If your boundary conditions don't match up to the model, it doesn't matter what tool you use, especially if it assumes a pistonic source. I use AkAbak for my circuit equivalence modeling and the remainder is hard work and multiple prototypes or modifications followed by measurement and listening trials. If your tool is axi-symmetric, there are many features that you miss out on including circular and higher order modes. I also use impedance tubes to determine characteristics of drivers to use in modeling. There is a lot that comes into play when you move beyond just wave guides and the rear chamber behind the diaphragm can matter a lot.
Any model, doesn't matter what its underlying workings, won't tell you what something sounds like. It's handy for characterizing some issues and in accelerating design choices but ultimately it's just a tool and one requires honesty and the willingness to chuck non-working prototypes. On domes I typically start off with 4 different angle conicals and experiment with different throats before moving over to other shapes. Some domes then don't like the sharp transition, others love it. In a model they'd all look the same but in practice you need to do the legwork for that tweeter.