Hi,
Good day to everyone.
I would like to get additional info on the value for friction stiffness factor in modal analysis. Recently for a project, i used a friction stiffness factor of 500 and got my first natural frequency as 4.5 hz. The client commented that use of 500 is not conservative and the value of 0 should be used.
When friction stiffness factor value of o was used the frequency was 1.1 hz which is below the requirement of 4 hz. I explained friction effects needs to be considered for modal analysis as it is based on engineering judgement and good engineering practice.The pipe will wag back and forth when friction stiffness effects are not considered and the fundamental frequency will be low which is not real.

Kindly give some additional thoughts on this subject.

Thanks & Regards,
SAM.

No, you can't do it this way. The "friction stiffness factor" is a tunning adjustment, to adjust your model under the applied load to match field observed response. If you don't have an actual system from which you have measured results, and an applied load, leave the "friction stiffness factor" at zero.

Using this parameter to force the 1st natural frequency to a value above 4.5 hz is an (incorrect) artificial adjustment with no justification.

I agree that the stiffness factor for friction is used to crank your model closer to reality and it's good to have the actual system to provide that reality.
But...
If you look at a mode of vibration that illustrates the frictionless slide on a +Y support that has a large normal load, it's hard to accept the "friction factor equals zero" results; especially when it requires additional money thrown at the system to stiffen it up.
Personally, and I have no physics to back me up, I like 1000. Perhaps only because it sounds more arbitrary than 500. The results, as far as I've seen, are similar.

Hi Dave,
Thanks for the details.
I was of the same thought as to when there is a large normal on a +Y support and a value of 0 stiffness factor used in modal analysis,it will make the pipe slide freely without any resistance whereas in reality, friction will resist the movement. The fundamental frequency will be low and in order to increase it, stress analysts add additional hold down, stop, guide supports and make the system more stiff.
A value of 1000 is reasonable and gives the realistic approach to the effect of friction stiffness in modal analysis.
Many thanks for the experts in sharing their knowledge.
Regards,
SAM.

Hi Dave,
Good day to you.
A puzzling question is "No physics to backup the value of 1000 used in Friction Stiffness factor"..... So how can we proceed with modal analysis citing an explanation for the value of stiffness factor used.

Would be helpful if your thoughts are shared on this topic.

Regards,
SAM

The 1000 factor just means friction is being applied / is being considered, based on...
- standard industry practice,
- best approximation of reality,
- experience,
- engineering judgement

Dear Clarke,

Thanks for the precise answers.

Regards,
SAM

I fully agree with Richard, that we cannot arbitarly select a value of 500 or 1000.The friction stiffness factor has great effect on the support loads in dyamic analysis(time history).If you use an arbitary value of 500 or 1000 ,even simple rest support will generate a horizontal loads much greater than contributed by friction factor (0.3 for steel to steel contact) which is unrealistic itself.
Frictional load=Normal force x .3 .
The factor of 500 or 1000 gives unrealistic loads and system response so i feel we should use value of 0 or 1 only .

Dear Learner,
Thanks for the details.

Kindly go through the dynamic modal analysis presentation given by Dave...It will clarify the value to be used for Stiffness factor.

As clarke pointed out, it is engineering judgement and doing analysis considering real scenario that finally prevails.

Again the arguement given by some is "Value of 0 is conservative" but it is wrong to use the value...I am sure on this..... We are not doing overdesign by being conservative but trying to simulate the real behaviour of pipe....

Learning is happening day in and out trying to match the real with virtual..

Regards,
SAM.

I'm reviving this old thread,

If we are assuming friction stiffness for Modal and/or Time History analysis, is it possible to design support base on the assumed stiffness?

Any idea?