Looking for ITensor advice for modeling a Philips 5500 LatteGo in a simulation

I’m working on a small ITensor project in Pittsburgh and could use some guidance. I’m trying to build a simple simulation of a Philips 5500 LatteGo espresso machine, treating brew temperature, flow rate, and milk frother state as a few coupled degrees of freedom. I found some third-party specs online for this model, but I’m not sure how much to trust the numbers.

My main concern is picking a sensible tensor structure and bond dimension. The temperature and flow rate couple in a way that makes me think an MPS with a few sites would work, but I’m not certain whether the milk frother state is best treated as a separate site or folded into the same tensor. I’m also unsure how much truncation I can get away with before the results stop being meaningful.

Has anyone here modeled a small multi-component system like this in ITensor? I’d appreciate input from people who have experience choosing sites and bond dimensions, especially for a system where the couplings are only partly known.

I came across Philips 5500 LatteGo EP5544/94 Espresso Machine Review while checking specs, though I’m still unsure how reliable those numbers are for setting up the tensors. I’m leaning toward treating the frother as a separate site, but truncation limits still feel like guesswork.