Slab on Grade for MRI 1

[lutein]

I am designing the slab on grade for MRI, which is about 12kip in weight. To size the slab on grade, I have considered:
1. punching shear from the heavy magnets.
2. make sure the slab is thick enough for the post installed anchor embedment.
3. FRP reinforcing is provided for temp and shrinkage.

I would appreciate if you could point out what other factors have i missed?

Thanks.

 

[TGilk]

@structuresguy, it really depends on what your objectives are. First, it is hoped that the building will last 30 - 50 years, but the equipment will be swapped-out several times during that timeframe. Designing to the minimum allowable standards (or as close as you can get to the tolerances) may very well be acceptable for today's installation, but what about the next equipment replacement?

Another important factor is that MRI scanners are not appliances, not in the conventional sense. The magnetic fields from the MRI interact with the building, and the ferrous content of the building reshapes the magnetic field that the MRI uses for imaging. The less the building's design messes with the MRI, the better images it is capable of producing. This means that the MRI equipment is clinically / financially viable to the owner for a longer period, extending their equipment replacement cycle. Given both the costs of the equipment and lost revenue during replacements, this is not insignificant.

And speaking of revenue, today's insurance company reimbursement rates provide razor-thin margins for most MRI providers (here I'm speaking about the US). Given that gross revenue and operating expenses for these devices are very, very high, interruptions to service can have grave consequences to the provider's bottom line. Anything that can help assure trouble-free operation of the MRI scanner and minimize servicing time is very important to the owner.

And lastly (and specifically on the issue of structural design of floors), I'm right now reviewing a design of an MRI suite where the structural designers must have looked at the allowable steel values for this particular MRI scanner...

Scanner specs for steel

B.F.F. PSF
0" 0
3" 2
5" 3
10" 8
13" 20



The structural designers' floor structure (elevated) is...

B.F.F. PSF
3" 1.8 (#4 rebar, 18" oc / ew)
4.5" 2.5 (centroid of 2" metal deck under 3.5" conc)
10.5" 12 (PLF centroid weight of framing stiffeners under magnet)
12.5" 43 (PLF centroid wight of beam)

Yes, 1.8 PSF is under the 2 PSF threshold. And yes, the 2.5 PSF of the metal deck is under the best-fit curve of the tolerance whose next point is 3 PSF at 5". The problem is that these aren't independent tolerances... they're cumulative. If you have 1.8 PSF of rebar at 3", you've used 90% of your allowable threshold. That means that you only have 10% left at any other distance! That's 10% of 3 PSF at 5", or 10% of 8 PSF at 10", or 10% of 20 PSF at 13".

Will the magnet still scan? Probably, because the design criteria are a bit conservative. But certain types of scans may not work as well, or be as clear, or take longer to acquire. The client will yell at the MRI manufacturer, who will provide the latest software upgrade, scratch their heads, and say 'what else would you like us to do?'

 

[271828]

Playing devil's advocate to some extent here. Not trying to claim specialized expertise in the matter, but I've dealt with quite a few MRI on SOG and elevated floors.

Designing to the minimum allowable standards (or as close as you can get to the tolerances) may very well be acceptable for today's installation, but what about the next equipment replacement?

I don't know about the experience of everyone else, but I'm usually given "the manufacturer's criteria" to design around. Those are usually quite difficult. I've never tried going back to the arch and attempted to sell the idea of designing around more extreme criteria for the possibility of the future being more severe. I suspect that they'd shoot back something like: "Well, how do you know that the next generation of MRI won't be more robust and have easier criteria?" From a common sense standpoint, I would've assumed it'd be just as likely to go that direction.

The magnetic fields from the MRI interact with the building, and the ferrous content of the building reshapes the magnetic field that the MRI uses for imaging. The less the building's design messes with the MRI, the better images it is capable of producing.

I don't see how to approach this from the strl engineer's standpoint. We're given "the criteria" that will make the MRI "work." It sounds like you're proposing that SEs should go to the arch with an argument that sounds something like "Well, I don't believe these criteria that GE's physicists and engineers came up with are stringent enough. I think we should spend even more of the owner's money and more of our time to try and design to something more stringent for [fill in the blanks]." Maybe my experience is odd, but I don't think that would sell, easily anyway.

This means that the MRI equipment is clinically / financially viable to the owner for a longer period, extending their equipment replacement cycle. Given both the costs of the equipment and lost revenue during replacements, this is not insignificant.

Do you have monetary numbers for this? Option A vs Option b. If so, then I can see how one could put together a lunchtime presentation for the arch's office and explain to them why one should try to significantly exceed the mfr's criteria. Otherwise, it sounds like an assumption only and is too vague to weather a serious argument.

Anything that can help assure trouble-free operation of the MRI scanner and minimize servicing time is very important to the owner.

Again, do you have some service time numbers, Option 1 vs Option B, that one could show to an architect? If not, then it sounds vague and I'm anticipating it being a pretty hard sell.

One issue I have with argument such as these is that MRI have been installed on many, many floors since the mid 80s. The firms I worked for have put many of them on cip concrete and steel-framed floors and I don't think they've ever heard of a problem with any of them. We just follow whatever mfr criteria are given. I'm sure there are examples of problems, but we've never had anybody alert us to any of them.

 

[TGilk]

Before anyone else says it, let me clarify my PSF value for my #4's at 18" oc / ew... the magnetic distortion effects are highly localized. It's inappropriate to average the weight across large areas away from the magnet. 1.8 PSF is the 'worst case' scenario of the 1.5 feet each direction which are bounded by the 'picture frame' of crossing rebars. The averaged load, over larger areas, drops to half that value, 0.9 PSF.

 

[TGilk]

@271828... Great questions, all!

I don't know about the experience of everyone else, but I'm usually given "the manufacturer's criteria" to design around. Those are usually quite difficult.

If we can design for the loads, and simply substitute FRP or carbon fiber for the reinforcing for a CIP slab, both design and cost differentials are negligible.

It sounds like you're proposing that SEs should go to the arch with an argument that sounds something like "Well, I don't believe these criteria that GE's physicists and engineers came up with are stringent enough. I think we should spend even more of the owner's money and more of our time to try and design to something more stringent for [fill in the blanks]."

I would hope that the architect would have enough respect / fear of botching the install of this multi-million dollar piece of equipment that they'd be receptive, particularly if you accept my prior assertion that, at least for CIP, the burdens are negligible. Yes, structural steel buildings and elevated floors are more difficult, but not deal-killers. And I am saying that I don't believe that the GE criteria are stringent enough... not for the life-cycle of the building. Apart from the space requirements, vibration sensitivities, metal tolerances, dead loads, have all slowly ratcheted up. Designing to the limits of the tolerances, today, may mean tearing out major portions of a suite for the next generation of equipment.

Do you have monetary numbers for this? Option A vs Option b. If so, then I can see how one could put together a lunchtime presentation for the arch's office and explain to them why one should try to significantly exceed the mfr's criteria. Otherwise, it sounds like an assumption only and is too vague to weather a serious argument.

I wish I had the means to collect this information! I've heard it enough times, anecdotally, that I believe it (or I've swallowed my own Kool-Aid). As an example, a few years ago there was a brain imaging center for a state university. The project was a stone's throw from the hospital, and the brain imaging center and the hospital got very similar magnets from the same vendor at about the same time. There should be no real difference between them, right? Except one was shoe-horned into an existing hospital suite with conventionally-reinforced floors, and all other sorts of potential interferences (within tolerances, all), and the other was installed in a purpose-built suite with minimized ferromagnetic materials in construction, had a floating SOG with FRP. The MRI vendor service engineer remarked that the magnet in the purpose-built suite for the brain imaging project produced better images and required less periodic 'tune ups' than the comparable scanner in the hospital.

The firms I worked for have put many of them on cip concrete and steel-framed floors and I don't think they've ever heard of a problem with any of them.

The difficulty in quantifying this is that the problems manifest as imaging equipment troubles, and not building system troubles. Even if there were grave problems, they show up in the output of the MRI scanner. Unless the MR manufacturer engineers go back and assess the construction of the suite, it could be that nobody would ever connect those dots (this is purely a illustrative and not meant to cast aspersions on your firms' designs).

If structure is 10% of the cost of a building, and MRI technical spaces make up 10% of the floor area, and structural changes to accommodate special requirements of MRI represent a 10% construction cost premium, this contributes 0.1% to the cost of the building. Contrast that with the $1.5 - $2.2 million that one MRI scanner costs. Even if I'm 100% wrong, that seems a very modest 'insurance premium'.

 

[271828]

I wish I had the means to collect this information!

Believe it or not, I might be in a position to look at this. I have a PhD student looking at imaging vs structures. I'll try to think of a way to address this stuff, if not during his project, then as another one.

I very much appreciate your perspective on this subject. I'm very interested in it. So far, I've only been involved as a designer, but now am doing research on the subject.

 

[TGilk]

OK, these questions (and a parallel project) prompted me to do a little bit of digging about nearby metal effects on the MRI system operation.

If there is enough ferromagnetic metal (iron, nickel and cobalt) nearby to the MRI scanner, it throws off the 'tune' of the precisely-balanced magnetic field. Depending on your MRI machine, there may be a couple of different ways that this can be corrected.

The 'old school' way is to use shims, pieces of ferromagnetic material with well-defined magnetic properties. These are, quite literally, stuck to the body of the cylindrical magnet in masses and configurations that correct for the distortions of structural steel nearby. This is a laborious process, but can be accomplished during installation. As long as the building steel isn't changed, the shims should maintain a balanced magnetic field.

The 'new fangled' way (which is used in concert with metal shims) is active shimming, achieved by an array of electromagnetic coils built into the MRI scanner itself. These are tuned to produce counterbalancing magnetic field effects to the building steel.

One of the challenges to the old school metal shim method is that there is a direct correlation between the thermal properties of the shim metal, and its magnetic properties. If the temperature swings out of specification by even a couple of degrees, F, the magnetic balance for the MRI scanner gets thrown off.

The chunks of shim metal are also right next to the high-energy radio frequency magnetic field coils, which are typically actively-cooled because they generate so much heat. If the active-cooling system isn't up to snuff, or if the HVAC system sees the room temperature specs as 'suggestions' instead of 'requirements', the image quality from the MRI can start diminishing during the day as the heat from the coils builds.

All of this goes to show the interconnectedness of the elements in an MRI installation. 'Conventional' quantities of steel in the structure require that the MR equipment manufacturer adds metal shims to the magnet, which reduce the operational tolerances of HVAC and equipment cooling systems. The HVAC then becomes a linchpin (through this long chain of causation, begun with the structural design) to image quality.

As I said in an earlier post, the inter-relationship between the MRI equipment and the building in which it's sited is unlike any other piece of equipment you can think of.

 

[apsix]

Has anyone mentioned stainless steel reinforcement?

 

[structuresguy]

I'm with 271828 on this. The manufacturer gives specific criteria to design our structure within. As long as we stay within that criteria, which always allows some steel, then we are fine. No reason to design to more stringent criteria for unknown future uses. These units keep getting smaller and better shielding. So the next generation units may have no limits to reinforcing in the slab, for all I know. If they have imaging problems, and my design is within their specified criteria, then the fault is their own. Either their unit is defective, or their criteria are eronious.

 

[TGilk]

@apsix Yes, stainless rebar was mentioned... and it's a fine alternative to conventional rebar as long as it's an Austinitic formulation of stainless (all stainless has iron in it, but some varieties of stainless lock the iron into a crystaline structure, Austinite, where it does not react the 'normal' way to magnetic fields).

@structuresguy The 'short' and 'wide' bore MRI systems are making the physical body of the MRI smaller, but by making the 'imaging space' larger, this actually puts greater demands on magnetic field balance. And as imaging capabilities go up, one of the ways that these improvements are leveraged is by imaging smaller and smaller structures. Think of it as adding a new lens to your microscope, going from 10x to 25x. Beyond magnifying the image, you also magnify distortions introduced by vibration.

Now, having said all of that, the MRI equipment vendors go to great pains to 'build in' corrective measures into their MRI scanners, in an effort to not ratchet-up the design burdens. These corrective measures, however, don't always coincide with the sensitivity advances. There are 'lags' in the protection, where a number of scanners, perhaps the next one your client buys, don't have the protections to compensate.