Question - I-Beam Stiffening/Strengthening, Bolted connections rather than welded? 3

[human909]

Hi all. I though I'd throw this question out to the masses for a few ideas.

I need to seriously stiffen some members due to some vibrational issue. Cost of steel isn't really the issue it is more labour and looking for the easier and quickest solution to stiffen some existing beams. I'm mostly looking at three options and are after opinions but other suggestion are welcome.

I also want to know how effective bolting beams together can be and whether adequately rigid shear connection can be achieve to ensure that the beam acts compositely with the added member. I would presume slip critical bolting would be required and so make the whole process more difficult and possibly impractical.

Options I'm considering:
1: Welded T piece
2: Welded Wide flange (advantage being downwards facing welds)
3: Bolted member identical to existing (advantage being avoids site welds, but might be negated but number of bolts and slip critical requirements)

Overall I'm thinking option 2 is best but I'd like further input, particularly regarding option 3. Thanks.

 

[dik]

I was involved with a vibration problem about 30 years back and the vibration consultant adhered an anisotropic rubber material between the existing steel beam and a steel plate 'glued' to the bottom... it wasn't a serious vibration problem, but the solution worked like a charm...

So strange to see the singularity approaching while the entire planet is rapidly turning into a hellscape. -John Coates

-Dik

 

[human909]

I was involved with a vibration problem about 30 years back and the vibration consultant adhered an anisotropic rubber material between the existing steel beam and a steel plate 'glued' to the bottom... it wasn't a serious vibration problem, but the solution worked like a charm...

Thanks and I appreciate the input. That sounds like a solution for higher frequency and lower amplitude.

I believe** that the amplitude of theses vibrations are far too large to be solved by rubber alone. It is either more stiffness, more weight, or more isolation with dampening. To put things in perspective the machine requires 80mm clearance around it. Steady state operation probably has ~20mm of amplitude to it. So yeah, if you want to dampen things properly you are going to have to have a decent mount of travel in you "spring and dashpot" setup.

**I'm stretching my area of expertise here. So I'm trying to pretext everything here with a suitable level of lack of certainty. I've dealt with this variety of equipment at least a dozen times. However retrofitting a solution is a new challenge. On countless matters of structural engineering I have more to learn. Dynamic behaviour, bring it on! But if I find myself out of my depth then I have some external consultants on speed dial that might be able to help (though charge appropriately).


Hopefully the interstate site visit will shed more light on things. You can get plenty of answer by just looking at things. Even more answers by touching and following the vibrations. And modern mobile phone accelerometers with the right software can do wonders in putting some numbers to your observations.

This issue has to be solved. So as far as I'm concerned it is an interesting structural challenge. My default position is to throw a bunch of steel at the problem to stiffen things up and hope it goes away. But it might involve a little more thought than that.


Oh and this thread seems to have attracted some great input from some regular forums members that I respect highly. Thanks. Like I have said I will provide updates even if it is just for the benefit of the community rather than asking more questions.

 

[TLHS]

It sounds like amplitudes are high, implying that frequency is reasonably low. Just FYI, especially for higher frequency loads, the local load path can end up being an issue. I've seen installations where you've got a pump skid bolted down to something nice and solid and it turns out that the vibrational issues are coming from the fact that the local pump skid beams are vibrating torsionally or the fact that you don't have stiffeners in the skid and the top flanges are bending as the anchors pull on them ends up causing problems.

My gut says that the first option is better.

Do you have a nice solid diaphragm or a stiff load path to a stiff horizontal bracing system to help with anything lateral?

edit: Man, if you've got 20mm of movement you've got a ways to go. Is the 80mm of movement in non-steady state just in ramp up/ramp down, or does it happen during other times?

 

[CrabbyT]

I'll let CrabbyT speak for himself but I know where my concerns for cracked welds come from.

In the context of my comment, I'm assuming OP is talking about a big reciprocating pump, or a conveyor that moves substantially massive things which also just so happens to be accidentally tuned to the frequency of a structure, and causes vibration to the extent that personnel in the area feel more than uneasy, as well as vibration that causes huge cracks which necessitate carbon fiber wraps around the pedestals that support the conveyor (oddly specific enough... LoL?)

Genuinely curious. Where do your concerns for cracked welds come from?

 

[human909]

It sounds like amplitudes are high, implying that frequency is reasonably low.
...
edit: Man, if you've got 20mm of movement you've got a ways to go. Is the 80mm of movement in non-steady state just in ramp up/ramp down, or does it happen during other times?

The 20mm movement is a guess at the movement of the equipment, based on previous observation of this type of machine. The 80mm is straight of the equipment spec sheet and I'd presume it is the non steady state ramp up/ramp down.

The frequency of the floor during operation is 16hz and I believe about 1.25mm of amplitude but that is VERY much to be confirmed. It certainly is highly unpleasant for those standing nearby.

Just FYI, especially for higher frequency loads, the local load path can end up being an issue. I've seen installations where you've got a pump skid bolted down to something nice and solid and it turns out that the vibrational issues are coming from the fact that the local pump skid beams are vibrating torsionally or the fact that you don't have stiffeners in the skid and the top flanges are bending as the anchors pull on them ends up causing problems.

Yes. I will very much be focussing on this. The equipment is on a 'stiff' equipment stand. But I do have queries about the connections of this to the floor and also its role in accentuating any lateral movment. (Lateral vibration of the equipment is about 30% of vertical vibration, but it could be part of the issue.)

Do you have a nice solid diaphragm or a stiff load path to a stiff horizontal bracing system to help with anything lateral?

The flooring is a steel deck. It should be stiff enough, but this will be looked at. Last time I had a more minor issue with a vibrational piece of equipment I traced the problem back to an unbraced bay! (Though this was never fixed and I was not responsible for the steel.)

I'm assuming OP is talking about a big reciprocating pump, or a conveyor that moves substantially massive things which also just so happens to be accidentally tuned to the frequency of a structure

It is a sieve. The contents of which aren't too heavy but when it weighs almost 3T and it is designed to vibrate as much as possible problems can arise.

 

[HTURKAK]

Dear human909 (Structural)(OP),

I was not around for the last four months. Having worked more than 30 yrs ago for the design of coal washing plant ( with vibrating sieves ) , i would like to share my opinions..

- The sieve supporting beams should not be rigidly connected to the frame of the bldg str . Moment connections is not preferred to avoid vibration of the whole str.

- The sieve should have spring dampers ( you may search for spring dampers suitable for mining application)

- The frequency of the supporting beams should be almost double the frequency of the sieve.

- The mass of the supporting beams more important rather than inertia . In this case, you may consider adding welded heavy plates to the flanges .

Good luck

 

[human909]

UPDATE.

I've visited the site where this is an issue. The behaviour is so severe that you can visually see the vibrations in a cup of water on a bench in an office that is 50m away and completely isolated from the source by 50m of earth and 20m of elevation. The same frequency was confirmed by measurement.

The RMS acceleration is generally around 3.5m/s^2 around the equipment, some places noticeably worse. Actually displacement is far less and can be readily calculated from the 3.5m/s^2 and 16hz, but regardless it is a severe vibration even if we are talking sub 1mm displacements.

Stiffening really does not seem to be a suitable and viable solution. Based on harmonic computation (I let the computer do most of the work) you really need a hell of alot of stiffening to get any benefit. Adding mass and reducing the natural frequency of the floor is more viable, just as effective and less expensive.

Spring dampers seem to be the best starting point. Spring dampers AND added mass are extremely effective (around 4T per Sieve for a total of 8T additional mass).

 

[TLHS]

Yeah, natural frequency being proportional to the square root of the stiffness really gets you when this is your situation. If you're trying to push out the other side of the resonance frequency and give yourself buffer past it, it quickly becomes impractical. You end up needing to stiffen by an order of magnitude or more to get in the reasonable range. Undertuning and adding mass makes sense. So does adding isolation.

The real question I have, though, is whether this actually resolvable? A screen of any appreciable size on shared steel doesn't feel right at all to me. I've only touched giant ones, so maybe my sense of the impact is a little exaggerated, but I question whether you can resolve this to everyone's satisfaction. I'd be doing some extensive expectation management with the client on this one.

Can you mess with the frequency of the screen at all to see what it does? It likely doesn't have the necessary controls, but if it has instrumentation that measures it you might be able to watch it on ramp up or ramp down. It'd be nice to know if you're seeing resonance in the frequency zone you're at, and if not where in the frequency range you're seeing it.

If you think it should be resonating at half the operation frequency and it's resonating at 3/4 of it or 1/4 of it, that could definitely inform your decision making.

 

[human909]

Yeah, natural frequency being proportional to the square root of the stiffness really gets you when this is your situation. If you're trying to push out the other side of the resonance frequency and give yourself buffer past it, it quickly becomes impractical. You end up needing to stiffen by an order of magnitude or more to get in the reasonable range. Undertuning and adding mass makes sense. So does adding isolation.

Excatly. And when you have a non simple criss-cross of primary, secondary and tertiary beams making up the floor we'd have to stiffen a large number of members.

The real question I have, though, is whether this actually resolvable? A screen of any appreciable size on shared steel doesn't feel right at all to me.

These screens on shared steel are a 'typical' installation. I would have seen dozens of these and by far the majority are on shared steel. These particular ones happen to be bigger and run at 16hz rather than the typical 24hz. It is the 16hz aspect that is causing the BIG issues.

I've only touched giant ones, so maybe my sense of the impact is a little exaggerated, but I question whether you can resolve this to everyone's satisfaction. I'd be doing some extensive expectation management with the client on this one.

Good advice. I've already expected to do so. I expect to recommend isolators to give XX% reduction and also suggest considerations of mass to provide a multiplicative YY% reduction. Based on previous experience I would expect isolators to be sufficient for the client. But I'd fully expect floor vibrations to be still quite noticeable, just not the bone jarring amounts it is currently.

Can you mess with the frequency of the screen at all to see what it does? It likely doesn't have the necessary controls, but if it has instrumentation that measures it you might be able to watch it on ramp up or ramp down. It'd be nice to know if you're seeing resonance in the frequency zone you're at, and if not where in the frequency range you're seeing it.

The sieve is resonating at 16.5hz, nominally it is 16hz. This is pretty much as expected for a 6 pole motor running eccentric masses.

I did consider messing with the frequency with a VSD. But I'm pretty sure the equipment itself is tuned to run at 16hz. So start changing the frequency of the vibrator will compromise operation. I will be following up with this to ensure that is the correct assessment. The other option is to reduce the driving force by adjusting the weights on the vibrator. That is an easy change, but that simply affects the amplitude and compromises the operation of the equipment. Some reduction maybe 25%-50% could be accomplished here. But I'm needing the big changes of the order of 90-95% reduction.

I believe I am on the right track. And isolators should be able to achieve a 90% reduction. I can go higher if required through mass addition and tweaking the machine.

 

[human909]

A quick update on this one. "Simple" spring and elastomer dampers were installed and significantly reduced the vibrations. I have yet to get back on site to measure the effect but the client is happy. I want to measure the values and see how close to the theoretical 92% reduction we have achieve.

Note even with a 92% reduction in energy transmission, the remaining 8% is still not insignificant. But this is a factory that where vibration is expected.

So what started off as a beam strengthening thought bubble went an entirely different route...