Pole top, pad mounted transformers

A distribution substation can house two or three power transformers. Each of those transformers can feed multiple buses. Each bus can serve multiple feeders fanning out in different directions. Each feeder branches into laterals, and each lateral eventually ends at a small transformer, pole-top or pad-mounted, usually somewhere between 25 and 75 kVA. These transformers step power down to the voltage a building can use. That's five layers of equipment between the substation that shows up on a hosting capacity map and the specific piece of hardware that determines whether a building can add a load.

A pair of transformers at a substation is usually sized and planned so that if one fails, the other can carry the whole substation's load for some period. The feeders coming off that substation aren't planned the same way. A feeder outage is usually restored by switching customers onto an adjacent circuit rather than by pre-reserved standby capacity. That means the headroom at a substation isn't a shared pool that each feeder can freely draw from. The headroom is contingency margin, held in reserve for a specific failure mode.

The inverse mistake is just as common. Adding up the stated headroom on every feeder off a substation and expecting it to equal the substation transformer's own rated capacity ignores that the feeders don't all peak at the same hour, so their combined worst-case draw is smaller than the sum of their individual peaks. The hierarchy is a series of separately engineered constraints.

Reporting requirements

The equipment at the bottom of that hierarchy isn’t covered by any published dataset. A Michigan Public Service Commission staff memo on hosting capacity maps states that service transformer results can't be readily produced from the hosting capacity calculations utilities currently run, because doing so would mean integrating the analysis with the utility's GIS system and, in some cases, its advanced metering infrastructure. These are two systems that generally weren't made to intertwine for this purpose.

Dominion Energy's hosting capacity tool documentation has a utility-scale hosting capacity map calculated from distribution lines and other facilities "not including any distribution transformers," while a separate residential generation map does incorporate the individual distribution transformer's size, using whichever is lower between that and the upstream line limit. Dominion had to put out a second, separate map just to bring the last transformer into the picture.

That last transformer is usually the limiting factor for a specific site. A feeder can show large headroom on a public map and still leave a building stuck behind a 50 kVA pole transformer that was sized decades ago for the load that used to be there. The US Department of Energy's atlas of hosting capacity maps notes that these maps provide general transparency into the distribution grid but don't address site-specific interconnection questions.

Nevertheless, feeder- and substation-level hosting capacity is still useful information about where the system has room, and it's usually the only thing available without directly asking the utility. But treating any published capacity figure as an answer about a specific address is a category error. The question to ask about any number on a hosting capacity map isn't "is there room here," but "room at which layer".