavier but you may be able to take advantage of improved fatigue, and lower maintenance costs.

 

[Stress_Eng]

One company I worked at had an entry in the drawing, identifying the 'CLASS' of part. Class 1 was 'Structural', in that the part had to be certified (means of compliance, be it by analysis, test, etc). Then there was CLASS 2 and CLASS 3. Class 3 was non-structural, non-load bearing, etc. Having a classification process permits a 'procedure' to be put in place for when changes to drawings take place, e.g. a change to class 1 parts would need recertification. Also, if a part was manufactured that doesn't meet drawing requirements, the 'procedure' would require that part to go through a 'concessions' process. I'm sure such a 'procedure' could include other quality / certification aspects.