Q_Goest said:
Hi Mech_Eng,
There's a lesson I learned that I've found is difficult to get across to other engineers, especially younger ones (not to imply that you're a young engineer, I know you're not). That lesson is that accepting a product for industrial service, whether it's a valve, a pump, or an entire system, on the basis that it is somehow 'certified' by the manufacturer, is dangerous.
You bring up a good point, how do you know the certification that the part is within spec is valid? This depends on the design criticality of the part. For a screw holding a cable tie to a wall, it probably doesn't matter; just call out a McMaster screw and you're fine, it isn't safety critical. For a bolt pattern holding a flange on to a 100kJ pressure vessel, you'll want material reports, test reports on the certified fasteners, torque specs (or load indicating washers), redundant load paths, etc. etc.
It is always a good idea to ask for proof of the certification (certified material test reports, load test reports, etc.) This first of helps prove to the company you mean business, and second of all if there is a hole in their testing you will see it. It's also important that you know what you're asking for when you call out a spec. This means buying the spec, and reading it! If you call it out, you'd better know what you're getting in to!
Q_Goest said:
The company I work for has over 15,000 employees worldwide. They won't purchase valves, machinery, pipe flanges from a third world country, or any product without a thorough review. That's not to say we never do so, but it becomes painfully obvious that many suppliers of 'certified' products haven't gone through sufficient testing and don't have enough field experience with their products to be certain that the product is properly designed for the service it's put into.
It's similar at my company (~2500 employees). For non-safety critical designs, you're open to use most vendors. For safety critical designs like pressure vessels or heavy frameworks, you have to use a company-approved vendor, and our engineering standards require safety notes/calculations and certifications.
You want to be able to trace original design intent from concept (I'm putting a bolt here to hold this part) to design (it will see this expected load based on my calculations, and will have this safety factor) to the bolt used (material, certified material test report). If the bolt fails and it turns out you incorrectly calculated the load it will see, you're at fault (and probably fired). If you call out a bolt but didn't ask for certified test reports, you're still at fault. If the bolt fails because the company gave you counterfeit ones and falsified test reports, you did your homework and they're screwed.
Still, you want your design to have a reasonable safety factor to help prevent rare events; an FMEA can help with that too.
Q_Goest said:
Take for example ASME Code testing of relief valves. The RV manufacturer brings a half dozen valves to a test lab and they pop test them, chart the discharge coefficient and assuming all the paperwork has been submitted properly and inspections of company procedures are approved, the valve gets the ASME stamp of approval. Not to say it's easy, but the valve doesn't need to be tested and shown to withstand repeated openings. Doesn't need to show that it won't have problems with galvanic corrosion, etc... I've seen top of the line RV's fall apart after a few uses. I won't mention manufacturer names, but ASME relief valves that fail after repeated use are not uncommon.
Did they fail safe (as in open)? I don't think the ASME code requires a blowoff valve to survive multiple openings... If the valve opens even once a safety investigation should probably figure out why it happened and make sure it doesn't happen again.
This goes back to knowing what the spec asks for and means, and accounting for it in your design intent. If your design intent requires that a valve open and close a lot, perhaps another kind of valve should be specified.
Q_Goest said:
The same thing goes with just about any industry. There's a manufacturer of valves in the propane industry for example that has a huge market share. This particular manufacturer has a terrible reputation with our company, and we've stopped using most of their products. Yet they are certified by UL, ASME and other industry standards.
Yes I can see if a company has had multiple QC issues, it would be removed from the qualified vendor list.
Q_Goest said:
I'm not suggesting that all suppliers are poor suppliers. The vast majority of valve manufacturers make solid stuff. But there's too many valves with various certifications on the market today that are downright dangerous. Evaluating components such as these to ensure they're properly designed doesn't take long, and doesn't make the company that investigates the design liable either. The supplier is still liable. It just reduces the number of bad valves in your system.
It's good advice to not take someone's word for granted. In most of the stuff I do it isn't practical to try and evaluate the design of a part based on our own engineering standards, there are just too many components in our systems and time is always against us. So instead, we rely on certified test reports on safety-critical items, engineering design standards (multiple checks by multiple engineers), safety factors, and qualified suppliers.
It is critical to always know exactly what the specification/certification MEANS, and how it affects your design intent. Also have in mind what the trail of accountability is; you don't want your design to fail and kill someone with the trail ending at you...
