Thick Footing at Heavy Machine (No impact - High Deflection Control) 1

[structural3]

I need help to design a thick foundation to control equipment deflection, I understand that there are a lot of publications covering dynamics support design, but in this specific case, the impact load is very minimal from machine, and vendor specifies 10^-6 order of magnitude accuracy (in inches) for the equipment deflection. How can I achieve this requirement by design a thick concrete foundation (I am sure there are other way, if you know please recommend it)? (machine wt =150k and footing size is about 100'x40')

Assuming sombody uses very thick foundation as a model to design this base foundation, I would imagine that typical mat foundation design software could not handle the design due to the high thickness requirement. Since mat foot typically design footing as 4 noded plate element up to certain depth(I think thickness can go up to around 5'),I would need to model it as 8 noded solid element. Please recommed a software or approach to accomplish this goal.

What will be the design criteria for rebar design, assuming concrete thickness is 16' deep, can I still use 0.0018 for temperature/shrinkage design. I feel plain concrete may work.

IF you know any reference talk about this topic, please let me know.

Thanks very much

Janet

 

[CTW]

Can you post what the machine is/does?

What is the footprint of the machine?

 

[cessna98j]

Hi Janet - that is a lot of concrete! It seems to be excessive given the relatively small machine weight, especially if impact loads are minimal. I've typically seen a 2:1 concrete weight to machine weight ratio to keep deflections under control, and that's for machines with substantial dynamic loading. You've got about 9,600K for you 150K machine. Granted, I've never had to deal with such precision in deflection so take my opinion with a grain of salt. Either way, I would think that you could get by with such deflections using less concrete.

 

[paddingtongreen]

Is that 0.000001" of deflection of any point relative to any other point on the slab? Is it in an environmentally controlled situation?

I'd go back to the vendor and say, "Okay, you've told me what you want, now tell me what you actually need".

Michael.
Timing has a lot to do with the outcome of a rain dance.

 

[PMR06]

10^-6 inches?! Are you sure?

Just to get a feel for the magnitude of that... if a 200# man were to stand on a 1ft cube of concrete, he would "squash" it 0.0000046 inches... that's the level of accuracy they are talking about?

I hope the construction workers don't step near the final machine... might deflect it excessively!

 

[ToadJones]

Ironically, I have been a part of a project with a similarly massive amount of concrete for a horizontal boring mill.
I am not sure the deflection requirement was so stringent, however.
But, for precision machining on mills where the workpiece and milling head are not on the same base, the support deflection needs to be VERY small.

 

[ishvaaag]

First you need a clear statement of the requirement, differential 1/1000000 between what positions, what distance. Then you need some rational evaluation of the equivalent force if you plan to dimension the thing by statics, or the forcing functions applied at what points. A complete model including machine, foundation and soil is more within the scope of mechanical engineers working for nasa that the average structural engineer. If you plan to use equivalent forces, almost anything from plate with shear deformation to upper precision software can be used for your evaluation. Solid elements are OK, and RISA 3D, SAP 2000 better, or the ANSYS, ABAQUS etc will do fine.

You need also very precise measures of the underground soil properties, especially if tilting by whatever the cause is paramount. Think that in foundations spans between columns, or in spans of beams themselves, center to column points, 1/250 distortion is one of the most commonly heard values to avoid non-structural cracking. If your statement was also center to end, you have a 4000 times more strict requirement; if deflection to span, a 2000 times more strict requirement. The mere natural variability in the ground, combined with thermal effects may forbid to deliver the wanted amobility.

Most likely you need to control thermal effects, to get only controlled exposure, or bracketed exposure to thermal movement. Think that heatflow is going to have a daily and seasonal pattern and excess of heat or cold from one side means tilt. Some of the named software, I would have to look RISA-3D manual, may not do this well. If you limit the temperature gradient by thermal insulation you will also limit the tilt. Also, for thicknesses over 1 m, this insulation also will prove extremely beneficial to forfeit inner shrinkage cracking. You need to control to under 20º C the gradient anywhere within the mass in the hydration process to do that. So enlisting someone with good mechanical and thermal FEM modeling abilities is almost a must.

In all, it maybe difficult to meet the requirement. Devices like the very large array telescope mitigates the imprecisions that a single telescope has by distance between supports. The same absolute imprecision gets minimized by the effect of such distance upon the measures. A degree of control of such precision for locale differential settlement of such degree for a 60 tonne machine with some degree of mechanical effects must be thought, I think, state of the art.

 

[ishvaaag]

... and if inviable the machine must include controls not only to lock to its target, but to its moving support.

 

[tai99]

The problem is not the foundation. It (the machine) needs a support system that can retain its original position upon any type of loading. Maybe you can find help from ME and/or IE.

 

[structural3]

I agree that problem may not be achieved by specifying thick foundation, but at this time I could not find a better way to achieve the goal to meet vendor's high deflection criteria. (Please recommend another way if are are aware of it)

As structural engineer, the typical way we handle equipment foundation is by specifying foundation weight to be 2.5 to 5 times of the equipment weight, this includes some units with dynamic load. adding damping pad will be effective for high vibration case.

In this case, vendor specification requires that machines sitting on the finish floor level (on our foundation) can not deflect more than 10^-6 inches. I reviewed some design which specified thick concrete foundation which may end up to be 20' thick of concrete. I feel if this is the correct approach, then just replace concrete with steel plate then thickness can be reduced to be 2' since E of steel is 10 times higher than concrete's E. But designer thinks that the foundation is not purely under compressive, it involves bending etc., the spring constant is given by geotechnical engineer. I have the same feeling this design is very questionable.

Ishvaasg seems know well about this kind of design, I agree, that this specification may be achieved by mechanical engineer's design, however, as a civil/structural engineer, I have not used Ansys and Abqus recently. I am very familiar with Risa-3D and know SAP 2000 may not handle this except use SAFE program (but SAFE is typically a MAT foundation design, and it typically will involve plate element instead of solid element), Risa 3D was used to model thick foundation but, the feature of solid element model is very limited. Can you make more detail suggestion by giving some good references. I want to know if it's possible to use RISA but at this time I could not persuade myself to be comfortable with this program. (think about my plate element is 5'Lengthx5'Widthx18'Depth, depths is three times of horizontal dimension, Risa technical support says it's ok that this ratio can go up to 3 to 5, based on FEM book from KJ Bathe (but I could not locate this definition in his book).

Attached at the bottom of this email is the similar data for machine I asked.

Thanks

Janet

SPEEDMAT boring and milling machines
SPEEDMAT is the last family developed by Pama in the sector of boring and milling
centres. The highest productivity is guaranteed by high spindle torque and powers
combined with the high feed rates of the axes. A solution patented by Pama assures the
automatic compensation of the thermal expansions of the boring spindle.
Some further characteristics enhancing the great machine accuracy are: the thermosymmetric
structures with central head, the air/oil lubricated and thermally stabilized
spindle line, the linear guides with re-circulating roller pads, the preloaded kinematic
chains enabling high accelerations and axis speeds up to 25 m/min.
The SPEEDMAT family includes three base models (SPEEDMAT 2, SPEEDMAT 3 and
SPEEDMAT 4) with boring bar diameters from 130 to 160 mm, x axis travels up to 4500
mm, Y axis travels up to 3000 mm, Z axis up to 2700 mm and W boring axis up to 800 mm,
max. power of 65 kW . The machine configuration is T-shaped, with symmetric structures
made of cast iron assuring the thermo-symmetric behaviour of the structure. The rotary
table (max. loading capacity 25000 kg) is equipped with preloaded hydrostatic bearing.
The new SPEEDMAT line is characterized by a large versatility and can be completed by
additional modules and accessories making it a range of machining centres whose
automation level can be graduated, so that they can be used both in the machining of
single pieces and in mass production. Automatic head change devices and accessories for
5 axis complex machining, rack or chain-type tool magazines, pallet change automatic
systems are available.

 

[structural3]

ishvaaag,

Can you elaborate what you recommended as follows, especially about thermal effect. the machines is sitting inside of the building and is located in California w/o extreme weather condition. Please recommend a good reference to a structural engineer to handle this situation.

Thanks

Janet
--------------------------------------
Most likely you need to control thermal effects, to get only controlled exposure, or bracketed exposure to thermal movement. Think that heatflow is going to have a daily and seasonal pattern and excess of heat or cold from one side means tilt. Some of the named software, I would have to look RISA-3D manual, may not do this well. If you limit the temperature gradient by thermal insulation you will also limit the tilt. Also, for thicknesses over 1 m, this insulation also will prove extremely beneficial to forfeit inner shrinkage cracking. You need to control to under 20º C the gradient anywhere within the mass in the hydration process to do that(what does this mean, if concrete is 20' thick, how long will the curing process end so deflection will not be effected). So enlisting someone with good mechanical and thermal FEM modeling abilities is almost a must (Can you recommend someone?).

 

[ishvaaag]

Respect the part of the gradient in the hydration process the reference I find is one in an ACI Special publication,

SP-152, Design and Performance of Mat Foundations, State of the Art Review, Edward J. Ulrich editor, ACI 1995

the paper is SP 152-7 Mass Concrete Pour Techniques for the Mat Foundation of 1100 Alakea Plaza, by Dar, Bravo, Sarwar and Samada

Another, ACI SP-139, Durable Concrete in Hot Climates, Cameron McInnis editor,

paper SP 139-9, Concreting of thick section in the tropics, by Tam, Swaddiwudhipong, Mani and Lee

is a bit more specific on FEM approach then taken for modelization of a thick mat rebar included.

Respect modelization I think SAP 2000 can still be used, only that by not using something like SAFE you'll be left with the stresses. I once did the thermal analysis for a ceramic "arab" curved tile for roofs to ascertain what incidence the thermal exposure could be having in some ruptures of the same that were being studied. So you can have stresses, both in some axes and principal and decide if safe for the case for some criteria of maximum tensile strength, tresca or whatever.

Respect what happens with the heat passed or taken from the ground there is certainly some stational variation, that for horizontal homogenous soils should be vertical, and say for an embedded block that some exposed upper machine may be shading in part can make irregular and so some gradient extant from some high temperature point to some lower temperature point. By the description of the machine it turns that is likely it won't be in this kind of exposition but in a building enclosure. So, fortunately, in general, both seasonal and daily variations of temperature, that are seen not to be high in general nor penetrate much towards the conditioned parts of the building won't be much and I think will pale in comparison with the thermal ones of the hydration process. So placing the machine say 3 or 4 m from outer walls in a conditioned building will allow you to dismiss that part of the deformation since whatever it takes will stay stationary for a conditioned building, that is where such a precision machine must be except other conditions so mandate.

Respect Temperature variations and their effect see

Effects of Column Exposure in Tall Structures, Mark Fintel and Fazlur Khan, Portland Cement Association. unspecified year, taken from ACI Journal of 1965

Respect penetration of the variation of the temperature in the ground see the freely downladable program Climate Consultant.

Control of the hydration process will be helped by isolation of the pour. However, having compressive material in the sides of a tall footing, mat or pile is contrary to the favourable aspect that passive push must have against tilting of the foundation and so should in this particular case be discarded in favour of other lower heat of hydration processes, adding ice, whatever. Then you need to help yourself in your model adding also horizontal compressive only springs counteracting the tilt to represent the passive push reaction.

A review of what ACI has respect hot temperature pours, and a search in google for hot temperature concreting will help good advice about this thing, for certainly there are some books about (I have at least another, but not familiar enough with it to say in what would help to this design).

 

[ishvaaag]

errata ... 2nd paragraph from below ... compressible material.

 

[steve1]

My experience is in engineering foundations for small turbine generators. These are installed on mat foundations. Manufacturers tell us to provide a "rigid" foundation without specifying what that means. Our mechanical engineers tell us to limit the foundation deflection to less than .25 inch under operating conditions. We use RISA Foundation for these projects. The key parameters for us are the values for subgrade modulus and the loads that are a result of operating conditions. We do not consider the dead weight of the machine as an operating condition as any settlement associated with that has already occured prior to the machine operating. Generally we have a torque that we apply as a set of out of plane couples at the anchor locations and compute our deflection based on that. We generally vary the subgrade modulus over a range of values. We also use the 3.5 factor in determining a preliminary foundation size.

 

[structural3]

steve1, How thick is your mat foundation, as I estimated, if thickness is less than five feet, it's legible to model by Risa as four noded element, once the width/depth ratio (element width, les assume is 5', that means depth can not exceed 5'), I don't think it can be defined as plate, instead it shall be solid element which requires 8 noded element in FEM model and Risa is very limited for this modelling (Risa does not even give node forces). I agree the dead load shall not be considered. I can revise the Risa model to see the results. Can anybody clarify about KJ Bathe's assumption (which Risa claimed), saying "plate element can be defined as Depth/Width ratio up to 3 or even 5, I could not believe, I thought, plate is plate, the thickness has to be small comparable to element width, once it passes ratio of 1, it shall be defined as solid element, let along ration of 3 to 5.

ToadJones, can you tell me how do you solve the problem based on your experience working on horizontal boring mill.

Shall I post this message if Civil/Structural guy can not help for the solution?

Thanks very much for all your help.

Janet

 

[ToadJones]

The foundation mass was spec'd by the mill provider. We had to detail reinforcement, anchorage.
They offered no explanation as to why such mass was required.

 

[tai99]

The precision you are seeking is far beyond the anticipated tollerances of general civil-structural works. As suggested before, you will need more expert helps from other sources -the owner, equipment vendor, engineering consultant in related field...

 

[CTW]

Initially I thought the word "deflection" was being used to describe "settlement", but now I think that's not the case.

Janet, please correct me if I'm mistaken about the terminology.

Is the vendor referring to the horizontal alignment of the machine? I don't see why it would matter if the machine foundation settled uniformly as long the horizontal alignment remained accurate.

 

[westheimer1234]

if this is a compressor foundation, i dont think staadpro or risa 3d is the a good software for this.. i have done this using staadpro both using plate and solid elements using time history.. honestly i have no clue what i was doing and was just using a previous calculation as reference.. no choice coz we didnt have anyone to do it.. when i moved to other companies,
experts on this use in house excel programs..
get a hold of Arya book..

you should be very, very careful during construction as i know a case in my previous company that the foundation cracked.. they had to hire experts to redo the foundation.. the crack was caused by temperature.. if remember right they put some device to control the temperature and add ice whenever required..

 

[structural3]

It's differential deflection (due to tilt etc. while operating to equipment) that we are concerned about, not overall settlement.

ToadJones, how much rebar are you putting in the foundation?