Plate design with max deflection below 5 microns

[Anurag K]

Hey everyone,

I have a project that I am working where I am designing a plate mounted/fixed to a heavy duty rotary actuator; the actuator is fixed. A 100kg blocked is clamped to it and upon it a 300kg weight is clamped. The whole plate length is 1.6m long with the actuator sitting 0.4m from one end.

My objective as given by my supervisor is to design a plate which when loaded has max deflection of 5 microns.
I modeled it as a cantilever beam and did some hand calculations and the thickness came out to be 350mm which was an immediate no. The plate thickness has to be within 100mm

How do I model this in SolidWorks CAE, what boundary conditions do I give it? I was suggested by my senior engineer that when two parts are clamped tight, it adds some stiffness. How do I factor all that in? I have attached a picture for reference.

TL;DR - How to model the Plate in the below picture accurately in SolidWorks CAE in the way it is mounted and clamped. Please suggest.

Material of Plate is A36 HRS

I really appreciate you all for taking time to read this length post.

 

[rb1957]

I think it won't work within the constraints given to you.

If you're serious about controlling deflections, then you'll probably need a different design, like a plate with stiffeners.



"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[btrueblood]

I agree with Greg, that the stiffness of the actuator itself, its bearings or whatever allows it to "actuate", will dominate the deflection.

 

[rb1957]

but the actuator is "fixed" ?

but we need to consider the weight of the plate (maybe others did, I didn't in my quick calc). At 100mm thick I think the 1.2m portion of the plate weighs 1000 lbs (500kg), the 1.6m plate is close to 1500 lbs

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[ThomasH]

The above design is part of a precision level CNC grinding machine to grind parts that are going to be used in an engine.
Therefore the max deflection is set at 5 microns; as the max deflection increases the parts won't be accurate to their design.

Are you sure that the 5 microns are a deflection and not a very stringent tolerance? Just to try to understand this, if you don't meet the deflection limit, what will happen?

I think you need to describe this better to get any reasonable answers. Just based on your figure, the dimensions are in meters and deflection is 5 microns. It seems a bit odd.

Is this a school assignment? I can't help asking.

 

[rb1957]

I took "micron" as 1/1000th millimeter (as opposed to the imperial micron, 1/1000th inch)

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[Anurag K]

@ThomasH The above design is part of a precision level CNC grinding machine to grind parts that are going to be used in an engine.
Therefore, the max deflection is set at 5 microns as the max deflection increases the parts won't be accurate to their design.

 

[GregLocock]

Fine, but until you can provide a sketch with the details of dimensions and boundary conditions then I don't think anybody can help you further than rb1957 has. I agree 5 microns is tight, but if a typical engine tolerance is of the order of 1 thou, that's 20% of that. To measure 5 microns in 2m will need a temperature controlled room and some fancy metrology gear, Faro arms won't cut it.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[GregLocock]

For comparison here's the spec on a 110mm thick optical table

horlabs' 110 mm (4.3") thick Nexus® breadboards are offered in a variety of lengths and widths from 24" x 36" (600 mm x 900 mm) to 48" x 72" (1200 mm x 1800 mm). They feature all-steel construction, excellent thermal stability, and broadband damping optimized for each breadboard size. The 5 mm thick stainless steel top and bottom skins have a matte finish and the top skin is precision-machined with a surface flatness of ±0.1 mm (±0.004") over any 600 mm x 600 mm area.

So, 20 times the tolerance you are proposing over a much smaller area.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[ThomasH]

I took "micron" as 1/1000th millimeter (as opposed to the imperial micron, 1/1000th inch)

So did I and since the scale of the plate is meters and the deflection limit is microns (1 millionth of a meter). As Greg pointed out, how can you check this? A small change in temperature will will be enough to cause problems.

 

[Anurag K]

The plate is mounted to the actuator using M8x1.25 bolts in bolt-hole-diameter of 280mm, qty 10.

@ThomasH
There is no way to check that deflection. If it's deflected way past the max. deflection it will be reflected in the part that is being grind on the machine - poor quality finish, out of tolerance specs.

 

[rb1957]

Ok, back to your original post ...

I agree with your hand calc (that 100mm ain't going to cut it) ... so modelling hoping for a good result is futile.

You're asking how to model a very simple structure. How much experience do you have ? have you read some of the manuals ? done some of the tutorials ??
The very simple structure lends itself to a simple 1D element model, but the scale (500mm) also lends itself to 3D solid models.

As for modelling the clamped interface, ask yourself what this clamping does (to the structure) ?
It is as though an infinitely rigid component is attached to the beam ... a rigid element might work. An easy way would be to have a single element under the mass.
As for the attachment to the "actuator", I'd model as a single fixed point. You "could" model all 10 fasteners and go to town on it, but I think that's a lot of wasted effort.

"Hoffen wir mal, dass alles gut geht !"
General Paulus, Nov 1942, outside Stalingrad after the launch of Operation Uranus.

 

[IRstuff]

There are commercial autocollimators with 0.1 arcsec capabilities, so measuring 5 um/1 m ~ 1 arcsec is feasible. There is also literature claiming 1 picoradian deflection measurement.

TTFN (ta ta for now)
I can do absolutely anything. I'm an expert!

https://www.youtube.com/watch?v=BKorP55Aqvg

faq731-376 forum1529 Entire Forum list

http://www.eng-tips.com/forumlist.cfm