No of cycles in a sine test

[izax1]

I wonder: How many cycles are accumulated in a sine test? If we are sweeping through frequencies and we are hitting some natural frequencies in the process, will the number of (fatigue) cycles be the fn*60 cycles pr minute? Since the time at peak stress will be rather short, will that be real fatigue cycles, or should we count cycles within the rms stress range. sqrt(2)/fn-sqrt(2)*fn. The latmter will be more cycles at lower stress level. Anyone with any experience here? We ar sweeping at 2 Oct/min.

Appreciate any input from experienced test engineers.

 

[jimkinney]

You could calculate the actual number of cycles at each frequency fairly easily from the sweep time, but for fatigue calculations we conservatively use the highest frequency and assume the entire test was at this frequency. It's quicker too.

Jim

Jim Kinney
Kennedy Space Center, FL

 

[izax1]

Yes, but do you use the frequency for the peak response and count the cycles at the peak only, or do you count cycles for å broader band of frequencies like the rms value. Theoreticaly it will only dwell (or not dwell at all) for a very short time at the stress peak. The structure will surely accumulate cycles on each side of the peak. I suppose it also must be connected to the Factor of Safety used for the Fatigue life calculations?

 

[jimkinney]

Typically we test 20-2000Hz at 1-4 octaves per minute. So 100 to 400 sec * 2000 Hz = 200000 to 800000 cycles at peak stress.

Usually our loads are low enough that we don't worry about fatigue, but we have to check the box that we looked at it.

If it were critical, then you are correct about counting around peak responses. Search rainflow counting for another method.

Jim

Jim Kinney
Kennedy Space Center, FL

 

[izax1]

Thanks for your input.