Add "missing" 2024.2 upgrades: metadata and load curves

[jpvelez]

What

Generate the following upgrades to ResStock's 2024.2 release, and make them available via bsf:

17. ENERGY STAR heat pump with elec backup + Full Appliance Electrification with Efficiency​
18. High efficiency cold-climate air-to-air heat pump with electric backup + Full Appliance Electrification with Efficiency​
19. Ultra high efficiency heat pump with elec backup + Full Appliance Electrification with Efficiency​
20. ENERGY STAR heat pump with existing system as backup​ + Full Appliance Electrification with Efficiency​
21. Geothermal heat pump​ + Full Appliance Electrification with Efficiency​
22. Full Appliance Electrification with Efficiency (i.e. appliances only)

Why

In the ResStock 2024.2 release, the 16 upgrades that NREL provided cover the following cases:

• Heat pumps alone (Upgrades 1-5)
• Light Touch Envelope alone (Update 16)
• Heat pumps + Light Touch Envelope (Upgrades 6-10)
• Heat pumps + Light Touch Envelope + Full Appliance Electrification with Efficiency​ (Upgrades 11-15)

But they don't provide:

• Full Appliance Electrification with Efficiency​ (1 upgrade)
• Heat pumps + Full Appliance Electrification with Efficiency​ (5 upgrades)

In many analyses, we want to calculate what happens to a home's bills if they fully electrify! But currently, we can't do that without removing the effects of Light Touch Envelope. Rather than having ResStock users do all of this manually over and over, let's do it once and publish it centrally via bsf.

How

It is possible to construct the time series for these upgrades by grabbing different columns of subsets from the existing upgrades, and combining them into new upgrades that have all the columns that are expected from each release.

Rather than pre-computing the time series for these "missing upgrades" for every bldg_id in the 2024.2 release, which would be expensive to compute and especially to host the resulting data, let's use a similar approach we used to implement hourly, daily, and monthly load curves (see #87): downloading 15 minute data and computing locally from there.

So, to compute load_curve_15min for upgrade 17, you would identify which upgrades have the columns you need to cobble together, and then efficiently procure those columns and stick them together.

If you need to compute load_curve_hourly for upgrade 17, you would compute load_curve_15min for upgrade 17 and then aggregate, doing this as efficiently as possible.

It'll take some thinking to do this efficiently!

Deliverables

• Docs page documenting these additional columns, why they exist, and how we constructed them
• Users should see these 6 additional upgrades under the 2024.2 release, properly documented, and denoted as "computed by Switchbox" with a link to docs page
• If user selects one of these 6 additional upgrades, they should be able to download correct data for the following file types:
  • metadata: Building characteristics and metadata
  • load_curve_15min: 15-minute resolution load profiles
  • load_curve_hourly: Hourly resolution load profiles
  • load_curve_daily: Daily resolution load profiles
  • load_curve_monthly: Monthly resolution load profiles
  • load_curve_annual: Annual resolution load profiles

[jpvelez]

@alexhyunminlee what would be involved in also doing this for hpxml and schedule file types?

[alexhyunminlee]

For upgrades 17 to 21 above:

17. ENERGY STAR heat pump with elec backup + Full Appliance Electrification with Efficiency​
18. High efficiency cold-climate air-to-air heat pump with electric backup + Full Appliance Electrification with Efficiency​
19. Ultra high efficiency heat pump with elec backup + Full Appliance Electrification with Efficiency​
20. ENERGY STAR heat pump with existing system as backup​ + Full Appliance Electrification with Efficiency​
21. Geothermal heat pump​ + Full Appliance Electrification with Efficiency​

I'm thinking of keeping the existing values for all the columns from Upgrade 11: ENERGY STAR heat pump with elec backup + Light Touch Envelope + Full Appliance Electrification with Efficiency, and replacing the following columns with the respective heating/cooling system upgrades:

out.electricity.cooling.energy_consumption - f64
out.electricity.cooling.energy_consumption_intensity - f64
out.electricity.cooling_fans_pumps.energy_consumption - f64
out.electricity.cooling_fans_pumps.energy_consumption_intensity - f64
out.electricity.heating.energy_consumption - f64
out.electricity.heating.energy_consumption_intensity - f64
out.electricity.heating_fans_pumps.energy_consumption - f64
out.electricity.heating_fans_pumps.energy_consumption_intensity - f64
out.electricity.heating_hp_bkup.energy_consumption - f64
out.electricity.heating_hp_bkup.energy_consumption_intensity - f64
out.electricity.heating_hp_bkup_fa.energy_consumption - f64
out.electricity.heating_hp_bkup_fa.energy_consumption_intensity - f64
out.electricity.mech_vent.energy_consumption - f64
out.electricity.mech_vent.energy_consumption_intensity - f64
out.fuel_oil.heating.energy_consumption - f64
out.fuel_oil.heating.energy_consumption_intensity - f64
out.fuel_oil.heating_hp_bkup.energy_consumption - f64
out.fuel_oil.heating_hp_bkup.energy_consumption_intensity - f64
out.natural_gas.heating.energy_consumption - f64
out.natural_gas.heating.energy_consumption_intensity - f64
out.natural_gas.heating_hp_bkup.energy_consumption - f64
out.natural_gas.heating_hp_bkup.energy_consumption_intensity - f64
out.propane.heating.energy_consumption - f64
out.propane.heating.energy_consumption_intensity - f64
out.propane.heating_hp_bkup.energy_consumption - f64
out.propane.heating_hp_bkup.energy_consumption_intensity - f64
out.load.cooling.energy_delivered.kbtu - f64
out.load.heating.energy_delivered.kbtu - f64
out.zone_mean_air_temp.air_source_heat_pump_airloop_ret_air_zone.c - f64
out.zone_mean_air_temp.attic_unvented.c - f64
out.zone_mean_air_temp.attic_vented.c - f64
out.zone_mean_air_temp.basement_unconditioned.c - f64
out.zone_mean_air_temp.central_ac_airloop_ret_air_zone.c - f64
out.zone_mean_air_temp.central_ac_and_furnace_airloop_ret_air_zone.c - f64
out.zone_mean_air_temp.conditioned_space.c - f64
out.zone_mean_air_temp.crawlspace_unvented.c - f64
out.zone_mean_air_temp.crawlspace_vented.c - f64
out.zone_mean_air_temp.furnace_airloop_ret_air_zone.c - f64
out.zone_mean_air_temp.garage.c - f64
out.zone_mean_air_temp.ground_source_heat_pump_airloop_ret_air_zone.c - f64

---

And then I would recalculate the following columns:

out.electricity.net.energy_consumption - f64
out.electricity.net.energy_consumption_intensity - f64
out.electricity.total.energy_consumption - f64
out.electricity.total.energy_consumption_intensity - f64
out.fuel_oil.total.energy_consumption - f64
out.fuel_oil.total.energy_consumption_intensity - f64
out.natural_gas.total.energy_consumption - f64
out.natural_gas.total.energy_consumption_intensity - f64
out.propane.total.energy_consumption - f64
out.propane.total.energy_consumption_intensity - f64
out.site_energy.net.energy_consumption - f64
out.site_energy.net.energy_consumption_intensity - f64
out.site_energy.total.energy_consumption - f64
out.site_energy.total.energy_consumption_intensity - f64

---

Questions:

1. Difference between out.site_energy.net.energy_consumption and out.site_energy.total.energy_consumption
2. what are the columns with "lrmer", e.g. out.total.lrmer_lowrecost_15__electricity.co2e_kg - f64
3. Again, difference between out.electricity.net.energy_consumption - f64 and out.electricity.total.energy_consumption - f64

Thoughts on the course of action: @jpvelez , @brmurray ?

[brmurray]

data_dictionary_2024_labeled.csv

In a previous attempt at this I added a column called "functional group" to the ResStock 2024.2 Data Dictionary. The groups were [appliance, cooling, heating, hot_water, total], but to align with this ticket, I suggest we just use [appliance, hvac, total]. Note that not every data column (rows in the attached data_dictionary_2024.tsv) has a functional group.

By reading in the data_dictionary_2024_labelled.csv, you can create groups of loads to mix and match.

Upgrade	HVAC	Appliances	Shell	hvac_source	appliances_source
0	current	current	current	upgrade0	upgrade0
1	hp_min	current	current	upgrade1
2	hp_high	current	current	upgrade2
3	hp_ultra	current	current	upgrade3
4	hp_min, current backup	current	current	upgrade4
5	hp_geo	current	current	upgrade5
6	hp_min	current	shell_light
7	hp_high	current	shell_light
8	hp_ultra	current	shell_light
9	hp_min, current backup	current	shell_light
10	hp_geo	current	shell_light
11	hp_min	electric_high	shell_light		upgrade11
12	hp_high	electric_high	shell_light		upgrade12
13	hp_ultra	electric_high	shell_light		upgrade13
14	hp_min, current backup	electric_high	shell_light		upgrade14
15	hp_geo	electric_high	shell_light		upgrade15
16	current	current	shell_light	upgrade0
==17==	hp_min	electric_high	current	upgrade1	upgrade11
==18==	hp_high	electric_high	current	upgrade2	upgrade12
==19==	hp_ultra	electric_high	current	upgrade3	upgrade13
==20==	hp_min, current backup	electric_high	current	upgrade4	upgrade14
==21==	hp_geo	electric_high	current	upgrade5	upgrade15

[brmurray]

Addressing those specific questions...

1. Difference between out.site_energy.net.energy_consumption and out.site_energy.total.energy_consumption

   • site_energy.total is gross energy consumption, 'net' subtracts behind-the-meter PV
   • From the data dictionary...
     • out.site_energy.net.energy_consumption = Total site energy consumed [less] any power produced by PV or generators
     • out.site_energy.total.energy_consumption = Total site energy consumed

2. what are the columns with "lrmer", e.g. out.total.lrmer_lowrecost_15__electricity.co2e_kg

   • 'lrmer' = "Long Run Marginal Emissions Rate"
   • This is a scalar used to estimate emissions-per-kWh
   • Emissions = kWh consumed x emissions rate per generation type x generation_mix
     • The scalar for emissions rate per generation type comes from ANSI/RESNET
     • The generation mix (time-varying)is from NREL scenarios "using Cambium 2022 high renewable cost scenario with long run marginal emissions rate levelized by 3% over 15 years"

3. Again, difference between out.electricity.net.energy_consumption - f64 and out.electricity.total.energy_consumption - f64

   • same as the site_energy answer

In summary, when we mix-and-match scenario, I think we have to ignore the columns that are totaled across the entire year and site, since we can't parse out the hvac vs appliance bits. That is, we leave these columns blank, in my opinion.