Calculating strength of bolt on longitudinal flange of a cylinder

[Naylor7]

Hello guys,

I´m currently trying to analyse the strength of a bolted joint between two halves of a cylinder. The cylinder is subjected to torque on one end and fixed on the other end.
The top and bottom halves of the cylinder are connected by means of a flange and bolted connection as shown in the attached file.
Please help

 

[racookpe1978]

No, don't do it.

The gasket will be twisted and forced tighter at one area where the two flanges are forced together by the torque; and be twsited and lifted apart at another area. So, you will get too little force clamping the gasket material in one region (it will leak), too much at another area (it will deform and leak), and a motion in the areas between (it will leak).

Lock the two PV halves together with at least 4x keys or tapered dowel pins (machined in placed AFTER the final torque of the bolted halves is complete). The keys take the torque, the gasket material has enough problems just trying to stop the leak.

If you think you can take a metal-metal fit, look at the flange requirements (sizes, thickness, precision flatness, manufacturing expense, size of bolts, bolt-torquing precision and methods and expense, etc) needed to seal the turbine casing and compressor casing flanges.

 

[electricpete]

I didn't see any reference to a pressure retaining function. Maybe op can clarify the nature of the application.

=====================================
(2B)+(2B)' ?

 

[EdStainless]

IF there is no internal pressure then it should work.
If this is rigid enough to not deform then the load on the bolts are simply derived from the moment. However I would expect some deformation and the bending/shear loads would need to be added to the bolt load.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[Naylor7]

Thanks for the replies.
There is no internal pressure present. The assembled cylinder will serve as a torque-tube and therefore only needs to be able to resist the torsion applied to it. There may be marginal deformations (80 mrad) within the tube, however accuracy is very important.

 

[Jboggs]

If accuracy is important I would strongly recommend against this approach. Is there really a good reason that the cylinder has to be constructed from two half-shells? As the torque is applied, the actual motion the fasteners will be resisting is a sliding action. They will not be able to completely eliminate all relative motion between the two halves. If you must approach it this way then I would recommend you use at least four LARGE dowel pins (press fit on both sides) to resist the relative lateral motion. Then the only load on the screws is the one for which they are designed - tension. That means the connecting flanges must be pretty substantial to securely grip those dowels.

Are you sure about your units? 66,577 kNm? That's over 49 million foot-pounds. That's a lot! this is going to be a very substantial structure. You better make sure your stress analysis is accurate and thorough.

 

[Screwman1]

I see problems with yielding on the holes in bearing if the fit is tight enough to transmit the torsion through the bolts and if the fit is loose the whole this will be a sliding mess on the faying surface.