Hosting capacity maps vs distribution load serving capacity maps

Indiana Michigan Power publishes the assumptions behind its hosting capacity map, and they include

• thermal loading limits set at 100% of summer equipment ratings,
• only distributed energy resources between 150 kW and 7,500 kW evaluated,
• each circuit refreshed on its own annual schedule,
• the results are for preliminary review only and don't replace an interconnection application

If available capacity comes from a hosting capacity analysis, it was very likely calculated under minimum daytime load conditions, not peak. Hosting capacity analysis was created to answer a generation question: how much solar or other distributed generation can this circuit absorb before it causes a problem? The worst case for that problem, such as reverse power flow pushing voltage too high or protection equipment losing visibility into downstream faults, happens when the circuit's demand is lowest and the sun is out, not when everyone's air conditioning is running. So the analysis stress tests the lightest load hour of the day.

Adding load goes in the opposite direction, because the worst case for serving new demand is peak, not minimum. Hawaiian Electric's planning materials explain the classic principle, writing “distribution systems are planned to serve the predicted peak kVA, not the daytime minimum”. Whether the circuit has thermal headroom left in the afternoon on the hottest day of summer is the information needed to determine whether a new load fits.

This discrepancy has been contended over the past decade. A 2017 industry white paper from SEIA on grid modernization describes the older, cruder screening rules that hosting capacity analysis was made to replace. California's Rule 21 has used 15% of a line section's peak annual load as a threshold for triggering deeper interconnection review since the late 1990s, and an alternative measure used a percentage of minimum daytime load instead, since that's the hour when solar export causes the most trouble. The same white paper cites Sandia National Labs research finding almost no correlation between a circuit's peak load and its hosting capacity.

So, before treating a published capacity number as room for a new load, find out what condition it was tested with. A number that matters for site selection has to answer a peak load, load addition question, and if the source should clarify which one you're looking at.