Understand a dynamic load representation 2

[nivoo_boss]

Hey everyone!

So I'm designing a factory that has a vibrating equipment on one floor. The load is given like this:

Do you have any thoughts what that "6 pol" might mean?

I'm not too familiar with dynamic analysis but am trying to educate myself a little in Robot Structural Analysis.

 

[human909]

According to my calculations stiffening the floor would have made things worse. As the floor's natural frequency was below 16hz.

Also in the thread you posted someone's advice was to get the natural frequency of the floor to basically double to that of 16 Hz - seems pretty insane to achieve that with steel beams.

Exactly, not practical.

With concrete there is more weight (lower natural frequency) and more natural damping so in such cases you be OK. My understanding is that in Europe it is common for such factories to be concrete floor. Where I'm located this is uncommon.

Detailing the dampers was relatively straight forward. I used these:

https://www.mecanocaucho.com/

 

[nivoo_boss]

Many thanks!

I think I'll still go with the high IPE400 beams just in case.

 

[ThomasH]

The frequency for the load is 16 Hz. You say that the frequency for the system is 8 Hz, I assume that is the lowest but are there any other modes can can be affected by this?

You mention time history analysis, In my opinion, vibrations are often better to study in frequency domain. And the mentioned concrete will add mass but it can also add stiffness.

 

[nivoo_boss]

...but are there any other modes can can be affected by this?

At least the technical documentation for the machine I have does not mention any.

 

[ThomasH]

At least the technical documentation for the machine I have does not mention any.

I mean that the structure will have more than one natural frequency.

 

[canwesteng]

You need to do some sensitivity analysis to see what happens if the concrete is composite. This kind of work usually requires several models. I'm not sure how robot works, but other FEA will need to have captured enough natural frequencies to properly get the dynamic amplification. I would try and capture either enough mass participation or at least natural frequencies up to 32 Hz

 

[ThomasH]

Isn't there a risk for resonance during start-up and stop of the machine? How often does that happen?

I would run an analysis for several frequencies up to 32 Hz and and check how the system responds.

 

[dgeesaman]

the supplier could not answer what the function for the dynamic load might be or what would be the allowable deflections for the support structure.

This is a red flag to me. I'm not a building structural designer but I am a machinery designer and it's essential for every part of the system to have this kind of data. Our equipment applies large varying overhung loads to the mounting surface and it's not at a regular frequency. There needs to be a crisp agreement between the machinery supplier and the mounting structure supplier. We require deflection limits for our machinery mounting surfaces and we provide loads to be applied to the structure to confirm suitability. There are other ways but the load/deflection is certainly valid and easy to communicate.
Perhaps the sizing is not yet available, or your immediate customer does not appreciate the importance of this data, but some exchange must happen. If nobody can speak to the actual loads, I would state on my drawings which assumed loads were considered in your design and your deflection limits.

It also makes sense to run resonance calcs if you have adequate data.

 

[ThomasH]

This is a red flag to me.

I had missed the comment that the supplier could not supply the load or the requirements. But I would consider that to be a problem.

I have had the problem that if a supplier is uncertain, they set the requirements so high that they are difficult or impossible to meet. The reasoning seemed to by: If things don't work out, it is not their fault .

 

[human909]

Remember the "supplier" here is likely not the manufacturer. (which is a well known Swedish company who has an excellent reputation). I see these sieves regularly across a broad number of industries.

nivoo_boss: "the supplier wants the openings between the steel beams cast in concrete under this thing"
Reinforces the above. Sounds like the 'supplier' in this case isn't that manufacturer. It would be abnormal for this equipment manufacturer to be involved in designating the required floor openings.

Isn't there a risk for resonance during start-up and stop of the machine? How often does that happen?
I would run an analysis for several frequencies up to 32 Hz and and check how the system responds.

Yes there is a risk but the consequences are unlikely to be problematic. I'd be focussed on the behaviour under steady state because that is where the machine spends 99.9% of the time. Regarding the start-stop patterns. That really depends on how they are used but having them running 24/7 is certainly quite common.

I would run an analysis for several frequencies up to 32 Hz and and check how the system responds.

I've never seen the equipment not vibrate at the specified frequency except during start up and slow down where it operates at lower frequencies for around 5-20s. The motor would get up to speed in 1 or 2 seconds. The sieve takes a little longer to reach steady state.

This is a red flag to me. I'm not a building structural designer... There needs to be a crisp agreement between the machinery supplier and the mounting structure supplier. We require deflection limits for our machinery mounting surfaces

This isn't exactly a delicate piece of manufacturing machinery. In fact, quite the opposite. The mounting surfaces and deflection limits of the structure are not really relevant to the sieve, it will quite happily vibrate and deflect because that is what it is DESIGNED to do.

The problem is if it induces excessive resonance in the structure. (As per my last thread, I've seen situations that are so bad it shakes the entire structure a 30mx30mx30x facility and could felt over 50m away through the earth.) Though in most installations I've seen or implemented myself there is some hard to avoid local floor vibration, but it doesn't extend beyond that.

If nobody can speak to the actual loads

The loads are listed in the first post and in the image posted two days ago. The function is sinusoidal, though the horizontal and vertical aren't in the same phase.

 

[WARose]

Isn't there a risk for resonance during start-up and stop of the machine? How often does that happen?

There is a very real possibility that you will pass through a resonance frequency (for either the equipment or the supporting structure) during start up. (Note I say "a" resonance.....since there are multiple frequencies involved for such a system.) I normally check the displacements with a variety of charts that include the short-term performance of the machinery. I typically do a time-history analysis to facilitate this.

 

[human909]

The maximum displacement of the beams under this dynamic load is 0,4 mm according to the analysis.

Whoah... Hang on a second! Have you thought about what 0.4mm means at 16hz! If I'm not mistaken, 0.4mm is quite significant. Try viewing the results in terms of acceleration rather than displacement.

Your displacement results are quite similar to my calculated displacement results in the other thread where the vibration was extremely 'severe'.

I don't want my floor or structure experiencing accelerations of 4m/s^2! These values where of the order I measured in the 'severe' case previously mentioned.

But this is achieved with some quite mean beams - IPE400 sections while the span is only about 5 m

I haven't done the calculation but that sounds pretty stiff which might make things worse in this situation. Hence your results.

 

[canwesteng]

The main concern during ramp up would be when it runs through the natural frequency of the springs or dampers it sits on, if it has any. This can create a large force that the structure needs to resist. The actual vibrations will be short lived so are unlikely to damage the equipment or cause concern to occupants.

 

[WARose]

The actual vibrations will be short lived so are unlikely to damage the equipment or cause concern to occupants.

Depends on the equipment.....and also the number of times it ramps up & down during it's use. I was brought it on one project (this is a specialty area of mine) where the equipment was having maintenance issues because of trasient motion that was not checked. This thing started up & ramped down at least several times every few hours. That adds up over a service life.

 

[ThomasH]

Yes there is a risk but the consequences are unlikely to be problematic. I'd be focussed on the behaviour under steady state because that is where the machine spends 99.9% of the time.

I've never seen the equipment not vibrate at the specified frequency except during start up and slow down where it operates at lower frequencies for around 5-20s.

Since the machine works at 16 Hz and from what I have seen the structure has it's first natural frequency at 8 Hz, there is a risk for resonance. Is that a problem? Probably not but I would check.

My approach would probably be something like this:

When I said that I would analyze several frequencies, I would run it in frequency domain, instead of time-steps i would use frequency steps. That means one analysis, "steady state", from 0 Hz to 32 Hz, step 0.1 Hz and also include all natural frecuencies in that range. Then I would plot displacement, velocity and/or acceleration (something) as function of frequency for the relevant positions on the floor. That would give me the level of the vibrations for the floor during start/stop and running and also an indication regarding the sensitivity since I would use the same load for all frequencies. Strong vibrations during start/stop can motivate a time-history approach since steady-state can be very conservative for that.

This is by no means the only "correct" approach but I often use this for vibration analysis. Depending on the results, it may suffice or more things may need to be investigated.

As for the displacement 0.4 mm, that is huge. I would be more comfortable with 0.4 mm/s . Are you sure about that calculation?

 

[human909]

As for the displacement 0.4 mm, that is huge. I would be more comfortable with 0.4 mm/s smile. Are you sure about that calculation?

I'm not sure whether you were question me or the OP here. The OP calculated values are 0.4mm as stated. My calculated values in the case of the existing installation was of similar magnitude.

I'm also sure about my measurements. In some locations the measured floor acceleration was 1.5g RMS with max observed at 3.0g. As already mentioned these vibrations could be seen, felt and measure over 50m away through the earth.

Since the machine works at 16 Hz and from what I have seen the structure has it's first natural frequency at 8 Hz, there is a risk for resonance. Is that a problem? Probably not but I would check.

Why check for the minor problem when the major problem is still present? The major issue here is going to be during steady state when there is continuing energy being pumping into the system. 8hz isn't far enough away from 16hz for it not to be a problem in steady state.

 

[ThomasH]

I'm not sure whether you were question me or the OP here. The OP calculated values are 0.4mm as stated. My calculated values in the case of the existing installation was of similar magnitude.

I apologize, that was far from obvious . I don't question your harmonic plots. I got the same results.

But I find the results so high that I wonder if they are correct, question for OP. If they are correct I believe the design needs some modification.

Why check for the minor problem when the major problem is still present? The major issue here is going to be during steady state when there is continuing energy being pumping into the system. 8hz isn't far enough away from 16hz for it not to be a problem in steady state.

I would check both 8 Hz and 16 Hz, and a number of other frequencies. All in one analysis. I don't quite understand what you mean with the minor and the major problem?

 

[human909]

I apologize, that was far from obvious. I don't question your harmonic plots. I got the same results.

And I apologize for if my responses have come off as rude in anyway. I suppose I'm a little forthright in my responses because I've dealt extensively with this EXACT machine as well as many of their variants. (The smaller 24hz machines are far more forgiving for most structures.) I've been involved in the installation of dozens of them and also been involved in rectifications in cases where others haven't suitably addressed these issues.

But I find the results so high that I wonder if they are correct, question for OP. If they are correct I believe the design needs some modification.

I agree that the OP should recheck his results to ensure he has accounted for all the relevant masses suitable stiffness. But I'm not shocked by his results. These 16hz machines are particularly bad.

I would check both 8 Hz and 16 Hz, and a number of other frequencies. All in one analysis. I don't quite understand what you mean with the minor and the major problem?

I'm referring startup/shutdown being the minor problem. I see the major problem as being the steady state as there is enough time and energy for the local vibration problem (the floor around the sieve) to become a global problem (engaging other significant masses in the structure) and causing serviceability issues or worse.


All that said I must say I still have plenty to learn from yourself and others and their approach to this type of analysis. I'm a long way from being an expert in structural dynamics.

 

[ThomasH]

And I apologize for if my responses have come off as rude in anyway. I suppose I'm a little forthright in my responses because I've dealt extensively with this EXACT machine as well as many of their variants.

No problem . I haven't perceived you as rude and I don't mind fortright. I haven't dealt with this exact machine before. But I was involved in a project several years ago with something similar but smaller. I don't remenber any significant issues, but it was smaller.

I have worked with machine induced vibrations but more with comfort vibrations from people walking on floors or slender pedestrian bridges. In that context these vibrations are very high. For the machines I have dealt with the acceptable vibrations can also be very small.

I'm referring startup/shutdown being the minor problem. I see the major problem as being the steady state as there is enough time and energy for the local vibration problem (the floor around the sieve) to become a global problem (engaging other significant masses in the structure) and causing serviceability issues or worse.

I agree that startup/shutdown is probably the minor problem and steady state the major problem. But if I assume that the result 0.4 mm is correct, that is a huge value for a non-resonant vibration. What will happen during startup when resonance occurs? And when we start to change the system, is it best to change the mass or the stiffness. And that will change the natural frequencie(s) but how will it change the vibrations, startup and steady state? That is the reason why I like include several frequencies in the analysis. And I would like to know what vibration level is acceptable for the machine to function properly. Maybe the governing criteria actually is an office floor 40 m from the machine .

Regarding learning, I think there is always room for learning more. And I think discussions in this forum is one way of doing that .

 

[human909]

I have worked with machine induced vibrations but more with comfort vibrations from people walking on floors or slender pedestrian bridges. In that context these vibrations are very high. For the machines I have dealt with the acceptable vibrations can also be very small.

In that aspect I've never dealt in any depth with walking induced vibrations. The only depth I've gone too is "basic rules of thumb" based on deflection limits and a few experiences where I've directly felt the discomfort of excessive deflection in a couple of my own designs and learnt from that. (One advantage of design and construct experience IMO, is you can readily experience and if necessary fix any design shortcomings regarding serviceability.)

What will happen during startup when resonance occurs?

Not much in my experience. This experience isn't calculative, simply experience standing next to the damn things. The these get get up to speed very quickly so the time crossing the resonance frequency is almost negligible. During shut down, it takes a little longer and from memory local vibrations do increase somewhat briefly (1-5seconds). But no more energy is being added so I'd argue it is a non issue. (though inherent in that argument is a certain amount of local inertia)

And when we start to change the system, is it best to change the mass or the stiffness.

In most installations this is likely to be a difficult issue to fix by adding stiffness. Removing stiffness could affect other serviceability/strength criteria. Adding mass is certainly a solution and one that I have considered in rectification works. I suspect, but haven't done calculations to confirm that if you have concrete flooring you stand a much better chance of acceptable behaviour due to the extra dead load. I'm used to steel construction with 5mm floor plate.

I've found the easiest way is through dampers. I've used dampers with a natural frequency of ~4hz with great success. I now specify these for all these 16hz installations. It isn't worth the engineering time or potential headaches if dampers aren't used.