MTBUR/MTBF for variable populations

[s170609]

Gents,

I am trying to establish the best method for working out MTBUR when the component population varies within the time period. An example is when an aircraft comes into, or leaves service mid-year. When caculating the MTBUR for a group of components, their number changes as well as the total time flown.

instead of the formula used: MTBUR = (QPA x total hours)/ failures

QPA: quantity per aircraft

I should simply add the hours of all the components within the time period and divide by number of failures. Is this correct?

Another idea was to add the failure rates of the different aircraft then convert to an MTBUR..

Any advice would be welcome.

Cheers, Peter.

 

[IRstuff]

Individual failure rates are ostensibly more precise, and more useful in doing predictions and failure rate consistency verifications. A global failure rate might be adequate for generally planning and logistics support situations.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[FranAle]

Peter:
Sometimes is quite useful to segregate (isolate) the sample or population by ranges (same or similar failure rates) or by types inside de configuration of the aircraft, for example by their ATA 100 codification if it is commercial or by WUC(work unit code) for military.
Remember that mechanical, electrical or thermo machines´s behavior normally fails according their own history in a system or subsystem.

So after this, you may aplied more focused statistic and get more reliable figures.

Regards,

FranALe

 

[s170609]

Thanks for your Help.

Although FranAle you may have missed the point. My population is already homogenous - I'm not talking about all components from all aircraft systems (ata chapters); I'm talking about calculations where that population varies within the time period. Current formulas used assume a fixed population per unit time.

I think I'll look at intergrating the failure rates from smaller time periods... seems obvious now, but wasnt sure if there was an established method for dealing with this problem as I will be computerising alot of it..

Cheers, Peter.

 

[IRstuff]

OK, I wasn't paying sufficient attention. If we assume constant failure rate populations, then the MTBUR at any given time is essentially inversely proportional to the number of aircraft in service at at that given time. Classic MTBF is 1/failures_per_million_hours, so if a platform has a failure rate of 2000/million hours --> 500 hr MTBF, then 10 platforms have 20000/million hours --> 50 hr MTBF.

TTFN

FAQ731-376
Chinese prisoner wins Nobel Peace Prize

 

[s170609]

Thanks IRstuff, that's a good piece of info