Yes, I did consider it, but did not invest that much time. Frankly, I newer could warm to Modelica, don’t know the exact reason. For this specific task/project, just to have one more language / environment besides the “main language for the everyday technical computing” was something I did not want, and it is quite different from a “normal programming language”, so not that quick to learn.
Ok, Modia is written in Julia, but it still adheres to Modelica’s philosophy (which, again, is perfectly fine in itself) and provides a single way of doing things (e.g. always using the IDA solver, or PyPlot for plotting). Also, it was (and still is) in a somewhat “experimental” state and does not provide any systematic documentation (AFAIK). All of this is no problem for a new, mostly cummunity-driven project trying out new ideas… but it’s not something you’d use to solve a real-world problem as fast as possible and need to be flexible to adapt to whatever may get in your way. So I decided to just use DiffEqs as base simulation framework and build a minimal “framework” around it, providing exactly (and only) what I needed (see this post).
In that post I also mention that ModelingToolkit.jl seems to be a very promising approach to “unify” such modeling frameworks by providing a “core engine”.