COLD FITTING KNOW HOW? 2

[cranekiran]

Dear all,

I have a shaft(ASTM A-479 type 410) of dia 12.85mm which has to be INTERFERENCE FITTED ONLY into a bore of 12.5 mm.I want to cool the shaft to sub-zero temperature so that it will shrink in dia. while keeping the cylinder(which has the bore)in room temp.Is it possible to fit the shaft in this manner?;If so,how will I know which is the right temperature for my purpose?.I cannot heat the cylinder(DUCTILE IRON)bore since it has a rubber encapsulation on it's outer dia. Does someone know this type of cold fitting method?I need a solution pronto ;My work is now held up for this very problem.I would be very thankfull for any response!

 

[gearguru]

0.35mm is too much interference for the 12.5 mm dia.
There are methods to calculate the shrinkage by cooling down the shaft, but I believe it won't work for so small dia and so huge interference. You simply do not have enough material to shrink it that much.
gearguru

 

You need to hire an engineer to perform a calculation of thermal dilatation of the shaft. It is possible.

 

[cranekiran]

Thanks for the 1st response:-

I was also thinking that maybe 0.35mm interference is a bit high,but then what could be the right amount of interference?See,the cylinder is a butterfly disc which has to operate at 200 p.s.i (20 bars approx.)so I have got to think in the operational point of view too.thanks again!

Thanks for the 2nd one:-

Is there any book or website that tells us more about the thermal dilation?Can a exact amount of shrinkage be calculated in ANSYS?

I would still want to know the process of cold shrinking.I am told they use dry ice,Any confirmation?

 

[gearguru]

Think again about using the press fit only!
For the reasonable interference you can use the ISO system of fits, it is always good starting point for selecting any fit. You'll find them in Machinery's handbook. Select the press fit (something like 12.5 H7/u6) and find (in the chart) the actual tolerances for 12.5 H7 and 12.5 u6.
Then using these actual tolerances the forces you can apply to your parts can be calculated.
There is one more problem - the ductile iron's behaviour under heavy stress...

I do not now about web pages, but there must be some, and books too; I do not know ANSYS, and the book which I use for my calculations would be unreadable for you...

The formula from my book is:

ti = te - (deltaDmax + v)/(alpha * D)
ti (temperature needed to shrink, degree C)
te (ambient temperature, degree C)
v (clearance between parts when putting together, mm)
deltaDmax (max. interference, mm)
alpha (coeff. of thermal dilatation, 1/degree C)
D ( part diameter, mm)

My books also show: alpha = 11*10^-6 for steel,
23*10^-6 for Alu
(Translated into a non-engineer language: if you change the steel part's temperature by 1C, its diameter changes by 11 millionths of the original dia...)
These coeffs. are valid for a limited temperature range only. Therefore please take my calculation below more like a joke than a serious engineering work.
Also per my book: dry ice can lower the part temperature to -70C.
gearguru

 

[gearguru]

OOPS!
I did not paste the rest of my aswer - here it is:

v.v.:
I did not say that shrinking shafts can not be used for press-fitting parts; I said that 0.35mm is huge interference compared to the 12.5mm hole. Per my quick calculation, to shrink the 12.85mm shaft by 0.35mm needs to change the temperature by -2476 C what is not that easy. (when I was in the school, the absolute zero used to be -273.16C... )
The shaft dia is crucial in the amount of the shrinkage one can achieve.
gearguru

 

[GregLocock]

1) Assume a coefficient of friction of say .1, and then assume that only half the butterfly is present, at full pressure. Then work out the interference needed to generate the requisite torque. The elastic and dimensional properties of both the shaft and the butterfly will be important.

2) It sounds like at the very least you need to run to a bookstore (or Amazon) and buy Shigley's Mechanical Enginering Design (or whatever it is called), and then read it and understand it.

3) The exact shrinkage can be modelled in ANSYS. I am dumbfounded that anyone who has access to that program, and expects to get accurate results, is asking questions like this (sorry if that sounds rude)

4) Dry ice is one possibility, liquid nitrogen is also a very common freezing agent. It is very cheap, and very cold. For added excitement you can heat the butterfly in an oven as well.

5) perhaps more practically - get a series of shafts turnd up, starting at the nominal hole bore. The diameter of each should be .02 mm greater than the previous one. Try fitting each shaft to a butterfly. You will learn much.
Cheers

Greg Locock

 

[cranekiran]

Dear gearguru,

The info. that you have given me is really important for me,kindly give me few days time to analyze the details before I get back & give my results.Thanks once again,I am really overwhelmed with the detailed response you have provided!

 

[gearguru]

cranekiran,
Please be aware, that I only responded to your question about the shrinking the shaft for the assembly purposes. The interference of 0.35mm cought my eye, I wanted to prove that it is too much.
I fully agree with Greg how the problem should be solved. The original calculation should start with the load (axial, if any, or torque, or both) which has to be carried. Testing after the calculations is necessary, because there are too many variables in the calculations, in material properties, and in the manufacturing tolerance. The tolerances and reasonable interference were the main reason why I recommended to look on the ISO fits. Both - tolerances for bore and shaft and interference (or clearance, depends what hole/bore combination is picked) are "prescribed" in ISO fits. It is very good starting point when designing, the tolerances there are very reasonable for parts made from steel. I'd like to stress the words starting point. (There can be another advantages, but this would be a topic for another thread.)
gearguru

 

[gbent]

An old machinist rule for press fits is .001 inch/inch diameter. Press fits greater than .002 inch/inch become very difficult to achieve. For shrinking methods, liquid nitrogen works great for small parts. It's alot of fun to play with also. Dry ice works well also. With dry ice, baths with either acetone or alcohol improve cooling rates. The equipment to handle dry ice (none, just wrap it in newspaper and put it in the freezer)is much cheaper if this is a one time application. Also, the coeficients for thermal expansion as normally listed in handbooks do not always apply at subzero temperatures. Subzero parts can also have large ice buildups before assembly, depending on ambient conditions. The condensation freezes immediately.

 

[cranekiran]

gearguru,

O.K;you were right,0.35 is too much interference, so now the nominal dia of shaft is also kept at 12.5mm.I have decided to go in for H7(BASIC HOLE) & r6 combination.There is an interference of 0.02 to 0.03.but,the formula you kindly provided is still confusing to me.

DeltaDmax=interference=0.02
clearance=no clearance(because you have interference)=0,
ti = te - (deltaDmax + v)/(alpha * D)
so you have,

ti=36-0.02/0.000012*12.5

ti=-97.3 degree centigrade which is still too low.

Am I doing something wrong here?
Also,I am still very thankfull for the response you have given.

 

[cranekiran]

Dear greglocock,

you are correct in mentioning that factors like torque and the friction factors should be taken into account here.And the reason I had a doubt about ANSYS is because I am familiar with only the structural part of this vast and powerfull software.Now,I will have to start looking into the thermal part of it too.Thanks!I really do appriciate your response.