Right now, this method calculates:

• Unmet shower energy (mcpdelta T when T_delivered < 105 F)
• Unmet shower time
• Total shower time

Unmet shower energy I feel is the most complete metric, since it accounts for how far off from the desired temperature you are. However, no one really has a sense for if say 2 kWh of unmet showers is too many, so this way you can calculate the % of shower time where loads are unmet.

We could also add metrics related to ALL hot water usage, or all fixtures. We'd want to account for the different desired delivered temperature for appliances, but it wouldn't be hard to add metrics like Unmet Energy for all fixtures.

@shorowit: I just did a major refactor of this, no more duplicate code. Thanks for the suggestions. It'd be good to know if you think this is a decent way to organize things at a minimum if/when you have a chance to take a look, we're looking to use this with the 120 V HPWH relatively soon.

I know one of the questions here was why we're not seeing unmet loads if we just drop the tank size to 1 gal, and I think I know why.

If we have a 1 gal WH but 4.5 kW elements, we basically have a small tankless water heater. Pretty much like the under sink units you see some manufacturers make, and the reason the "very small" UEF bin exists. So something like these: https://www.rheem.com/group/rheem-tankless-electric-water-heaters-professional-classic-series-tankless-electric-rtex-3-5-rtex-06-kw/

These units can support flow rates up ~ 1 gal/min, depending on mains. Here's an example from those spec sheets (the number is element capacity in kW):

Even with stochastic events but hourly timesteps, you might end up with low enough flow rates that the "tankless" water heater we use for testing can keep up with the loads. So I think it's not actually unexpected that we might not see unmet loads if you just reduce the tank volume, depending on what the draw profile looks like. From our "base-dhw-undersized", here's the fixture flow rate in gal/min:

So a test case with 1 gal but 4.5 kW elements might actually keep up with the shower draws, and we'd get 0 unmet loads. I think it's much better for our test cases to just use small elements/burners, rather than small tanks. But to make sure this works right, Joe helped make a new test file that has a 1 gal tank and 4.5 kW elements. With hourly timesteps/events, that still produces 0 unmet loads, but with 10 minute timesteps/events you do see a tiny bit of unmet loads (one hour). So we DO now have a test case for this if we want to make sure both situations will lead to unmet loads.

@shorowit: the other big outstanding question here was splitting out SSB from fixtures, since this is sensitive to shower flow rates. I'd vote that we not do that for now, but we could consider it in the future. I'm actually curious what feedback we get about this feature, since there's no real standard way to quantify unmet WH loads, and if people agree with me that we really should just focus on showers.