1Pw3 + 3 DC expansions = 11kW/54kWh
For most people this will be about right
TL;DR: you can’t ignore the cost of that… it’s a relevant factor.
—
A few years ago I did a horribly complex spreadsheeting exercise to do a “what if” analysis of having different battery sizes. This analysis used my actual historical 5-minute energy flows (solar generation, household consumption) to redirect the energy flows as if I had a smaller or larger battery - and see what happens.
I did this as an economic cost/benefit analysis, assuming a base cost for a PW2 with 0 kWh (i.e. the cost of the electronics in it, gateway and the enclosure) plus additional $/kWh for battery capacity. How much extra grid usage did more (or less) battery capacity displace, hence what were the resultant grid costs?
The results were unexpected. It showed that the economic optimum (for me) was a battery of about 7kWh - i.e. half a PW2. Below that the grid draw started to increase a lot and the battery provided less and less benefit as it did not take much to fully charge it, and equally it didn’t take much to fully discharge it.
Above 7 kWh, the extra capacity was used less and less often, and because it has a fixed cost per kWh, it became increasingly dilutive in terms of ROI.
Adding additional capacity above 1xPW2 was economic insanity. I modelled up to adding 100xPW2. It really was a case of rapidly diminishing returns, chasing an increasingly long tail of solar generation / consumption edge cases, with quite expensive capacity installed that was hardly ever charged or discharged.
People with larger solar arrays and higher consumption would have an economic optimum larger than 7 kWh, but given how long the ‘statistical tail’ is, I posit that even in that situation it wouldn’t take much extra capacity to optimise the ROI.