Quote:
Originally Posted by jmblur
Let me put it this way  all engineering calculations are approximate. ALL OF THEM. The equations engineers use are simplified to closely model the situation at hand, but there's way too much uncertainty to ever get 100% correct answers.
The best you can do is to get a statistically good chance that an engineered object will not fail under normal design constraints. Things will fail due to things out of your control. Maybe the seal on the top of the pipe had a bird nesting on it, and the bird picked out an oring seal. Maybe there was a small inclusion in the metal at one of the welds that lead to a small crack, large enough to let water in. Maybe you just had dumb luck and the normal, random discontinuities and slip planes in this particular pipe happened to line up in a way that's extremely weak to the method of loading, which leads to a crack the lets water in. You can't control for everything  the best you can do is put a factor of safety on it and hope statistics are in your favor.
Now, there were certainly things they could have done to mitigate this risk  but hindsight is 20/20. I'll be interested to see HOW water got in in the first place  that's the failure right there. The pipe splitting apart due to ice inside it was pretty much inevitable, and splitting at a weld is a very normal place for a failure due to residual stresses and material property inconsistencies, which can lead to poor grain structure.
Now, I'm not saying this was designed correctly  or inspected correctly  but you can't engineer a perfect, thiswillneverfail product unless it's way overbuilt (which quickly becomes cost prohibitive) and part of an extremely simple system.

I agree with some of your points, in particular the "approximate" nature of engineering calculations. Except in this case that is not the issue we're dealing with. The tower wasn't designed to withstand the force of expanding water/ice, so the failure has nothing to do with a poorly designed connection or weld. The design issue with this tower is that no consideration was made for the situation where water gets into the tower. It's got nothing to do with calculations, strengths of materials, or poor welding / fabrication practices.
In school and in industry I've met and worked with many engineers who I would consider brilliant designers. Their calculations, and their methods to approaching complex design problems are spoton. However, not every problem can be solved with calculations.
As mentioned many times already in this thread, the towers need drain holes. Ultimately you could design the tower with rainhats and make it double wall, and all sorts of things to keep rain out, but even if all those things were done, why not still put in a drain hole(s)? Engineering is all about risk mitigation.