PSD chart v's simple frequency

[styleruk]

Hello

Vibration being my weak area in engineering, I call upon thee to help please.

Can anyone help, I'm having trouble comparing a chart that shows a PSD path against a simple frequency and speed.
I've done a test on a componant to this:
33.3 Hz at 43.12 m/s2
20 Hz to 200 Hz sweep (20-minute reciprocating) at 21.56 m/s2
20 Hz to 200 Hz sweep (20-minute reciprocating) at 9.8 m/s2

I have to try to compare this to a chart that I have been given (see attached), I do not quite understand if this chart is worse than the frequencies I have tested to. At first I though if I squared the PSD numbers it would give me the speed, but I know this is way off. I know this chart shows a random frequency and is not as simple as a staight comparison, however, there must be a rough way of determining which is worse.

Can anyone help me here, I would like it if someone could tell me if the chart is less or more severe. I would like it more if someone could lay out or point me to a simple method of converting the numbers, by simple I mean a method that I could understand. If anyone could point me towards some literature that would help me decide, that would help as well.

Linked file here:

http://files.engineering.com/getfil...963-4940-a370-e0b613033a4d&file=PSD_Chart.JPG

Thank You
Simon

 

[jeyaselvan]

I have attached publication, probably in line with you have asked for...

I think you are looking for " Sine or Random ? " . Hope fully this coudl throw you some information.

Good luck

Jeyaselvan

 

[styleruk]

Thank you Jeyaselvan for that quick response, it made for interesting reading.
It mentions in that document that there is a 'Q', factor. This is something I do not have available, should I have it? if not, what should it be?

Regards
Simon

 

[jeyaselvan]

Q factor is a kind of amplification you have while at resonance. For a viscous damping model,Q=1/(2*ceta), where 'ceta' is the damping ratio.

It will depend on whether your system is lightly or heavily damped. You always have to measure that for a given system.

Jeyaselvan

 

[BobM3]

It's hard to know if the PSD would be more damaging. It's possible that you missed a particular frequency that (because of a resonance) might have produced some significant damage. PSDs are generally used for random vibration. With random noise you'll have many (all) frequencies being produced at the same time. At any particular point of time some combinations may prove more damaging then having just one frequency driving at a time.

There is a way to calculate the average energy of each frequency in a PSD. You would have to know something about the sample rate and frequency resolution that was used to generate that particular PSD.

 

[styleruk]

BobM3

I agree. I'm finding it increasingly difficult to get an exact answer, so far my answer is:
The random vibration could be worse, depending on the harmonics of the part. Although at first glance it looks less damaging, it could in reality be a lot more damaging depending on how the part reacts.
I feel it is a harsher test because it could find things that a sine wave test vibration would not.

To add to this, surely with that chart I would need a randomiser of sorts. Something that would give me potential numbers and could be calculated using a statistical analysis program. Or again, is this an unknown untill I record this from the part?

I'm begining to get the picture now and is a lot clearer.
Si