Rules for endurance test

[PaulLag]

HI there
Hope everybody is safe
I hope to be posting in the right part of the forum, nonetheless I haven’t found a specific thread about endurance tests.
Please, I am contacting you concerning following topic.
A standard defines an endurance test targeted to simulate a minimum of 12 years of life and 8000 working hours.
A confidence is determined by the formula
C = 1-R^(*(r *Lv)^b
Where
C = confidence
R= reliability
r = factor for acceleration of stress according to Woehler
Lv = the number of lives (therefore the extension of the standard),
b is the Weibull slope
n is the number of samples

the standard assigns a specific number of cycles equivalent to the 12 years (ex 30.000 cycles)

following is my question:
Let’s assume I’d like to target the test in order to simulate the double of time, meaning:
24 years of life and 16000 working hours.
Is this possible to simply double the number of cycles – therefore 60.000 – and apply the same formula for confidence ?
Ex:
Let’s assume that with a reliability of 98% and confidence of 80%
Considering r=1, b=1, with 10 samples, I’d need to test 8 times the estimated life, therefore 240000 cycles in the case of 12years
In order to extend to 24 years, shall I test the same 10 samples for 480000 cycles ?

Please, could anybody suggest me a specific literature or standard related to endurance tests ?

I thank you in advance

 

[rb1957]

"the standard assigns a specific number of cycles equivalent to the 12 years (ex 30.000 cycles)" ... gust loads (on a building) ?

I would have thought that the length of the life affected some of your confidence parameters.

another day in paradise, or is paradise one day closer ?

 

[PaulLag]

Hello, thanks for your answer.

Indeed the aim of my question was to know if it is possible, starting from the endurance stated from the standard, how to adapt the standard requisite to assure a double life.

When I look at the confidence formula

I would say that confidence and reliability is something that we set.
Woheler factor and Weibull slope is something that should not change in my understanding

Said that the number of Lives is our output, I deem we need to act on the samples numbers.

the question is: is this argumentation solid even to a range that is double of what is stated in the standard ?

Many thanks

 

[rb1957]

"A standard defines an endurance test targeted to simulate a minimum of 12 years of life and 8000 working hours."

what standard ? government ? company ? project ??

who's certifying the work ? government inspector ? EOR ??

My concern would be how the varibales in the confidence calc vary with length of service ?
and/or the fatigue spectrum ... maybe it's simply doubling the cycles, maybe the cycles are dependent on time ??


another day in paradise, or is paradise one day closer ?

 

[PaulLag]

Hello

the standard is a SAE standard
Link

the aim would be to realize in a third party laboratory the test.
Fulfilling the points of the standard would fulfill the declared hours and years

 

[rb1957]

1) that's a link to the SAE site, and not the standard you're trying to meet.

2) If you claim to be compliant to the SAE standard, who checks ? who makes the statement that you're compliant ? If you want to extend the standard or interpret it (for say double the life it states) then these would be the people to talk to.

3) It seems really "odd" to me that a standard should define "a lifetime" as 12 years. In our business the standard sets what to do for the life that you define for your project. That said, doubling the life In years and cycles) seems reasonable. The question to me is how this change affects the confidence calc ... a longer life could have higher or lower confidence.

another day in paradise, or is paradise one day closer ?

 

[PaulLag]

apologizes.

The specific standard is J-28 42
Link

the confidence is given from the formula that I wrote in the opening thread.
Point is: this refers to the half of the years I would like to consider.

But if you could recommend me any other standard to refer concerning durability, your are highly welcome

Thanks in advance.

 

[rb1957]

ok, that's a link to the stub ... still none the wiser.

It sounds like they're saying a typical life for one of these things is 12 years, which is fine. It sounds like it's a question to SAE (or the authors of the spec).
As I said previously doubling the life means doubling the cycles sounds reasonable ... but are you compliant to the SAE spec if you do this ? Do you put some "weasel words" around your compliance ?

It's not the confidence formulae that I'd be concerned about ... but do the parameters change as the life gets longer ? Maybe they're already really conservative (they look that way to me). I guess you're concerned about the pressure cycles on the tubing ?

another day in paradise, or is paradise one day closer ?

 

[PaulLag]

Hello

As a matter of fact I was trying to understand more concerning endurance engineering, and specifically neither the laws that rule this branch engineering or some specific standard where endurance is stated.

I have searched a lot of standards, nonetheless the SAE is the sole one where a duration is stated.
For this reason I was first looking for a new one stating a longer period, or a law to correlate the different periods.

Moreover, knowing the theory behind endurance evaluation would have helped me to realize if it is possible or not to extend the various requirements in order to enlarge the stated duration period.

To be honest, differently from you, to me these are only a number of cycles, I have no knowledge to state these requirements are conservative or not.

And, yes, my concern refers either to pressure cycle in the tubing but also the thermal stress.

Thanks

 

[rb1957]

ok, so this is a different question. What it seems you are interested in is general fatigue "safe life" analysis ?

Research "safe life analysis" and Miner's rule to get started, maybe "Goodman diagram" too.

another day in paradise, or is paradise one day closer ?