Commercial electricity use is sometimes double counted, other fuel use is missing #394

jtlangevin · 2024-06-10T10:31:50Z

Per this spreadsheet check it appears that ~0.7 quads of what's currently under the commercial fuel type electricity end use MELs and technology other is actually non-electric energy use (in AEO, it's under the fuel types distillate, residual, LPG, steam coal, motor gas, and kerosene - so should probably be assigned to distillate in the Scout data). Also, it appears that Scout's electricity use for the unspecified building type is double counting water services under both the unspecified and MELs -> water services end uses, ~0.3 quads.

As a result, the Scout commercial electricity totals are about a quad higher than they should be, while the commercial non-electricity totals are ~0.7 quads lower.

This has important implications for full sector emissions and electricity analyses (electric loads overstated, emissions from non-electric end uses understated).

The text was updated successfully, but these errors were encountered:

* Remove secondary heating and wood stoves from being subject to exogenous switching rates, as they were not intended to be covered by these rates. * Allow users to set ‘tech_switch_to’ to the value ‘same’ and have the code calculate a switchover to the same technology under exogenous switching rates. This is sometimes necessary for accounting purposes – for example, if exogenous rates are represented and a baseline segment is included that is subject to those rates, but without a complementary measure in the analysis to switch that baseline segment to, the results will show a decline in that baseline segment without a complementary increase in a switched to segment, thus under-representing that baseline segment’s influence on the total results. * Add ASHP to technology list where zero costs are expected (in EIA new construction data ASHP costs are zero for cooling, all costs are in the heating end use.) * Remove water services as a distinct technology type as it appears to be counted already under “unspecified” (see issue #394). * Refactor calculation of hourly TSV load adjustment factors to ensure that each end use in the tsv_load.gz data will be leveraged (previous approach had resulted in skipping some edge case end uses in those data like clothes washing, pool pumps and pool heaters that are under ‘other’ in the Scout data). * Ensure that once cumulative measure-captured or cumulative competed stock fractions reach 1 in a given modeling year, these fractions are not updated in any subsequent years of the modeling time horizon. * Improve warning message for units conversion issues. * Add measure JSON check to ensure that measures that are set to fuel switch to electricity apply to non-electric fuels in the baseline.

* Remove secondary heating and wood stoves from being subject to exogenous switching rates, as they were not intended to be covered by these rates. * Allow users to set ‘tech_switch_to’ to the value ‘same’ and have the code calculate a switchover to the same technology under exogenous switching rates. This is sometimes necessary for accounting purposes – for example, if exogenous rates are represented and a baseline segment is included that is subject to those rates, but without a complementary measure in the analysis to switch that baseline segment to, the results will show a decline in that baseline segment without a complementary increase in a switched to segment, thus under-representing that baseline segment’s influence on the total results. * Add ASHP to technology list where zero costs are expected (in EIA new construction data ASHP costs are zero for cooling, all costs are in the heating end use.) * Remove water services as a distinct technology type as it appears to be counted already under “unspecified” (see issue #394). * Refactor calculation of hourly TSV load adjustment factors to ensure that each end use in the tsv_load.gz data will be leveraged (previous approach had resulted in skipping some edge case end uses in those data like clothes washing, pool pumps and pool heaters that are under ‘other’ in the Scout data). * Ensure that once cumulative measure-captured or cumulative competed stock fractions reach 1 in a given modeling year, these fractions are not updated in any subsequent years of the modeling time horizon. * Improve warning message for units conversion issues. * Add measure JSON check to ensure that measures that are set to fuel switch to electricity apply to non-electric fuels in the baseline. * Suppress check for 'other' and 'unspecified' building type input breakouts when these categories are also present in measure end uses (prioritize correct end use input breakouts, which are more commonly used.)

trynthink · 2024-06-13T23:16:15Z

From @jtlangevin, there appear to be three separate accounting issues.

The MELs->other branch is intended to represent whatever is leftover from end use # 10 ("other") in KDBOUT after the specific MELs techs are identified. But I can't seem to find a place here where those identified MELs are subtracted out from that other, thus it appears to just include them all (and thus they are double counted via this mseg). That's ~0.8 quads double counted electricity in 2050 (all MELs minus water services).

Similarly, in the unspecified end use, water services is included in that end use in KDBOUT (# 11) but I can't find where that is subtracted out from unspecified in that MELs block of the code. That's 0.2 quads double counted electricity in 2050 (water services).

KDBOUT includes other fossil fuels (# 4–8) under unspecified that don't make it into the Scout data because we don't have an other fuel for commercial (do we need to add, or perhaps this should this just be lumped with distillate since it's small and only really occurs in this end use so not worth the trouble?) So that's ~0.65 quads missing fossil in 2050.

Further investigation shows that for item 2, both "water services" and "telecom systems" are double-counted.

Update handling of commercial energy use data to address the three problems identified in issue #394: (1) for all commercial building types except 'unspecified,' the MELs technology type 'other' double-counts electricity use from all other specific MELs technologies; (2) for the 'unspecified' commercial building type, the 'unspecified' electric end use double-counts electricity use from the specific MELs technologies 'water services' and 'telecom systems;' and (3) energy use for fuels other than natural gas and distillate are not included for commercial buildings. Update stock and energy data files with revised values to incorporate these changes.

Update handling of commercial energy use data to address the three problems identified in issue #394: (1) for all commercial building types except 'unspecified,' the MELs technology type 'other' double-counts electricity use from all other specific MELs technologies; (2) for the 'unspecified' commercial building type, the 'unspecified' electric end use double-counts electricity use from the specific MELs technologies 'water services' and 'telecom systems;' and (3) energy use for fuels other than natural gas and distillate are not included for commercial buildings. Update stock and energy data files with revised values to incorporate these changes. Update cost, performance, and lifetime files to include commercial other fuel fields under the unspecified building type (though no data are available).

* Remove secondary heating and wood stoves from being subject to exogenous switching rates, as they were not intended to be covered by these rates. * Allow users to set ‘tech_switch_to’ to the value ‘same’ and have the code calculate a switchover to the same technology under exogenous switching rates. This is sometimes necessary for accounting purposes. For example, if exogenous rates are represented and a baseline segment is included that is subject to those rates, but without a complementary measure in the analysis to switch that baseline segment to, the results will show a decline in that baseline segment without a complementary increase in a switched to segment, thus under-representing that baseline segment’s influence on the total results. * Add ASHP to technology list where zero costs are expected (in EIA new construction data ASHP costs are zero for cooling, all costs are in the heating end use). * Remove water services as a distinct technology type as it appears to be counted already under “unspecified” (see issue #394). * Refactor calculation of hourly TSV load adjustment factors to ensure that each end use in the tsv_load.gz data will be leveraged (previous approach had resulted in skipping some edge case end uses in those data like clothes washing, pool pumps and pool heaters that are under ‘other’ in the Scout data). * Ensure that once cumulative measure-captured or cumulative competed stock fractions reach 1 in a given modeling year, these fractions are not updated in any subsequent years of the modeling time horizon. * Improve warning message for units conversion issues. * Add measure JSON check to ensure that measures that are set to fuel switch to electricity apply to non-electric fuels in the baseline. * Suppress check for 'other' and 'unspecified' building type input breakouts when these categories are also present in measure end uses (prioritize correct end use input breakouts, which are more commonly used).

Remove secondary heating and wood stoves from being subject to exogenous switching rates, as they were not intended to be covered by these rates. Allow users to set ‘tech_switch_to’ to the value ‘same’ and have the code calculate a switchover to the same technology under exogenous switching rates. This is sometimes necessary for accounting purposes – for example, if exogenous rates are represented and a baseline segment is included that is subject to those rates, but without a complementary measure in the analysis to switch that baseline segment to, the results will show a decline in that baseline segment without a complementary increase in a switched to segment, thus under-representing that baseline segment’s influence on the total results. Add ASHP to technology list where zero costs are expected (in EIA new construction data ASHP costs are zero for cooling, all costs are in the heating end use). Remove water services as a distinct technology type as it appears to be counted already under “unspecified” (see issue #394). Refactor calculation of hourly TSV load adjustment factors to ensure that each end use in the tsv_load.gz data will be leveraged (previous approach had resulted in skipping some edge case end uses in those data like clothes washing, pool pumps and pool heaters that are under ‘other’ in the Scout data). Ensure that once cumulative measure-captured or cumulative competed stock fractions reach 1 in a given modeling year, these fractions are not updated in any subsequent years of the modeling time horizon. Improve warning message for units conversion issues. Add measure JSON check to ensure that measures that are set to fuel switch to electricity apply to non-electric fuels in the baseline. Suppress check for 'other' and 'unspecified' building type input breakouts when these categories are also present in measure end uses (prioritize correct end use input breakouts, which are more commonly used).

* Remove secondary heating and wood stoves from being subject to exogenous switching rates, as they were not intended to be covered by these rates. * Allow users to set ‘tech_switch_to’ to the value ‘same’ and have the code calculate a switchover to the same technology under exogenous switching rates. This is sometimes necessary for accounting purposes – for example, if exogenous rates are represented and a baseline segment is included that is subject to those rates, but without a complementary measure in the analysis to switch that baseline segment to, the results will show a decline in that baseline segment without a complementary increase in a switched to segment, thus under-representing that baseline segment’s influence on the total results. * Add ASHP to technology list where zero costs are expected (in EIA new construction data ASHP costs are zero for cooling, all costs are in the heating end use.) * Remove water services as a distinct technology type as it appears to be counted already under “unspecified” (see issue #394). * Refactor calculation of hourly TSV load adjustment factors to ensure that each end use in the tsv_load.gz data will be leveraged (previous approach had resulted in skipping some edge case end uses in those data like clothes washing, pool pumps and pool heaters that are under ‘other’ in the Scout data). * Ensure that once cumulative measure-captured or cumulative competed stock fractions reach 1 in a given modeling year, these fractions are not updated in any subsequent years of the modeling time horizon. * Improve warning message for units conversion issues. * Add measure JSON check to ensure that measures that are set to fuel switch to electricity apply to non-electric fuels in the baseline. * Suppress check for 'other' and 'unspecified' building type input breakouts when these categories are also present in measure end uses (prioritize correct end use input breakouts, which are more commonly used.)

jtlangevin added this to the v1.0.0 milestone Jun 10, 2024

jtlangevin assigned trynthink Jun 10, 2024

trynthink changed the title ~~Commercial MELs -> other data are incorrectly assigned to electricity~~ Commercial electricity use is sometimes double counted, other fuel use is missing Jul 3, 2024

trynthink closed this as completed in 24ee5eb Aug 15, 2024

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Commercial electricity use is sometimes double counted, other fuel use is missing #394

Commercial electricity use is sometimes double counted, other fuel use is missing #394

jtlangevin commented Jun 10, 2024 •

edited

Loading

trynthink commented Jun 13, 2024 •

edited

Loading

Commercial electricity use is sometimes double counted, other fuel use is missing #394

Commercial electricity use is sometimes double counted, other fuel use is missing #394

Comments

jtlangevin commented Jun 10, 2024 • edited Loading

trynthink commented Jun 13, 2024 • edited Loading

jtlangevin commented Jun 10, 2024 •

edited

Loading

trynthink commented Jun 13, 2024 •

edited

Loading