23-06-2014, 09:48 AM
DEVELOPMENT OF THE GHG REFRIGERATION
AND AIR CONDITIONING MODEL
[attachment=65179]
Introduction
The objective of this project was to review and update DECC’s refrigeration and air conditioning (AC)
emissions model so that it can produce more accurate and transparent emission estimates within a
functional, flexible modelling framework that can also be used as a policy tool. More specifically, the aim of
the research was to:
1. Feed improved and more transparent GHG emissions estimates into the UK national GHG emissions
inventory report (as submitted to the UNFCCC) and improve emission projections for non-CO2 gases;
2. Allow for improved tracking of GHG emission reductions across specific refrigeration/AC sector subdivisions
(or equipment “end-uses”) to measure against national GHG reduction goals;
3. Enable testing of policy effectiveness and “what-if” scenarios; and
4. Project a consistent time series from 1990-2050.
As part of this work, ICF addressed suggestions provided by the UNFCCC on DECC’s 2010 inventory and
performed other model improvements based on the best available data and consideration of other country
inventories/best practices. These priority model improvements included: the addition of new end-uses (heat
pumps, other transport refrigeration, and other mobile AC); use of bottom-up data across all end-uses;
addition of ODS and natural refrigerants; incorporation of likely future market conditions pertaining to the
transition away from HFC refrigerants through 2050; and enhanced model functionality and transparency.
This report summarises the approach taken, detailed research findings, and associated model updates that
resulted from this effort. A comparison of the revised model output with the previous model output and UK
refrigerant production data is also provided. Finally, areas for additional research and model improvements
are identified.
Approach/Method
To update DECC’s refrigeration and AC emissions model from a largely top-down approach (based on total
refrigerant sales data) to a bottom-up approach (based on equipment stocks and average charge size from
available market data), ICF revised both the model structure and expanded and improved upon the input
assumptions for each end-use. Prior to making any updates, the previous model’s structure and
assumptions were reviewed to identify strengths, weaknesses, uncertainties, and areas for improvement.
ICF coordinated with DECC to clarify any questions and then prioritised efforts to most efficiently update the
model.
To revise the model structure, calculations were reviewed and updated to reduce redundancies and
streamline programming. The organisation of the model was also restructured to make it more user-friendly
and easier to follow. Finally, new output files were developed to enable DECC to use the model to inform
policy and decision-making. As part of this effort, ICF configured output tables to be consistent with section
2.F.1, “Refrigeration and Air-Conditioning Equipment,” of the Common Reporting Format (CRF) to facilitate
entry into the UK national GHG emissions inventory report, as submitted to the UNFCCC.
To expand and refine the end-use input assumptions, an extensive literature review was conducted and key
industry stakeholders were contacted. As a first step, literature from a broad range of sources was
consulted in June/July 2011, including those developed by government and non-government organisations,
trade associations, and other institutions. Simultaneously, ICF contacted priority industry stakeholders
(selected based on representation across all end-uses) to complement the information found in the
literature. Following the development of preliminary assumptions for all end-uses, ICF shared the draft
assumptions with a broader range of stakeholders in August/September 2011 to solicit additional industry
input. Appendix A provides a listing of the stakeholders contacted as well as a summary of those who
provided feedback. Additionally, a list of references used to form the assumptions for each end-use are
provided at the end of each subsection in Section 4.
In developing modelling input assumptions by end-use, ICF applied expert judgment to select appropriate
values when more than one estimate was provided by literature and/or stakeholders. In general, more
weight was given to estimates that are UK-specific and/or more recent. In cases of equal data quality where
numerous data points were available, values were selected based on the mid-point of the data range. Where
no UK-specific information was available, ICF relied on the 2000 Intergovernmental Panel on Climate
Change (IPCC) Good Practice Guidance default assumptions to estimate emissions. The 1996 and 2006
IPCC reports were also reviewed and considered, but the latter (most recent) assumptions could not be
adopted at this time without additional supporting information, per IPCC guidance. The table below
summarises the default assumptions from the IPCC 2000 Good Practice Guidance.
General Reorganisation and Streamlining
ICF has reorganised the model in order to facilitate continued improvement. The model now contains a
General Inputs sheet, which houses the model’s master refrigerant list, global warming potentials, and other
assumptions and inputs common to all end-uses, sheets for each end-use, and a series of calculation and
summary worksheets. Each worksheet is explained in detail below. The model also now uses a consistent
colour-coding scheme to facilitate future data entry and model updates. The colour scheme is summarised
on the introductory sheet of the model.
3.1.1 General Inputs Sheet
The General Inputs sheet houses the master list of refrigerants used throughout the tool and calculates their
global warming potentials (GWPs). The first table in the worksheet, Table A, shows the master refrigerant
list and calculates the GWP based on the refrigerant’s component chemicals. GWPs for the component
chemicals are entered into the second table on the worksheet, Table B. GWPs are entered for the IPCC’s
second, third, and fourth assessment reports, and space is also provided for users to enter an additional
GWP source.
The GWP source used throughout the model is selected from the drop-down menu in cell I11. Note that
values from the IPCC Second Assessment Report (SAR) are consistent with the values that must be used in
GHG inventories in the first commitment period of the Kyoto Protocol. Users may also enter additional
refrigerants at the bottom of the chemical list and, as needed, can enter information on any components of
that refrigerant in the empty columns at the right of Table A and GWP information for new components in the
empty rows at the bottom of Table B.
The General Inputs sheet also houses general unit conversion factors used throughout the model, as well as
assumptions that apply to all retrofits (i.e., replacement of equipment charge with a new refrigerant type).
3.1.2 Summary by End-Use
ICF reorganised the tool so that each end-use is represented on its own Excel worksheet (in the previous
system, each gas had its own worksheet). This allows users to easily review all the key assumptions for a
given end-use at the same time. Each end-use sheet is split into four parts: (1) data inputs used by
calculations; (2) disaggregated annual consumption by specific chemical; (3) retrofit assumptions; and (4)
market data specific to each end-use used to build up calculations. All parts are structured similarly for all
end-uses, and each of these parts is described in detail below.
Part 1 of the sheet contains the data inputs that are used by the calculations (see Figure 2). These include
the end-use name; lifetime; annual refill status; average charge size; and emission rates for manufacture,
operations, and disposal. The “Retrofit table start cell:” communicates the presence of retrofits to the model
and is not an entry, while “Post-2010 Leak Reductions” controls whether or not reductions in annual leakage
are assumed to occur in existing equipment beginning in 2010 as a result of the leak checking/repair
provisions specified under Regulation (EC) No. 842/2006 on certain fluorinated greenhouse gases (the Fgas
Regulation). Input cells are shaded in yellow for easy identification. Emissions rates are expressed as
fractions.
Blends
The original model listed each chemical formulation separately for each end-use, listing neat compositions
and blends. ICF has revised the model such that each end-use sheet only shows the blends and chemicals
in the forms that are actually used in the equipment. The component chemicals are disaggregated later in
the model flow as necessary to generate inventory outputs by chemical. This allows users to focus only on
the products used in an end-use when focusing on that particular end-use.
3.1.4 Future HFOs
HFOs (hydrofluoro-olefins), or unsaturated HFCs, are now beginning to be used as refrigerants in the UK
and will be used increasingly in future, as additional HFOs and HFO blends are developed and brought to
market. The specific HFOs/HFO blends that will successfully enter the various market segments will depend
on safety,4 efficiency, and cost. Given the current uncertainty regarding the future HFOs/HFO blends that will
be used across the various end-uses, three broad HFO “Types” have been entered into the model defined
by GWP, based on product information provided by Honeywell (2011).5 These HFO types are summarised
below in Table 4. It should be noted that the GWP assumed for the HFO already in use in light duty MACs
Adjusting Model Outputs
The model is capable of generating outputs for every row of the Calculations table (from the second half of
the Calculations sheet). Many of these rows represent intermediate calculation steps and have not been
configured to load into the model output at this time, as the model takes longer to run the more variables are
processed. However, if the user wants to generate additional outputs from the calculations table, they can
follow the following steps. Note that this involves opening the Macro that runs the model in Visual Basic
Editor. It is recommended that users have some familiarity with the Visual Basic Editor before making such
edits, as mistakes during this step can cause the model to run incorrectly.
Operational Loss Rate
Oko-Recherche et al. (2011) assume a leak rate of 1%. IPCC (2006) estimates a default leak rate of 1-15%
while IPCC (2000) estimates a default leak rate of 1-10%. IPCC (1996) does not specifically identify an
annual leak rate for stand-alone refrigeration units but provides an operational leak rate for the broad
category of “other stationary refrigeration and air conditioning equipment” of 3-17%. Recognising that some
leakage will occur, albeit very small, and that a small percentage of systems will suffer catastrophic damage,
ICF increased the previous assumption of 1% to 1.5% in 2010, decreasing it to 1% by 2020. ICF assumes
higher loss rates in earlier years—i.e., 3% in 1990 and 2% in 2000—based on the assumption that leak
tightness technology has improved over time. Loss rates towards the lower end of the IPCC (2000) range
have been selected, since this is more in line with the EU specific estimates from Oko-Recherche et al.
(2011). For this end-use, ICF also assumes that no servicing occurs; i.e., the amount of refrigerant in each
unit decreases over time due to annual leaks.
Uncertainty Analysis
The Tier 2 bottom-up analytical modelling approach used to estimate emissions from refrigeration/air
conditioning equipment is IPCC compliant. Although the DECC model is more comprehensive than the IPCC
default methodology, significant uncertainties still exist with regard to the levels of equipment sales,
equipment characteristics, and end-use emissions profiles that were used to estimate annual emissions for
the various compounds.
In order to calculate uncertainty, functional forms were developed to simplify some of the complex aspects of
the refrigeration and air-conditioning sector. In particular, because emissions are calculated based on the
entire lifetime of equipment, not just equipment put into commission in the current year, simplifying
equations were used. The functional forms used variables that included growth rates, lifetimes, emission
factors (manufacturing, operational, and disposal emission rates), refrigerant transitions, charge size,
disposal quantities, and new and existing stock. Uncertainty was estimated around each variable within the
functional forms based on ICF’s expert judgment, taking into account the range of estimates provided in the
literature and by industry stakeholders. A Monte Carlo simulation analysis was performed and uncertainty
bounds were generated using 10,000 simulations. The most significant sources of uncertainty for this source
category include the emission factors for centralised supermarket refrigeration systems and marine transport
refrigeration—two end uses with significant installed base (due to large stock and/or charge size).
AND AIR CONDITIONING MODEL
[attachment=65179]
Introduction
The objective of this project was to review and update DECC’s refrigeration and air conditioning (AC)
emissions model so that it can produce more accurate and transparent emission estimates within a
functional, flexible modelling framework that can also be used as a policy tool. More specifically, the aim of
the research was to:
1. Feed improved and more transparent GHG emissions estimates into the UK national GHG emissions
inventory report (as submitted to the UNFCCC) and improve emission projections for non-CO2 gases;
2. Allow for improved tracking of GHG emission reductions across specific refrigeration/AC sector subdivisions
(or equipment “end-uses”) to measure against national GHG reduction goals;
3. Enable testing of policy effectiveness and “what-if” scenarios; and
4. Project a consistent time series from 1990-2050.
As part of this work, ICF addressed suggestions provided by the UNFCCC on DECC’s 2010 inventory and
performed other model improvements based on the best available data and consideration of other country
inventories/best practices. These priority model improvements included: the addition of new end-uses (heat
pumps, other transport refrigeration, and other mobile AC); use of bottom-up data across all end-uses;
addition of ODS and natural refrigerants; incorporation of likely future market conditions pertaining to the
transition away from HFC refrigerants through 2050; and enhanced model functionality and transparency.
This report summarises the approach taken, detailed research findings, and associated model updates that
resulted from this effort. A comparison of the revised model output with the previous model output and UK
refrigerant production data is also provided. Finally, areas for additional research and model improvements
are identified.
Approach/Method
To update DECC’s refrigeration and AC emissions model from a largely top-down approach (based on total
refrigerant sales data) to a bottom-up approach (based on equipment stocks and average charge size from
available market data), ICF revised both the model structure and expanded and improved upon the input
assumptions for each end-use. Prior to making any updates, the previous model’s structure and
assumptions were reviewed to identify strengths, weaknesses, uncertainties, and areas for improvement.
ICF coordinated with DECC to clarify any questions and then prioritised efforts to most efficiently update the
model.
To revise the model structure, calculations were reviewed and updated to reduce redundancies and
streamline programming. The organisation of the model was also restructured to make it more user-friendly
and easier to follow. Finally, new output files were developed to enable DECC to use the model to inform
policy and decision-making. As part of this effort, ICF configured output tables to be consistent with section
2.F.1, “Refrigeration and Air-Conditioning Equipment,” of the Common Reporting Format (CRF) to facilitate
entry into the UK national GHG emissions inventory report, as submitted to the UNFCCC.
To expand and refine the end-use input assumptions, an extensive literature review was conducted and key
industry stakeholders were contacted. As a first step, literature from a broad range of sources was
consulted in June/July 2011, including those developed by government and non-government organisations,
trade associations, and other institutions. Simultaneously, ICF contacted priority industry stakeholders
(selected based on representation across all end-uses) to complement the information found in the
literature. Following the development of preliminary assumptions for all end-uses, ICF shared the draft
assumptions with a broader range of stakeholders in August/September 2011 to solicit additional industry
input. Appendix A provides a listing of the stakeholders contacted as well as a summary of those who
provided feedback. Additionally, a list of references used to form the assumptions for each end-use are
provided at the end of each subsection in Section 4.
In developing modelling input assumptions by end-use, ICF applied expert judgment to select appropriate
values when more than one estimate was provided by literature and/or stakeholders. In general, more
weight was given to estimates that are UK-specific and/or more recent. In cases of equal data quality where
numerous data points were available, values were selected based on the mid-point of the data range. Where
no UK-specific information was available, ICF relied on the 2000 Intergovernmental Panel on Climate
Change (IPCC) Good Practice Guidance default assumptions to estimate emissions. The 1996 and 2006
IPCC reports were also reviewed and considered, but the latter (most recent) assumptions could not be
adopted at this time without additional supporting information, per IPCC guidance. The table below
summarises the default assumptions from the IPCC 2000 Good Practice Guidance.
General Reorganisation and Streamlining
ICF has reorganised the model in order to facilitate continued improvement. The model now contains a
General Inputs sheet, which houses the model’s master refrigerant list, global warming potentials, and other
assumptions and inputs common to all end-uses, sheets for each end-use, and a series of calculation and
summary worksheets. Each worksheet is explained in detail below. The model also now uses a consistent
colour-coding scheme to facilitate future data entry and model updates. The colour scheme is summarised
on the introductory sheet of the model.
3.1.1 General Inputs Sheet
The General Inputs sheet houses the master list of refrigerants used throughout the tool and calculates their
global warming potentials (GWPs). The first table in the worksheet, Table A, shows the master refrigerant
list and calculates the GWP based on the refrigerant’s component chemicals. GWPs for the component
chemicals are entered into the second table on the worksheet, Table B. GWPs are entered for the IPCC’s
second, third, and fourth assessment reports, and space is also provided for users to enter an additional
GWP source.
The GWP source used throughout the model is selected from the drop-down menu in cell I11. Note that
values from the IPCC Second Assessment Report (SAR) are consistent with the values that must be used in
GHG inventories in the first commitment period of the Kyoto Protocol. Users may also enter additional
refrigerants at the bottom of the chemical list and, as needed, can enter information on any components of
that refrigerant in the empty columns at the right of Table A and GWP information for new components in the
empty rows at the bottom of Table B.
The General Inputs sheet also houses general unit conversion factors used throughout the model, as well as
assumptions that apply to all retrofits (i.e., replacement of equipment charge with a new refrigerant type).
3.1.2 Summary by End-Use
ICF reorganised the tool so that each end-use is represented on its own Excel worksheet (in the previous
system, each gas had its own worksheet). This allows users to easily review all the key assumptions for a
given end-use at the same time. Each end-use sheet is split into four parts: (1) data inputs used by
calculations; (2) disaggregated annual consumption by specific chemical; (3) retrofit assumptions; and (4)
market data specific to each end-use used to build up calculations. All parts are structured similarly for all
end-uses, and each of these parts is described in detail below.
Part 1 of the sheet contains the data inputs that are used by the calculations (see Figure 2). These include
the end-use name; lifetime; annual refill status; average charge size; and emission rates for manufacture,
operations, and disposal. The “Retrofit table start cell:” communicates the presence of retrofits to the model
and is not an entry, while “Post-2010 Leak Reductions” controls whether or not reductions in annual leakage
are assumed to occur in existing equipment beginning in 2010 as a result of the leak checking/repair
provisions specified under Regulation (EC) No. 842/2006 on certain fluorinated greenhouse gases (the Fgas
Regulation). Input cells are shaded in yellow for easy identification. Emissions rates are expressed as
fractions.
Blends
The original model listed each chemical formulation separately for each end-use, listing neat compositions
and blends. ICF has revised the model such that each end-use sheet only shows the blends and chemicals
in the forms that are actually used in the equipment. The component chemicals are disaggregated later in
the model flow as necessary to generate inventory outputs by chemical. This allows users to focus only on
the products used in an end-use when focusing on that particular end-use.
3.1.4 Future HFOs
HFOs (hydrofluoro-olefins), or unsaturated HFCs, are now beginning to be used as refrigerants in the UK
and will be used increasingly in future, as additional HFOs and HFO blends are developed and brought to
market. The specific HFOs/HFO blends that will successfully enter the various market segments will depend
on safety,4 efficiency, and cost. Given the current uncertainty regarding the future HFOs/HFO blends that will
be used across the various end-uses, three broad HFO “Types” have been entered into the model defined
by GWP, based on product information provided by Honeywell (2011).5 These HFO types are summarised
below in Table 4. It should be noted that the GWP assumed for the HFO already in use in light duty MACs
Adjusting Model Outputs
The model is capable of generating outputs for every row of the Calculations table (from the second half of
the Calculations sheet). Many of these rows represent intermediate calculation steps and have not been
configured to load into the model output at this time, as the model takes longer to run the more variables are
processed. However, if the user wants to generate additional outputs from the calculations table, they can
follow the following steps. Note that this involves opening the Macro that runs the model in Visual Basic
Editor. It is recommended that users have some familiarity with the Visual Basic Editor before making such
edits, as mistakes during this step can cause the model to run incorrectly.
Operational Loss Rate
Oko-Recherche et al. (2011) assume a leak rate of 1%. IPCC (2006) estimates a default leak rate of 1-15%
while IPCC (2000) estimates a default leak rate of 1-10%. IPCC (1996) does not specifically identify an
annual leak rate for stand-alone refrigeration units but provides an operational leak rate for the broad
category of “other stationary refrigeration and air conditioning equipment” of 3-17%. Recognising that some
leakage will occur, albeit very small, and that a small percentage of systems will suffer catastrophic damage,
ICF increased the previous assumption of 1% to 1.5% in 2010, decreasing it to 1% by 2020. ICF assumes
higher loss rates in earlier years—i.e., 3% in 1990 and 2% in 2000—based on the assumption that leak
tightness technology has improved over time. Loss rates towards the lower end of the IPCC (2000) range
have been selected, since this is more in line with the EU specific estimates from Oko-Recherche et al.
(2011). For this end-use, ICF also assumes that no servicing occurs; i.e., the amount of refrigerant in each
unit decreases over time due to annual leaks.
Uncertainty Analysis
The Tier 2 bottom-up analytical modelling approach used to estimate emissions from refrigeration/air
conditioning equipment is IPCC compliant. Although the DECC model is more comprehensive than the IPCC
default methodology, significant uncertainties still exist with regard to the levels of equipment sales,
equipment characteristics, and end-use emissions profiles that were used to estimate annual emissions for
the various compounds.
In order to calculate uncertainty, functional forms were developed to simplify some of the complex aspects of
the refrigeration and air-conditioning sector. In particular, because emissions are calculated based on the
entire lifetime of equipment, not just equipment put into commission in the current year, simplifying
equations were used. The functional forms used variables that included growth rates, lifetimes, emission
factors (manufacturing, operational, and disposal emission rates), refrigerant transitions, charge size,
disposal quantities, and new and existing stock. Uncertainty was estimated around each variable within the
functional forms based on ICF’s expert judgment, taking into account the range of estimates provided in the
literature and by industry stakeholders. A Monte Carlo simulation analysis was performed and uncertainty
bounds were generated using 10,000 simulations. The most significant sources of uncertainty for this source
category include the emission factors for centralised supermarket refrigeration systems and marine transport
refrigeration—two end uses with significant installed base (due to large stock and/or charge size).