Ref. reply #24
Rex,
While it seems easy to add a reinforcing strip to the model, it actually involves incorporating special contact elements between the original and the strip, which can be somewhat tedious and tricky to work with. Besides, it is easy to say what the effect will be--the maximum stress will be marginally reduced and the location will move radially inward to where the ID of the strip winds up. I don’t see much point in doing that.
Obviously, the 38ksi peak result of the elastic solution is useful to show where and to what degree there may be structural concerns, but 38 ksi would never be generated there since the yield strength of the aluminum is around 10 ksi. The material would yield and larger local strains generated. The problem is that the endload of the closure is having to be carried by the thin material at the radius in bending, which quickly generates high stresses--as can be seen since the same endload is carried in tension in the cylindrical wall at a much lower stress level. To transfer the load into the cylinder without the bending, the closure needs to be much stiffer--added ribs across the end or a thicker, stiffer endplate. (However, a thick/stiff endplate would not expand radially as fast as the cylinder wall, leading again to bending of the cylinder wall at the junction, although probably less so than the existing arrangement). The tapered junction of a traditional elliptical tank head would also tend to transfer the load but with more compatible radial displacements.
As pressure is put into the tank as it stands, the stress at the radius would rise to the yield point, at which time the end cap would begin to permanently deform (balloon), changing the geometry, relieving the bending and moving toward a more tensioned stress state. The question is, as pressure increases, whether the structure would obtain a stable geometry before the material ruptures. (I may play around with an elastic/plastic analysis version of the existing model to see how it goes).
In any case, it would be prudent to do a proof test of the tank USING WATER to some pressure in excess of the service pressure before putting it into service. Doing this in stages and observing deflections and any “permanent” distortion may give an indication of whether it can be made to work successfully as presently constituted.