Although I have to say that in my local power case I live in a semi-rural area in a small enclave of private homes, each either on their own transformer or at most sharing a transformer with one other home (my home enjoys its own transformer). There is no heavy industry within miles of my location and the closest large loads are irrigation pumps for agriculture, which are still over a mile away and are not active at this time of year. Don't discount the effect of the inductive reactance of transformers and long lengths of buried cables at smoothing out spikes and dips in electrical supply.
Having worked professionally in electrical power generation and distribution I'm very familiar with the daily load demand curve. The local variation in electrical demand represented by my own home and my neighbors is pretty minor (lights, TVs, kitchen appliances, computers, this time of the year heat pumps) and I generally listen (at least critically listen) during the same general time each evening.
So I guess what I'm trying to say is that yes, daily variation in power supply quality is a real thing and can affect audio quality, but in my case is much less than what some folks suffer through. As a mechanism for significant differences in what I personally hear from Roon and XDMS I'm not sure that electrical power quality in the main factor.
Is it possible that Roon sound quality is driven as much or more by the user-load on Roon's back-end driving how much data Roon's servers have to deliver?
Steve Z