This post has turned out to be rather lengthy, so I'll start with the question:
For equation based loading, is there a way to write the equation so that it is bounded, even in just one direction?
If you're intrigued and need more details, please read on...
There is a Femap Tips and Tricks demonstration called "Equation Based Loading" where a hydrostatic pressure load is applied to several surfaces of a liquid tank quarter model. The demonstration shows how an equation can be used to linearly vary the pressure from a maximum value at the bottom of the tank to zero at the fluid's surface. One vital detail is that the geometry is split horizontally at the liquid's surface; if the surface was not split, the load would decrease to zero as intended but then start to increase again above the intended fluid level. It can be a fair amount of work if the user wanted to test a slightly different fluid level because the geometry would have to be split and remeshed at the different fluid level. Analyzing the same model at different fluid levels is is a situation that I regularly encounter.
In my practice, I've been using the 2-point linear data surface instead of an equation. If a 2-point linear data surface was used with the liquid tank example, the user wouldn't have to split the side wall at the water level because the data surface would be structured to provide both upper and lower bounds for the applied load. To change the fluid depth, only the data surface's upper coordinate and its lower load multiplier would have to be edited. However, this approach also can cause problems: If the load is applied to a large number of surfaces, the time it takes to perform a force balance and output an analytical model increases significantly because the data surface must be evaluated for every surface to which the load is applied. Also, the Messages window becomes populated with one "Evaluating Data Surface [n]." message per surface to which the load is applied, making it harder to find messages that I actually want to see.
I've run tests on one of my models and with an equation it takes about 3.5 minutes to write the analysis model. With the same model and a 2-point linear data surface, it takes about 5.5 minutes. Going further, an Along Coordinates data surface with 10 points (for nonlinear loading of granular commodities) takes about 7.5 minutes to write. Thus, I'd really like to use an equation because I can output analytical models more quickly, but it's not worth it to me unless the equation can be bounded at the upper fluid level.
Whew. So...Does anybody know how to write an equation so that it is bounded? Or is there a different method that doesn't require geometry modification that is also fast?
Dear Dan (got your name from another posting here)
At SDC Verifier we spent a considerable amount of time solving this issue, and also the reverse one with a pressure applied to all the elements of the hull. (see figures below). As you already pointed out it is pretty hard to get all pressures correct and also in the correct direction. All tank pressure loads should be pointed outward, without flipping the orientation. (otherwise if 2 tanks share 1 wall this creates a problem)
To make it even more usefull, instead of iterating yourself, you can set the weight and density of the fluid and the correct waterlevel is calculated for you.
If you need this feature just once install a trial version which works for 45 days. If you use this more often consider buying the software and profit also from the other features you might also find usefull (report generation, plate buckling check, checks according to standards etc. see more on our website www.sdcverifier.com)
I hope this is helpfull for you.
Wouter van den Bos