Unfamiliar Reliability Term

[J.R. Strickland]

We recently received a Request for Information (RFI) for a potential contract that has the term "Inspection Reliability" in it. The requirement is listed as "...Inspection Reliability (120 days) = 0.95...". I know enough about reliablity engineering to be dangerous and have never heard of this term. Has anybody out there heard of it and what does it mean? I have "googled" it and that has not been much help. Thanks in advance.

 

[Steve Prevette]

"Inspection Reliability" is a term out there for - how good is your ability to detect a flaw, such as in this reference for bridge inspections https://rosap.ntl.bts.gov/view/dot/33883/dot_33883_DS1.pdf.

Quoting: As such, early investigations showed that inspections with only a single defect type gave enhanced defect detection and indicated that the inspection reliability decreases with each additional fault type. However, it is pointed out that the reliability of inspections with large numbers of fault types can potentially be increased by allowing longer inspection times. Gallwey and Drury also noted two other complexity factors that affect inspection reliability. The first of these is the number of separate points that must be inspected, and the second is the complexity of the standards by which defects must be measured.

Now would need a little more context to determine what the 120 days refers to. Perhaps the length of time to do the inspection? I would expect a full inspection of a large bridge could take 120 days . . .

 

[Bev D]

So many terms are mis-used and/or mis-appropriated with a hidden definition known only to the user. My suggestion would be to ask the potential Customer for clarity - in writing.

 

[Steve Prevette]

Don't know if the OP can share it, but even knowing what the RFI is for, and some more verbiage around the quote may help, but yes, even better is to ask the potential client.

 

[Miner]

Reliability methods may be used to analyze any "event" that occurs over time. In this particular case, that "event" is likely a failure to detect a defect.

There are numerous ways to specify reliability requirements, some being much better than others (e.g., MTBF being one of the worst). The one at play here (R) is one of the better ones. The reliability (R) may be specified in two different ways:

• R(time) = 0.##
• R(0.##) = time

The requirement (Reliability (120 days) = 0.95) follows the latter format. This would be interpreted as the Reliability at 120 days from time zero must be >= 0.95. Stated in another way, 95% of the defects must be successfully detected over a 120 day period.

 

[Steve Prevette]

Just to verify, we are not talking survival rate, but that at least 95% of any existing or new defects are detected within a 120 day interval on a certain population?

 

[Miner]

Just to verify, we are not talking survival rate, but that at least 95% of any existing or new defects are detected within a 120 day interval on a certain population?

Correct. The term "Survival Rate" is only appropriate when discussing product reliability or medical studies on people. When applying reliability tools to other "Events in time" another term should be used.

 

[J.R. Strickland]

Thanks to everyone for their feedback and thoughts. I've been travelling with very limited time for items not related to my travel.

The RFI (for handheld radio) also provides other reliability measurements in the same context of the inspection reliability...MTBF, MTTRF, and MTBCF. After reading everyone's comments, it does make sense that inspection reliability could be the ability to detect the failure in 120 days. In the end, as a couple of you have suggested, it seems to not be a very common term and I should probably just ask the customer requesting the information.