There are hundreds of pipe modeling packages. I've tested a couple of dozen and found significant problems with all of them. Some have node-count limitations that are too small. Some limit the calculations to equations that do not require friction factor (AGA, Panhandle, etc.). Many of them are "good enough" for most applications. The approach I take to evaluating a new piece of software is:
[ul]
[li]Look at the equations that it allows. Modern steel and all plastic is too smooth to ever get a relative roughness above the fully turbulent line on the Moody diagram, so AGA almost never gives good results these days. The Panhandle Reynolds Number range is pretty small. You need to think about what you are doing and determine the equation that is most appropriate. For gas flow I will always use an equation that explicitly uses a friction factor (which requires an additional iteration step) like the Isothermal Gas Flow Equation. This is even more important for liquid flow.[/li]
[li]Look at the method it uses to add a compressor or pump station. One well known program allows you to set the discharge pressure and hp and it floats the suction pressure. This is never the best choice.[/li]
[li]Look at the number of nodes. It should be at least an order of magnitude bigger than the biggest you can envision your system to grow.[/li]
[li]Don't look at the user interface. You can learn any user interface, and it is so easy to fall in love with the bling when the underlying arithmetic doesn't work.[/li]
[li]If you use a GIS, look to see if the model can accept a shape file to lay out the topography. Most can't, but this can save you a lot of struggle.[/li]
[/ul]
When you've found one that looks like it fits the bill:
[ul]
[li]Get a set of data from an installed system. Don't use averages, get a snapshot.[/li]
[li]Find a second set of data from the same system that is running at a different pressure.[/li]
[li]Build your model and tweak the levers that are available (usually it is pipe efficiency) until every source pressure is within 5% of actual data.[/li]
[li]Load your second set of flow rates and run the model without tweaking the controls. If the second run is within 10% at every point, then you can be confident that it has a chance of providing a valid evaluation of system modifications.[/li]
[li]If the second run is out of spec, adjust the levers to get all of the pressures within 5% and load the first dataset. If that run is within 10% then you have a model that is possible, but not high confidence.[/li]
[li]You can get a third dataset and see if it will match reality in either of the first two calibrated scenarios, but that lowers your confidence even further.[/li]
[/ul]
This is a lot of work, but you'll be living with this model for years and you'll never do the upfront work again, you'll just rely on model results. I've been using the same program since 1982. The user interface is amazingly awful, but the arithmetic is reliable. It was written for a 16 bit system and won't run on 64 bit. I'm looking for a replacement. Every one I've found so far has been really pretty, but they have all had one or more reason to disqualify them. I'm going to keep a 32 bit computer around until I find one that works.
The short answer to your question is that I've never run or tested Fluidflow3. They call themselves "The industry leading flow simulation software for pipe" so how could you go wrong with them? Well, 25 other companies also call themselves the industry leader so at least one of them is lying. I went to the FluidFlow web site and it is pretty. The video seems to be amazingly lacking in detail, but high on emotion. The module descriptions don't say much. I hate it when these web sites don't have a "marketing to geeks" section, but they don't seem to. I may eventually download one of the demo's, but not in any kind of time that would help you.
David Simpson, PE
MuleShoe Engineering
Law is the common force organized to act as an obstacle of injustice Frédéric Bastiat
