SRI SMDI Auto CLCA Study Final Report

TABLE OF CONTENTS
EXECUTIVE SUMMARY ii
1. BACKGROUND 1
2. GOAL OF THE STUDY 4
3. SCOPE OF THE STUDY 5

3.1 Product System Description 5
3.2 Functional Unit 5
3.3 System Boundary 5
3.4 Cut-off Criteria 6
4. LIFE CYCLE INVENTORY ANALYSIS 7

4.1 Data Collection and Sources 7
4.1.1 Material Inputs 7
4.1.2 Material Replacement Coefficients 9
4.1.3 Secondary Mass Savings 11
4.1.4 Lifetime Driving Distance 11
4.1.5 Fuel Reduction Values and Power Train Resizing 12
4.1.6 Recycling Rates 12
4.2 CLCA Model Calculation Procedures 13
4.3 Allocation 13
4.4 Data Validation and Data Quality Assessment 13
4.4.1 Time-related Coverage 13
4.4.2 Geographical Coverage 14
4.4.3 Technology Coverage 14
4.4.4 Precision 14
4.4.5 Completeness 14
4.4.6 Representativeness 15
4.4.7 Consistency 15
4.4.8 Reproducibility 15
4.4.9 Sources of Data 15
4.4.10 Uncertainty 15
4.5 Sensitivity Analysis 15
5. LIFE CYCLE IMPACT ASSESSMENT (LCIA) RESULTS 17

5.1 Base Case Results 17
5.2 Sensitivity Analysis Results 22
6. INTERPRETATION 29
6.1 Key Findings 29
6.2 Assumptions and Limitations 29
6.3 Conclusions 31
APPENDIX A – Consequential Life Cycle Assessment Model User Guide and

Project Report
APPENDIX B – Critical Review for this Study
i|Page
EXECUTIVE SUMMARY
Automakers in the U.S. have used vehicle weight reduction (“lightweighting”) as one of many
strategies to comply with mandated federal fuel efficiency targets. This strategy often
involves substituting one material for another; for instance, replacing mild steel with
aluminum or replacing mild steel with advanced high strength steel (AHSS). This
lightweighting process typically results in improved fuel efficiency and a corresponding
reduction in greenhouse gas (GHG) emissions from the vehicle’s use phase. However, such
lightweighting doesn’t necessarily result in an overall GHG savings, since the production
phase emissions for some lightweighting materials can counteract the improvement in the
tailpipe or use phase.
Aluminum is one of the materials that is often considered for lightweighting applications, and
in North America, the Aluminum Association has projected a significant increase in the use of
aluminum in vehicle body and closure panels. 1 The Steel Recycling Institute (SRI) and Steel
Market Development Institute (SMDI), both business units of the American Iron and Steel
Institute (AISI), conducted this study to assess the GHG emissions consequences of an
increase in the use of aluminum equal to that projected by the Aluminum Association, and
alternately, to assess the GHG emissions consequences of using AHSS instead of aluminum
to lightweight the same vehicle fleet.
These consequences were assessed using a spreadsheet model (hereinafter referred to as

the Model) developed by Dr. Roland Geyer, PhD, Associate Professor, Bren School of
Environmental Science and Management, University of California at Santa Barbara. The
report describing the model is titled “Consequential Life Cycle Assessment (CLCA) of
Replacing Steel with Aluminum in Vehicles: User Guide and Final Project Report”, dated
February 26, 2016. The Model and referenced report were independently reviewed by a
three-person panel. (See Appendix A for a copy of the report and review letter.)
The CLCA was conducted according to the requirements of ISO 14044:2006. Strictly
speaking, this study is a consequential life cycle greenhouse gas emissions (assessed as
global warming potentials, or GWP-100) assessment. However, for simplicity the acronym
“CLCA” is used in this report. It is also important to keep in mind that ISO 14044 was written
with attributional life cycle assessment (ALCA) in mind and therefore some interpretation
may be necessary when it is applied to CLCA.
The goal of this study is to identify and quantify the main GHG emissions consequences of a
significant increase in the use of aluminum for vehicle body and closure parts, and compare
these consequences to the GHG emissions consequences of the use of AHSS for the same
parts, as part of an overall vehicle fleet lightweighting strategy. The study considers light-
duty vehicles produced in North America between 2015 and 2053.
Four sets of consequences are modeled in the study:

1) Changes in the production levels of the steel and aluminum used in the modeled
body and closure parts, as well as secondary mass savings associated with these
production level changes;
2) Changes in the fuel economy of the vehicles due to mass reduction;
1Ducker Worldwide. “Aluminum Content in North American Light Vehicles 2016 to 2028.” Summary Report prepared for the
Aluminum Association. July 2017. http://www.drivealuminum.org/research-resources/ducker2017/.
ii | P a g e
3) Changes in the generation and use of steel and aluminum scrap from material
forming processes; and,
4) Changes in the generation and use of steel and aluminum scrap from vehicle end-of-
life management.
In order to develop the data required for the intended comparison of an increase in the use
of AHSS vs. aluminum, the Model’s calculations are run separately for two different sets of
assumptions. All basic input data is the same for both Model runs, except for the production
levels of AHSS or aluminum.
The baseline results of the assessment can be seen in the following graph:
The graph shows the GHG consequences of the baseline comparison. In short, the projected
increase in the use of aluminum for lightweighting of body and closure parts results in peak
cumulative GHG emissions of approximately 209 million metric tons, while a similar increase
in the use of AHSS results in an immediate and sustained decrease in overall GHG
emissions. In this baseline case, at the time of the peak increase in emissions for the
aluminum option (approximately Year 2038), the difference in GHG emissions between the
two options is approximately 332 million metric tons. This net difference continues to grow
throughout the study period until it reaches over 411 million metric tons in 2053. In other
words, a significant increase in the use of aluminum for lightweighting of light-duty vehicles,
as described in this report, results in an increase in overall GHG emissions of over 330
million tons in approximately 20 years time, when compared to similar lightweighting of the
same vehicle fleet with AHSS.
These results are based on baseline data and inputs. This study includes numerous
sensitivity scenarios which are described in Section 4.5.
iii | P a g e
1. BACKGROUND
Corporate Average Fuel Economy (CAFE) standards for U.S. vehicles were first established by
Congress in 1975 to reduce fuel consumption by cars and light trucks. These standards are
administered by the Department of Transportation’s National Highway Traffic Safety
Administration (NHTSA). NHTSA sets and enforces the standards, while the U.S.
Environmental Protection Agency (U.S. EPA) calculates average fuel economy levels for
vehicle manufacturers, and also sets related standards for greenhouse gas (GHG) emissions
associated with fuel combustion. The effect of these standards on average vehicle carbon
dioxide (CO2) emissions and fuel economy in the U.S. is shown in Figure 1. From model year
(MY) 2004 to MY 2016, CO2 emissions have decreased by 102 g/mi, or 22%, and fuel
economy has increased by 5.4 mpg, or 28%. 2
Figure 1. Adjusted* Vehicle Carbon Dioxide (CO2) Emissions and Fuel Economy Trends
Source: U.S. EPA, 2018. 3

* Adjusted CO2 and fuel economy values reflect real world performance and are not comparable to automaker standards
compliance levels. Adjusted CO2 values are, on average about 25% higher than the unadjusted, laboratory CO2 values that form
the starting point for GHG standards compliance, and adjusted fuel economy values are about 20% lower, on average, than
unadjusted fuel economy values that form the starting point for CAFÉ standards compliance.
These more recent improvements were achieved primarily through the development and
implementation of new engine and transmission technologies, such as gasoline direct
injection, turbocharging, continuously variable transmissions, and non-hybrid start/stop
technology. 4 Regulatory targets are currently set to become even more stringent through MY
2025, causing automakers to look to additional solutions, such as reducing the mass of
vehicles to decrease fuel consumption, commonly referred to as “lightweighting.” On April 2,
2018, the U.S. EPA’s Administrator announced the completion of the Midterm Evaluation
2 U.S. EPA, Light-Duty Vehicle CO2 and Fuel Economy Trends website https://www.epa.gov/fuel-economy-trends/highlights-co2-
and-fuel-economy-trends. Last accessed June 1, 2018.
3 Ibid.
4 U.S. EPA, “Light-Duty Automotive Technology, Carbon Dioxide Emissions, and Fuel Economy Trends: 1975 Through 2016”
Report. November 2016. pg. 47.

1|Page
process for cars and light trucks for model years 2022-2025, noting that the current
standards should be revised. 5 It is currently unclear what this revision will entail.
A common lightweighting strategy is designing vehicles to use materials that enable the
weight reduction of vehicle components and systems, while still maintaining functionality and
performance. These materials may include advanced high-strength steels (AHSS), aluminum,
carbon fiber composites, and magnesium. Each of these materials can contribute to vehicle
mass reduction, thereby helping to improve fuel economy; however, each does so at different
manufacturing cost levels and environmental impacts.
While the focus of federal regulations has been on the vehicle use phase (i.e., fuel economy),
the complete GHG and energy profile of a vehicle is only evident by considering the entire life
cycle. A vehicle’s life cycle can be subdivided into three distinct parts (or phases): production
(materials, auto parts, and vehicle assembly), use (driving and maintenance) and end-of-life
(recycling and/or disposal). Because the use of lightweight materials may increase vehicle
production emissions while improving use phase performance, a life cycle approach is
necessary to assess trade-offs and ensure improvements in one phase are not causing larger
impacts in another.
To evaluate these trade-offs, the Steel Recycling Institute (SRI) and Steel Market
Development Institute (SMDI), both business units of the American Iron and Steel Institute
(AISI), have taken a two-part approach. In November 2017, SRI and SMDI released an ISO
critically reviewed study 6 of the greenhouse gas (GHG) emissions and energy consumption
implications associated with lightweighting the body and closures of various MY 2016 vehicle
types with aluminum and AHSS. This more common “attributional” life cycle study found
that, even though the use phase currently dominates the life cycle results for most vehicle
types, AHSS-intensive vehicles had lower or equivalent life cycle emissions than aluminum-
intensive vehicles for every class of vehicles tested. The study demonstrated that a focus
only on tailpipe emissions is likely to produce unintended consequences of higher total GHG
emissions. Previously published studies, notably Das et al. 2014 7 and Bushi et al. 8, as well
as the various automotive lightweighting life cycle assessment (LCA) studies assessed by
Hottle et al. 2017 9, found the use of aluminum to have life cycle GHG advantages over steel
and even AHSS, largely due to the use phase fuel consumption savings of aluminum over
steel. The 2017 SRI/SMDI study reached a different conclusion largely due to differences in
the approach to recycling allocation methods (holistic assessment of possible methods vs.
the avoided burden end-of-life approach), accounting for imported primary aluminum, and
recycled content assumptions and was subjected to a rigorous ISO 14044 review by a panel
of four experts.
Since individual vehicle-to-vehicle comparisons, such as those in attributional life cycle
assessments (ALCAs), may not fully assess the GHG emissions implications of a large-scale
shift in the material composition of vehicles produced in North America, SRI and SMDI have
also commissioned an alternate approach to assessing these larger scale consequences,
termed “consequential” life cycle assessment (CLCA), which is the focus of this study.
5 U.S. EPA, News Release, “EPA Administrator Pruitt: GHG Emissions Standards for Cars and Light Trucks Should Be Revised.”
April 2, 2018. https://www.epa.gov/newsreleases/epa-administrator-pruitt-ghg-emissions-standards-cars-and-light-trucks-
should-be.
6 Steel Recycling Institute and Steel Market Development Institute. “Life Cycle Greenhouse Gas and Energy Study of Automotive
Lightweighting.” November 7, 2017. http://www.steelsustainability.org/automotive/auto-ghg/.

7 Das, S., "Life Cycle Energy and Environmental Assessment of Aluminum-Intensive Vehicle Design," SAE Int. J. Mater. Manf.
7(3):588-595, 2014, https://doi.org/10.4271/2014-01-1004.

8 Bushi, L., Skszek, T., and Wagner, D., "MMLV: Life Cycle Assessment," SAE Technical Paper 2015-01-1616, 2015,
doi:10.4271/2015-01-1616.
9 Hottle, T., Caffrey, C., McDonald, J., and Dodder, R. “Critical factors affecting life cycle assessments of material choice for
vehicle mass reduction,” Transportation Research Part D 56 (2017) 241–257. http://dx.doi.org/10.1016/j.trd.2017.08.010.

2|Page
In the field of life cycle assessment (LCA), consequential LCA is a relatively new and emerging
assessment method. For ALCA, a static or fixed inventory of inputs and outputs is
determined for all processes in the life cycle of a product or service and scaled linearly to a
functional unit. 10 ALCA typically utilizes global or national averages to model the involved
unit processes. For CLCA, marginal or incremental inventories are used to assess the
consequences of a change to the life cycle under study. In this study, CLCA is applied as an
assessment of the physical and economic processes that are affected by a well-defined
change to the studied product system, causing it to evolve over time from its initial state.
Key consequential parameters are identified and defined and the sensitivity of the results to
their variation is modeled.
10Koffler, C.; Geyer, R.; Volz, T. Life Cycle Inventory. In Environmental Life Cycle Assessment: Measuring the environmental
performance of products; American Center for Life Cycle Assessment: Vashon Island, Washington, 2014; pp 46–57.
3|Page
2. GOAL OF THE STUDY
SRI and SMDI conducted this study to assess the life cycle GHG emissions, assessed as 100-
year global warming potentials (GWP-100), resulting from a projected large-scale shift of
closure and body parts from primarily mild steel to aluminum in North American production
of light duty vehicles, and compare these GHG implications, or consequences, to a similar
shift from primarily mild steel to AHSS.
The intended application of the study is to develop comparative claims and assess trade-offs
across different life cycle stages on the basis of life cycle GWP-100 (hereafter referred to as
“GHG emissions” for convenience) for communication with internal and external
stakeholders. As described in Section 5, the focus on a single impact is a limitation of the
study. However, GHG emissions that cause climate change impacts are of high public and
institutional interest, and is among the currently most pressing environmental issues relative
to the production and operation of vehicles. Since the study assesses a single indicator, it is
not intended to support “comparative assertions” as defined by ISO 14040, Section 3.6, 11 or
claims of overall environmental superiority. Nevertheless, since it does compare the GHG
emissions implications of two options, the study was subjected to a critical review by a panel
of independent experts to demonstrate its conformance with ISO 14044 (see Appendix B).
The target audience includes all internal and external stakeholders with an interest in
understanding the GHG implications of large-scale automotive material substitution for
reasons of vehicle mass reduction. Due to its sole focus on climate change impacts, the
study does not investigate any potential trade-offs across different impact categories. Strictly
speaking, this study is a consequential life cycle GWP-100 assessment; however, for
simplicity the acronym CLCA will be used in the remainder of this report.
The CLCA model is highly parameterized and thus it is possible to conduct extensive
sensitivity analyses by varying key input parameters. Section 4.5 of the report includes
several of these sensitivity analyses.
This study has been conducted according to the requirements of the international standard
ISO 14044:2006 12. It is important to note that ISO 14044 was written with ALCA in mind,
and therefore, some interpretation of these requirements is necessary when applied to CLCA.
11 ISO 14040: Environmental management – Life cycle assessment – Principles and framework. Geneva: International
Organization for Standardization. 2006.
12 ISO 14044: Environmental management – Life cycle assessment – Requirements and guidelines. Geneva: International
Organization for Standardization. 2006.

4|Page
3. SCOPE OF THE STUDY
The following sections describe the project scope as defined to achieve the stated goals. This
scope is reflected in the consequential life cycle assessment model (the Model) developed by
Dr. Roland Geyer, PhD, Associate Professor, Bren School of Environmental Science and
Management, University of California at Santa Barbara. The computational structure and
methodology of the Model is described in the report titled “Consequential Life Cycle
Assessment (CLCA) of Replacing Steel with Aluminum in Vehicles: User Guide and Final
Project Report”, dated February 26, 2016. The Model and referenced report were
independently reviewed by a three-person panel. (See Appendix A for a copy of the report
and review letter.) The independent review concludes that the model “… conforms to the
LCA requirements of ISO 14040:2006 and 14044: 2006 standards.”
3.1 Product System Description
The product system in this study is the fleet of North American light duty vehicles that would
be lightweighted with aluminum (or alternatively, AHSS) to meet the aluminum industry’s
projected increase in aluminum use in body and closure applications in North American
vehicles. The projected increase in the use of aluminum is based on a publicly-available
study conducted by Ducker Worldwide and commissioned by the Aluminum Association. 13
For the AHSS scenario, the projected increase in the use of AHSS is based on lightweighting
the same body and closure parts in the vehicle fleet modeled in the aluminum case. Since
the weight reduction potential (i.e., Material Replacement Coefficient; see Section 4.1.2) for
AHSS to mild steel is lower than for aluminum to mild steel, the overall weight of the vehicle
fleet is higher and the fuel savings lower for the AHSS scenario versus aluminum.
3.2 Functional Unit
The functional unit of the study consists of the transportation services (total vehicle miles
traveled) provided between 2015 and 2053 by all light duty vehicles produced in North
America between 2015 and 2053. Only vehicles produced between 2015 and 2053 are
considered, and vehicle use or end-of-life (EOL) management after 2053 is excluded. This
definition of the functional unit means that not all included vehicles reach the end of their
lives within the study period. However, all vehicles lightweighted during the aluminum
industry’s projected growth curve 14 are modeled to reach their EOL and are recycled during
the study timeframe. The industry projection ends in 2028, so the aluminum content and
production volumes are held constant from 2028 to 2053. This assumption creates a
steady state annual emission profile by the end of the modelling period and improves the
interpretation of the results (see Section 3.3 for additional details).
An alternative functional unit would be to stop considering vehicle production at a given year,
but include vehicle use until all of the included vehicles reach end of life. However, this
creates annual GHG emissions that are somewhat artificial and difficult to interpret, since
vehicle production, use, and end-of-life goes on concurrently. Furthermore, the amount of
aluminum use in vehicles, once implemented, is relatively fixed without significant vehicle
redesign. Therefore, the functional unit definition described above was employed.
3.3 System Boundary
13 Ducker Worldwide. “Aluminum Content in North American Light Vehicles 2016 to 2028.” Summary Report prepared for the
Aluminum Association. July 2017. http://www.drivealuminum.org/research-resources/ducker2017/.
14 Ibid.
5|Page
The system boundary has been defined to capture the main GHG emissions changes caused
by the material substitution described in the reference flow of the study, where the reference
flow consists of two separate parts. The first part is a time series of annual North American
light vehicle production in vehicles per year from 2015 to 2053. For each year the total
production number is broken into eight different vehicle classes, including A/B, C, D, E, MPV
(passenger vans), SUV (sport-utility vehicles), VAN (utility vans), and PUP (pick-up trucks), and
five different power train types, including gasoline- and diesel-powered internal combustion
vehicles (ICVs), standard and plug-in hybrid electric vehicles (HEVs), and pure battery electric
vehicles (BEVs). Gasoline vehicles comprise the majority of the vehicle fleet, decreasing
slightly from 90.3% to 81.3% during the study period. All other power train types make up the
rest of the share, with the hybrid and electric power train vehicles growing modestly during
the study period. The fleet model calculations and assumptions are detailed in the CLCA
Model User Guide and Project Report found in Appendix A.
The second part of the reference flow is a time series of the annual amount of additional
aluminum body and closure parts, or the additional amount of AHSS body and closure parts,
used instead of mild steel body and closure parts in North American light vehicle production.
The total amount is further converted into average amount per vehicle, as a function of
power train type and vehicle class composition of each power train type.
Note that the input data in the aluminum case uses forecasts from a Ducker Worldwide study
(as previously described) for the years 2015 to 2028 and aluminum production levels are
then held constant between 2029 and 2053. The AHSS case is handled similarly. Keeping
production of aluminum or AHSS level for another 25 years beyond 2028 results in several
advantages from a modeling standpoint. First, it is a conservative modeling approach that
allows for complete consideration of the use phase and EOL recycling savings associated
with vehicles lightweighted during the industry projection timeframe. Second, it enables the
model to calculate a cross-over time, i.e. the time at which use phase savings have
completely made up for the GHG increase during vehicle production, and cumulative GHG
emissions have reached zero.
The aim of this CLCA study is to account for all significant changes caused by the reference
flow rather than to account for every significant part of a product or service life cycle, which is
the typical objective of an ALCA. Four distinct consequences (changes) within and outside of
the vehicle life cycles are modelled: 1) increases and/or decreases in the production of steel
and aluminum used in the modeled body and closure parts as well as secondary mass
savings, 2) fuel economy of the mass-reduced vehicles, 3) the generation and use of steel
and aluminum scrap from material forming processes (prompt or manufacturing scrap), and
4) the generation and use of steel and aluminum scrap from vehicle end-of-life management.
All life cycle stages or processes that do not experience significant GHG emission changes
due to the material substitution are omitted. Examples are the production of tires, vehicle
fluids, non-structural materials, as well as vehicle assembly.
3.4 Cut-off Criteria
The cut-off criterion of this study is to capture at least 95% of all GHG emission changes as
described in Section 3.3 and caused by the studied increase in aluminum production for use
in North American vehicles as a replacement of equivalent primarily mild steel closures and
body parts. For the AHSS scenario, the studied replacement is AHSS closures and body parts
for the same amount of primarily mild steel replaced in the aluminum option. In this context,
“equivalent” is defined as meaning that the material substitution does not affect the
technical specifications and safety rating of the vehicles.
6|Page
4. LIFE CYCLE INVENTORY ANALYSIS
The scope of the study results in the following set of unit processes:
• Cradle-to-gate production of primary aluminum ingots and integrated (blast furnace /
basic oxygen furnace, or BF/BOF) steel slabs used in vehicle production or displaced
due to production and end-of-life scrap recycling;
• Scrap-to-gate production of secondary aluminum ingots and electric arc furnace (EAF)
steel slabs used in vehicle production or from automotive production and end-of-life
scrap;
• Gate-to-gate aluminum rolling, extrusion, and casting and steel rolling and casting;
• Cradle-to-gate gasoline, diesel, and electricity production; and,
• In-vehicle gasoline and diesel combustion.
4.1 Data Collection and Sources
For many input parameters, initial values are sourced from the SRI/SMDI ALCA study titled
“Life Cycle Greenhouse Gas and Energy Study of Automotive Lightweighting” dated
November 7, 2017. These key input values and their justification are described below in
Sections 4.1.1 to 4.1.6. Several of these inputs vary over the study time period and are
indicated as such. All other inputs are described in the CLCA Model User Guide and Project
Report (see Appendix A). The SRI/SMDI 2017 ALCA study was subjected to a comprehensive
independent peer review by a panel of four subject matter experts. The review concluded in
part: “After an exhaustive three rounds of review of comments and responses by the panel
members and the AISI, based on the goals set forth to review this study, the review panel
concludes that the study conforms to ISO 14044:2006 as a comprehensive study that may
be disclosed to the public.”
Input data values for this study are based partly on a survey of relevant literature sources to
assess the validity of these inputs and their appropriateness for use in representing North
American conditions over the study time period. Data is primarily sourced from high quality
and reputable secondary sources, such as material associations and published literature.
This section describes the specifics for key data sources and parameter inputs. Several
input parameters are tested in sensitivity analyses to assess their influence on the baseline
results over time (see Section 4.5).
4.1.1 Material Inputs
Steel and aluminum comprise the majority of a vehicle’s bill of materials. Furthermore, the
replacement of steel with aluminum or AHSS in vehicle body and closures is modeled to only
affect the production levels of those two materials in this study. There is sufficient capacity
in North America to supply automotive steels for North American vehicle production and the
use of imported steel in automotive applications is very limited based on information from
North American steel industry experts. However, there is no North American steel slab or
finishing data currently available. As shown in Table 1, the steel input data used in this study
is based on the latest LCI data published by the World Steel Association and represents
global average production, which includes North American sites, between 2006 and 2009.
The global LCI data was benchmarked against aggregated cradle-to-steel mill gate North
American LCI datasets for hot-dip galvanized (2.16 kg CO2eq/kg) and hot rolled coil (2.00 kg
CO2eq/kg). These datasets are comprised of a blend of BF/BOF and EAF steel production
and are comparable to the global average values used in this study. The World Steel
7|Page
Association (worldsteel) released new global average LCI datasets in September 2017.
These datasets are aggregated on a cradle-to-gate basis and the unit process data necessary
for the modeling in this study has not been released. However, the LCI data used in the CLCA
Model was benchmarked against the new worldsteel data for hot-dip galvanized (2.7 kg
CO2eq/kg) and hot rolled coil (2.22 kg CO2eq/kg). These datasets are comprised of a blend
of BF/BOF and EAF steel production and are comparable to the global average values used
in this study. Furthermore, updated North American average LCI datasets are currently in-
progress, but not yet available.
Table 1. Cradle-to-gate GHG Intensity (GWP-100, IPCC AR5) for Steel and Aluminum Inputs
Material GHG Intensity Source
(kg CO2eq/kg)
Steel, BF/BOF slab 1.87 World Steel Association (WSA), 2010; representing global average
Steel, EAF slab 0.40 production between 2006-2009.
Mild steel finishing, 0.485
flat
Mild steel finishing, 0.29
long
AHSS finishing, flat 0.534 Calculated assuming a 10% increase over mild steel finishing.
AHSS finishing, long 0.319
Aluminum, primary 8.937 GaBi dataset: “RNA: Primary Aluminum Ingot AA, primary
ingot – North production, consumption mix”; reference year 2011.
American
Aluminum, primary 16.5 GaBi dataset: “GLO: Aluminum ingot mix, International Aluminium
ingot – imported Institute (IAI)”, reference year 2010.
Aluminum, secondary 0.68 GaBi dataset: “RNA: Secondary Aluminum Ingot AA, production mix,
ingot at producer”; reference year 2010.
Aluminum finishing, 0.5 GaBi dataset: “EU-27: Aluminium sheet ts <p-agg>”
rolling
Aluminum finishing, 0.73 GaBi dataset: “EU-27: Aluminium extrusion profile ts <p-agg>”
extrusion
Note: GaBi datasets are from the thinkstep GaBi Professional Database Service Pack 30, released July 2016.
http://www.gabi-software.com/international/databases/.
AHSS represents a broad spectrum of steel grades ranging from slightly higher strength
grades to significantly higher strength grades. While AHSS does not require a significantly
greater degree of processing versus other sheet steel products, specific LCI data does not
exist for AHSS products. To account for this uncertainty, the GHG intensity of the finishing
processes for mild steel have been increased by 10% for the base case, as noted in Table 1
above.
The model separately accounts for primary and secondary material production (in the
aluminum case) and EAF vs. integrated production (in the mild steel and AHSS cases), as
well as finishing processes, such as rolling, extruding, casting, and galvanizing. The
aluminum share of secondary production increases over time based on the amount of
prompt scrap and end-of-life scrap available for recycling back into automotive sheet via a
closed-loop recycling approximation (see Section 4.3). The share of flat steel produced by
the EAF route is 6% for the base case; however, the use of EAF-produced sheet in automotive
applications is increasing, so this value was tested in the sensitivity analyses (see Section
4.5).
Aluminum primary ingot GHG intensity values (shown in Table 2 above) are based on the
most recent North American average production mix data collected by the Aluminum
Association and global average production data collected by the International Aluminium
Institute. The CLCA model uses import shares for primary aluminum of 22% in 2015 and
34% in 2016, which represent the actual shares based on import, export, and production
8|Page
data from the Aluminium Association of Canada 15, U.S. Geological Survey 16, and the
Canadian International Merchandise Trade Database, part of Statistics Canada 17. The import
share (where “import share” refers to imports into North America, and does not include
imports from Canada into the United States or vice versa) is modeled to increase at a rate of
3.6% per year based on the increasing demand for aluminum in North American vehicles as
projected by the Ducker study commissioned by the Aluminum Association.
The CLCA model also includes a downward trend in imported primary aluminum GHG
intensity (reduction of 0.9% per year until 2029) to reflect expected improvements in
aluminum production processes and electricity grid mixes from a Hao, et al., 2015 study 18.
Using global average data to model imported primary aluminum ingots is a conservative
approach because it includes production from regions with relatively low GHG intensity, such
as North America, whereas the regions most likely to export primary aluminum to North
America have coal and natural gas-fueled production. 19 The global average also includes
primary aluminum production in China. While at the time of data collection, most primary
aluminum produced in China was consumed within the country, China has become a
significant exporter of primary aluminum over the last several years. 20 In fact, according to
Aluminum Association data, China now produces more than half of the world’s primary
aluminum and U.S. imports of semi-fabricated aluminum products from China grew 183%
between 2012 through 2015 before leveling off in 2016. 21 The majority of the Chinese
imports to the U.S. are comprised of sheet and plate products.
Furthermore, according to the Aluminum Association’s energy production accounting

methods, 75% of North American primary ingots are produced using electricity generated by
hydropower. 22 This accounting is embedded in the GHG intensity of North American
produced primary aluminum shown in Table 1. The validity of this assumption is not well
documented, but was maintained in the base case modeling for this study as a conservative
assumption. Secondary ingot data represents North American average production in 2010
based on data collected by the Aluminum Association.
4.1.2 Material Replacement Coefficients
The material replacement coefficients (MRCs) indicate the ratio of substituting one material
for another. For example, the MRC for advanced high strength steel (AHSS) is 0.75 lb
AHSS/lb mild steel. 23 This means, by switching from mild steel to AHSS in a given
application, 25% less steel by weight is needed. The MRCs used in this study are shown in
Table 2.
15 Aluminum Association of Canada. “Canadian Primary Aluminium Production”, October 2017. Available at:
http://aluminum.org/sites/default/files/CanadaPrimaryProduction092017.pdf.
16 U.S. Geological Survey. “2015 Minerals Yearbook: Aluminum” and “2017 Mineral Commodity Summary: Aluminum”.
Available at: https://minerals.usgs.gov/minerals/pubs/commodity/aluminum/.

17 Statistics Canada, Canadian International Merchandise Trade Database. “Merchandise imports and exports between
"Canada" and "World", by Harmonized System section, customs basis, September 2017”. Available at:
http://www5.statcan.gc.ca/cimt-cicm/home-accueil?lang=eng.
18 Hao H., Geng Y., and Hang W. “GHG emissions from primary aluminum production in China: Regional disparity and policy
implications,” Applied Energy, 2015. Published online: http://dx.doi.org/10.1016/j.apenergy.2015.05.056.

19 Accenture, LLC. “North American Primary Aluminum Smelter Study,” prepared for the American Iron and Steel Institute.
December 7, 2015.
20 Aluminum Association. “Getting Trade Right – Addressing Chinese Overcapacity.” Available at:
http://www.aluminum.org/getting-trade-right. Accessed October 2017.

21 Ibid.
22 Aluminum Association. “The Environmental Footprint of Semi-Finished Aluminum Products in North America: A Life Cycle
Assessment Report.” December 2013.

23 WorldAutoSteel. UltraLight Steel Auto Body – Advanced Vehicle Concepts (ULSAB-AVC) Programme. www.ulsab.org.
9|Page
Table 2. Material Replacement Coefficients for Steel and Aluminum.
Material MRC Unit Source
AHSS replacing 0.75 lb AHSS/ lb mild steel WorldAutoSteel ULSAB-AVC 24
mild steel
AHSS replacing 0.805 lb AHSS/lb average Calculated based on the
average steel steel average of the AHSS MRCs
(0.75 – 0.86) found in the SRI
2017 25 ALCA study.
Aluminum 0.67 lb aluminum/ lb mild A2Mac1 Automotive
replacing mild steel Benchmarking, 2017 26;
steel average of vehicle types.
Aluminum 0.782 lb aluminum/ lb Calculated based on the
replacing average average steel AHSS and aluminum mild
steel steel MRCs using a 50%
share of aluminum replacing
mild steel and 50% AHSS.
The aluminum to mild steel MRC is based on actual body structure and closure lightweighting
from a mass benchmarking study using the A2Mac1 tear-down database. 27 The study
entailed a detailed analysis of approximately 250 vehicles representing MY 2011-2015. The
MRC for replacing mild steel with aluminum represents average nominal designs. In this
CLCA, an MRC for replacing aluminum with average steel was used as the base case. As
newer vehicle models are developed, aluminum would need to replace lighter advanced
steels in addition to mild steel and in increasingly more mass efficient vehicle designs. In the
A2Mac1 study, mass efficient designs lightweighted with aluminum achieved significantly
less mass savings, i.e., MRCs increased to 0.78-0.79.
A recent paper by Hottle, et al., 2017 28 assessed several published automotive LCA studies,
which include aluminum to mild steel MRCs in the range of 0.46 – 0.74. Three of the five
non-SMDI studies with specific MRCs are on the high end of the range, including Dubreuil
2012 (0.74), Marretta 2012 (0.65), and Stasinopoulos 2012 (0.70). There are only two
studies with lower MRCs, specifically Raugei 2015 (0.5) and Baroth 2012 (0.46).
Furthermore, a paper by Kelly, et al., 2015 29 presents a wide range of MRC values for “Steel
to Wrought Al” with most between 0.48 – 0.72 for body or general part applications. There
are outliers of 0.29 and 0.99. Specific part or component-level MRCs are generally not
applicable to this study, as they would be applied across all components and systems for
every vehicle type in the fleet we are modeling. The aluminum to mild steel MRC used in the
study (0.67) is considered reasonable and was tested in a sensitivity analysis (see Section
4.5).
For AHSS, the selected baseline MRC is based on steel industry applied research through the
WorldAutoSteel ULSAB-AVC Programme and experience in vehicle designs as described in
24 Ibid.
26 A2Mac1 Automotive Benchmarking. Malen, D.; Nagaraj, B.; and Singher, B. “Automotive Mass Benchmarking.” February
2017.
27 Ibid.
28 Hottle, T., Caffrey, C., McDonald, J., and Dodder, R. “Critical factors affecting life cycle assessments of material choice for

29 Kelly, J.C., Sullivan, J.L., Burnham, A, and Elgowainy, A. “Impacts of Vehicle Weight Reduction via Material Substitution on
Life-Cycle Greenhouse Gas Emissions,” Environ. Sci. Technol. 2015, 49, 12535−12542. DOI: 10.1021/acs.est.5b03192.
10 | P a g e
the SRI 2017 ALCA study 30. The A2Mac1 database used to derive the aluminum MRCs does
not distinguish between different steel types, such as AHSS and mild steels; therefore, it
could not be used to discern a specific MRC for AHSS. The base case AHSS MRC is the
average of the range of MRC values presented in that study (0.75 – 0.86) and reflects the
fact that, like aluminum, it is anticipated that newer grades of AHSS will be replacing both
mild steel and earlier grades of AHSS during the study period.
The MRC values for both aluminum and AHSS are assessed in a sensitivity analysis
described in Section 4.5.
4.1.3 Secondary Mass Savings
Secondary mass savings (SMS) or secondary mass change indicates additional vehicle mass
that can be saved, for example in braking and suspension systems, as a result of the
reduction in mass due to primary lightweighting. This secondary savings, like primary mass
reduction, must fulfill performance and function requirements. SMS is typically expressed as
a percentage of primary weight savings. The primary mass reduction in this study occurs in
the body structure and closures. As a result, additional mass savings would come from other
systems, such as the chassis and braking systems. The composition of secondary savings
modeled in the base case is shown in Table 3.
Table 3. Composition of Secondary Mass Savings

Material Secondary savings (%)
Flat products 40%
Long products 30%
Castings 30%
Total 100%
The most commonly used SMS in literature surveys of previous studies 31,32 is 50%; however,
this value is typically applied at the individual component-level and is generally lacking
justification beyond industry rules of thumb and expert opinion. In a 2013 study by Malen et
al., 33 the estimated mass influence coefficients using analytical and regression methods is
0.342 and 0.406 ± 0.052 (34.2% and 40.6 ± 5.2%), respectively. These represent the sum
of various subsystem lightweighting potentials. A 2012 study by Alonzo, et al, 34 found the
mean theoretical SMS potential is 0.95 kg for every kg of primary mass savings; however,
when realistic manufacturing and design limitations were implemented, the SMS potential
decreased to a mean of 0.12 kg/kg. The base cases in this study use a SMS of 20% to
conservatively reflect realistic limitations on current vehicle design. This value is assessed in
a sensitivity analysis in Section 4.5.
4.1.4 Lifetime Driving Distance
The lifetime driving distance parameter refers to the total number of miles a vehicle can be
expected to be driven during its useful lifetime. A 2006 study by the National Highway Traffic
Safety Administration (NHTSA) reports lifetime mileage for passenger cars to be 152,137
31 Malen, D.; Gobbels, R; and Wohlecker, R. “Secondary Mass Changes in Vehicle Design Estimation and Application.” Prepared
for WorldAutoSteel. January 2013.

32 Alonso, E.; Lee, T.M.; Bjelkengren, C.; Roth, R.; and R. Kirchain. “Evaluating the Potential for Secondary Mass Savings in
Vehicle Lightweighting.” Environmental Science & Technology. 2012, 46, 2893−2901.

33 Malen, D.; Gobbels, R; and Wohlecker, R., 2013. Ibid.
34 Alonso E.; Lee, T.M.; Bjelkengren, C.; Roth, R.; and R. Kirchain, 2012. Ibid.
11 | P a g e
miles and light-duty trucks to be 179,954 miles. 35 The Oak Ridge National Laboratory
Transportation Data Book 36 and the GREET 2 (2016) vehicle life cycle model 37 use 150,000
miles for passenger vehicles and 180,000 miles for light-duty trucks. The base case
scenarios in this study use 251,500 km (approximately 156,000 miles), based on an
approximately 21% share of light-duty trucks and utility vans in the modeled vehicle fleet at
an average of 180,000 miles and a 79% share of passenger cars at 150,000 miles. This
parameter is also assessed by a sensitivity analysis in Section 4.5.
4.1.5 Fuel Reduction Values and Power Train Resizing
Fuel reduction values (FRVs) represent the amount of energy that is saved for a given
amount of mass savings and are expressed in units of liters/(100km*100kg). In this study,
FRVs are used to calculate the use phase GHG emissions savings from lightweighting. They
were derived from the Dr. Don Malen (University of Michigan) & Dr. Roland Geyer (University
of California, Santa Barbara) Power Train Model for different vehicle classes and power train
types with and without resizing. 38 FRVs for HEVs were derived from engine map simulations
conducted in 2010-2011 by Forschungsgesellschaft Kraftfahrwesen mbH Aachen (fka) 39, an
automotive engineering and strategic consulting firm. The Malen & Geyer Power Train model
was developed independently, and the model and supporting documentation will be
published separately.
FRVs with power train resizing assume a fully optimized engine design matched to the new
lightweighted vehicle, which is not typically demonstrated in practice. Automakers have a
fixed number of engine and transmission combinations available to incorporate into their
vehicle designs. Therefore, entire power train systems or power train components are often
used in several vehicle models and many vehicle models are offered with more than one
power train option. Since this study is of average vehicle types and considering the practical
limitations of power train resizing, an approximation of 25% power train resizing was selected
for the base case scenarios in this study. This calculation adds 25% of the difference in the
FRVs with and without power train resizing to the FRV without resizing for each vehicle type
to approximate the benefits of power train resizing realized by an average vehicle. While an
arbitrary value, it is most likely that the benefit of power train resizing is closer to 0% than
50% or 100%.
The amount of power train resizing was assessed in a sensitivity analysis as described in
Section 4.5.
4.1.6 Recycling Rates
Recycling rates are modeled separately for each material and for prompt and end-of-life
scrap in the Model. Values used in this study are shown in Table 4. These values remain
constant over time due to lack of information to support an alternative assumption; however,
35 National Highway Traffic Safety Administration. “Vehicle Survivability and Travel Mileage Schedules”. DOT HS 809 953
Technical Report. January 2006.
36 Davis, S.; Williams, S.; Boundy, R. Oak Ridge National Laboratory, Center for Transportation Analysis, Energy and
Transportation Services Division. Transportation Energy Data Book, Edition 35, ORNL-6992. October 2016. pp. 3-17 and 3-18.
37 Argonne National Laboratory. Greenhouse gases, Regulated Emissions, and Energy use in Transportation (GREET) 2 Vehicle
Life Cycle Model, 2016 Update, Revision 1. January 2017. Available at: https://greet.es.anl.gov/index.php.
38 Geyer, Roland. “Life Cycle Energy and Greenhouse Gas (GHG) Assessments of Automotive Material Substitution: User Guide
for Version 5 of the UCSB Automotive Energy and GHG Model.” University of California at Santa Barbara, CA, on behalf of
WorldAutoSteel. March 2017.
39 FKA simulations in 2010-2011 based on their 2007 report “Determination of Weight Elasticity of Fuel Economy for
Conventional ICE Vehicles, Hybrid Vehicles and Fuel Cell Vehicles,” Report 55510, Forschungsgesellschaft Kraftfahrwesen
(FKA), Aachen, Germany
12 | P a g e
the amount of material recycled over time changes as a function of aluminum and steel
production and vehicles reaching their end-of-life during the study period.
Table 4. Recycling Rates for Aluminum and Steel

Prompt Scrap Recycling Rate End-of-life Scrap
Steel Aluminum Steel Aluminum
Collection Rate 99% 99% 97% 97%
Shredder Rate - - 98% 90%
4.2 CLCA Model Calculation Procedures
All of the calculation procedures used in the Model are detailed in the report “Consequential
Life Cycle Assessment (CLCA) of Replacing Steel with Aluminum in Vehicles: User Guide and
Final Project Report”, dated February 26, 2016, which can be found in Appendix A.
4.3 Allocation
Consequential LCA does not require allocation procedures since all significant changes in
process activities are included through system expansion. In this study, system expansion is
required when the changes in scrap inputs to metal production and scrap outputs from
material forming and vehicle end-of-life management are recycled in an open loop. As a
conservative assumption, aluminum was modeled as being recycled in a closed loop
process, which was tested as a sensitivity case (see Section 4.5). For steel scrap, open loop
recycling was assumed for the scope of this study, which is specific to the automotive sector.
On a more general scale, steel scrap recycling can be approximated as a closed loop system;
however, in practice, it is readily recycled from one product system or sector to another
based on where scrap is demanded. This is considered a conservative approach as the use
of closed loop recycling for aluminum provides the lowest cumulative GHG results as can be
seen in the sensitivity analysis results (see Section 5.2). The Model system expansion
equations are detailed in Section 4 (Appendix A: Consequential system expansion) of the
Model report, which is included with this report as Appendix A.
4.4 Data Validation and Data Quality Assessment
Data quality was assessed for the inventory data used as inputs to the model. Specific
discussion of key input parameters is included in Sections 4.1.1-4.1.6. Validation was
performed through benchmarking against published literature sources and by sensitivity
analyses (see Section 4.5).
ISO 14044 includes a set of data quality requirements to address time-related, geographical,
and technology coverage; data precision, completeness, and representativeness;
consistency, reproducibility, sources, and uncertainty. These criteria are described below in
the context of this study.
4.4.1 Time-related Coverage
This criterion addresses the age and timeframe of data. This study utilizes the most current
secondary data available from reputable sources, such as industry associations and the GaBi
professional database, which is subjected to a rigorous quality assurance process, is widely
distributed, and has been used globally in many critically reviewed and published LCA
studies. Most data inputs are less than 5 years old. The steel input data is approximately 10
years old. New global steel LCI data was released in September 2017, but is available only
13 | P a g e
for an aggregated cradle-to-gate scope. As described in Section 4.1.1, the LCI data used in
this study was benchmarked against the new global average data and was found to be
reasonably comparable. New North American steel LCI data is currently being collected;
however, this data was not yet available at the time of this study. The aluminum input data is
6-7 years old.
Reasonable assumptions have been made about whether input data will vary over the
project time horizon or remain relatively constant. Projecting future processes is a necessary
part of this and many consequential LCAs. To address inherent uncertainty, a sensitivity
analysis has been conducted (see Section 4.5).
4.4.2 Geographical Coverage
This criterion addresses the geography from which data was collected for unit processes to
satisfy the goal of the study. The study focuses on North American production of average MY
2016 vehicles. The driving cycle used to model the use phase driving conditions is the U.S.
combined cycle. Electricity and fuel inputs represent U.S. average production, which is
appropriate for a fleet-level study of average vehicles. Aluminum inputs represent the
percentage split of North American and imported primary aluminum and all North American-
produced secondary ingot. The share of imports changes over time as described in Section
4.1.1. Global average data was used to model steel inputs due to the aggregated nature of
North American average production LCIs for steel. However, the global average data was
benchmarked against the North American average data and was found to be a good
approximation.
4.4.3 Technology Coverage
This criterion addresses the technology mix represented in the data. For material inputs, the
amount of primary and secondary material is modeled separately to reflect the different
technologies used in those production routes. For aluminum, the model assumes an
increase in secondary content over the study period as a result of an increase in both prompt
and end-of-life aluminum scrap becoming available for recycling back into automotive sheet.
For steel production, the split of integrated and EAF production for automotive applications is
independent of the availability of automotive scrap since steel is readily recycled from one
product system or sector into another without loss of quality. Use phase fuel and electricity
inputs are representative of average U.S. production technologies.
4.4.4 Precision
The precision of the data is a measure of the variability in the values for each data input (e.g.
variance). Data precision is addressed in this study through the use of individual parameter
sensitivity analyses.
4.4.5 Completeness
Completeness addresses the percentage of flow that is measured or estimated. As described

in Section 3.4, the cut-off criterion of this study is to capture at least 95% of all GHG
emission changes caused by the studied replacement of primarily mild steel closures and
body parts with equivalent aluminum closures and body parts, or equivalent AHSS closures
and body parts. The recycling stage is based on the amount of material recovered and
recycled during auto part manufacturing (i.e., stamping and forming) and at vehicle end-of-
life. Overall data completeness is also judged to be high in the context of this study as data is
14 | P a g e
derived from reputable secondary sources, such as industry associations and the GaBi
professional database.
4.4.6 Representativeness
Data representativeness entails a qualitative assessment of the degree to which the dataset
reflects the true population of interest. In reference to the discussion in Sections 4.4.1 to
4.4.3, the time, geographic, and technology coverage of the data used in this study is
considered representative for the assessment of North American light-duty vehicle
production.
4.4.7 Consistency
Consistency entails a qualitative assessment of whether the study methodology is applied

uniformly to the various components of the analysis. The data used in this study for material
inputs represents a consistent scope and was collected from consistent sources, including
industry associations and the GaBi database.
4.4.8 Reproducibility
This criterion entails a qualitative assessment of the extent to which information about the
methodology and data values would allow an independent practitioner to reproduce the
results reported in the study. The use of the CLCA Model, along with the use of documented
sources for input values, would allow any practitioner to replicate the base case results and
sensitivities presented in this study.
4.4.9 Sources of Data
This data quality criterion simply requires an assessment of data sources in the context of
data quality. The input data for this study is sourced from reputable secondary data sources,
such as industry associations and the GaBi professional database. These sources are
described in detail in Section 4.1.1 to 4.1.6.
4.4.10 Uncertainty
This criterion simply requires an assessment of data uncertainty. Data uncertainty is

addressed in this study through the use of individual parameter sensitivity analyses.
4.5 Sensitivity Analysis
Sensitivity analysis is typically included in life cycle studies to determine the influence of key
assumptions, methods, and input data on study results. A sensitivity analysis compares the
base case results with results obtained using variations on key parameters or assumptions.
In the absence of published literature, these variations are often quantified by adjusting data
by a specified range, such as ±15%.
Individual sensitivity tests were conducted by varying single parameters in the base case
scenario to assess the influence of key parameter assumptions on the results. The scenarios
included in this study are listed below. Results of these assessments are presented in
Section 5.2.
15 | P a g e
• S1: Aluminum Material Replacement Coefficient Varied from 0.5 – 0.8 lb Alum./lb
Mild Steel and AHSS MRC Varied from 0.75 – 0.86 lb AHSS/lb Average Steel.
• S2: AHSS Greenhouse Gas Intensity Increased by 25%.
• S3: Open loop, closed loop prompt scrap, and closed loop prompt and EOL scrap
scenarios considered for both aluminum and AHSS.
• S4: Secondary Mass Savings varied from 0% to 50%% for both aluminum and AHSS.
• S5: Percentage of power train resizing varied from 0% to 50% for both aluminum and
AHSS.
• S6: Growth rate of imported primary aluminum varied from 0% to 6%.
• S7: Flat steel stamping yield varied from 50% to 60%.
• S8: Aluminum sheet stamping yield varied from 50% to 55%.
• S9: Imported primary aluminum GHG intensity varied ±25% from baseline.
• S10: Mean vehicle lifetime varied from 10 years to 16 years.
• S11: Vehicle Lifetime Driving Distance Varied from 200,000 km (Approx. 124,000
Miles) to 300,000 km (Approx. 186,000 Miles).
16 | P a g e
5. LIFE CYCLE IMPACT ASSESSMENT (LCIA) RESULTS
The CLCA Model calculates cumulative GHG emissions from all life cycle phases over the
period from 2015 to 2053 for the considered scenario. These GHG emissions are assessed
as global warming potentials measured on a 100-year time horizon (GWP-100), in kg carbon
dioxide equivalents (CO2eq) using GWP characterization factors from the International Panel
on Climate Change (IPCC) 5th Assessment Report 40. These methods are the most up-to-date
with high international acceptance. While this is a limitation of the study, GWP-100 has high
environmental relevance specific to assessing climate change impacts, which is of high
public and institutional interest, and is among the currently most pressing environmental
issues, especially relative to the production and operation of vehicles. The focus on GHG
emissions is consistent with the majority of automotive LCA studies published in literature. 41
Furthermore, the use of a custom-built Excel model for this study was not conducive to
assessing other LCIA indicators simultaneously.
This section contains the results of studied scenarios for GWP-100 on the basis of the
chosen functional unit. The reported results represent impact potentials and should be
interpreted as approximations of environmental impacts that could occur if the emissions
followed the underlying impact pathway. LCIA results are therefore relative expressions only
and do not predict impacts on category endpoints (or actual impacts), the exceeding of
thresholds, safety margins, or risks.
5.1 Base Case Results
This section presents the results of the base case scenarios comparing an increase in the
use of aluminum versus a similar increase in the use of AHSS for lightweighting of vehicles.
Results are presented in graphical format, along with a value for peak net emissions and a
“breakeven year” for the aluminum option.
As shown in Figure 2, the peak cumulative emissions for the baseline aluminum option are
209 million metric tons CO2eq, occurring at about the year 2038, and the “breakeven year”,
where cumulative GHG emissions return to zero, is calculated as 2084. The AHSS option
results in cumulative GHG emissions savings of over 123 million metric tons in 2038,
increasing to over 261 million metric tons CO2eq by 2053. (See accompanying Model files for
the base case detailed results.)
40 International Panel on Climate Change (IPCC). 5th Assessment Report (AR5). “Climate Change 2013: The
Physical Science Basis.” 2013.

41Hottle, T., Caffrey, C., McDonald, J., and Dodder, R. “Critical factors affecting life cycle assessments of material choice for
17 | P a g e
Figure 2. Base Case Results for Aluminum and AHSS Options
As supplemental information to the baseline results shown above in Figure 2, the following
figures show the annual difference in emissions, both net and by life cycle phase, for the
AHSS and Aluminum lightweighting scenarios at 5 different points in time. Since the AHSS
lightweighting scenario is modeled as an open-loop system, the values for each phase and
the net result are taken directly from the model’s results. The Aluminum lightweighting
scenario is modeled as a closed-loop system, so the model does not separately report a
recycling credit. Instead, the significant credits assigned to aluminum recycling are implicitly
included in the model’s avoided burden calculations. However, by combining the material
manufacturing footprint from the Aluminum open-loop sensitivity (which models 98% primary
aluminum use) with the use phase footprint (which is identical in both closed and open loop
cases) and the final emissions from the baseline closed-loop scenario, the aluminum
recycling credit was calculated for use in these graphs.
These results show that for the AHSS lightweighting scenario, both the manufacturing and
use phase annual footprints are negative, resulting in a negative net footprint which grows
every year. By contrast, in the Aluminum lightweighting scenario the production emissions
are extremely high every year, and at no point are there greater annual emissions reductions
than in the AHSS scenario.
18 | P a g e
NOTE: Over 99% of the vehicles produced in 2020 will reach end-of-life before the end of the
study period (2053), and 100% of the use phase benefits for vehicles produced in 2020 are
accounted for within the model.
19 | P a g e
NOTE: 48% of the vehicles produced in 2040 will reach end-of-life before the end of the
20 | P a g e
NOTE: Less than 1% of the vehicles produced in 2050 will reach end-of-life before the end of
the study period (2053), and 23% of the use phase benefits for vehicles produced in 2050
are accounted for within the model.
21 | P a g e
5.2 Sensitivity Analysis Results
As described in Section 4.5, a sensitivity analysis was conducted for key scenarios to test the
influence of variations in assumptions, methods, and data on the study conclusions. The
result of each such analysis is shown in graphical format below.
Figure S1: Aluminum Material Replacement Coefficient Varied from 0.5 – 0.8 lb Alum./lb,
Mild Steel and AHSS MRC Varied from 0.75 – 0.86 lb AHSS/lb Average Steel
Figure S2: AHSS Greenhouse Gas Intensity Increased by 25%
22 | P a g e
Figure S3: Open Loop, Closed Loop Prompt Scrap, and Closed Loop Prompt and EOL Scrap
Scenarios Considered for Both Aluminum and AHSS
Figure S4: Secondary Mass Savings Varied from 0% to 50% for Both Aluminum and AHSS
23 | P a g e
Figure S5: Percentage of Power Train Resizing Varied from 0% to 50% for Both Aluminum
and AHSS
Figure S6: Growth Rate of Imported Aluminum Varied from 0% to 6%
24 | P a g e
Figure S7: Flat Steel Stamping Yield Varied from 50% to 60%
Figure S8: Aluminum Sheet Stamping Yield Varied from 50% to 55%
25 | P a g e
Figure S9: Imported Aluminum GHG Intensity Varied ±25% from Baseline
Figure S10: Mean Vehicle Lifetime Varied from 10 Years to 16 Years
26 | P a g e
Figure S11: Vehicle Lifetime Driving Distance Varied from 200,000 km (Approx. 124,000
Miles) to 300,000 km (Approx. 186,000 Miles)
These sensitivity analyses show that while varying key inputs does have an effect on the gap
in emissions between the aluminum and AHSS scenarios, in no case do they fundamentally
alter the shape of either curve. The fundamental finding of an increase in emissions
followed by a gradual reduction for aluminum, and an immediate and sustained decrease for
AHSS, remains unaltered.
The sensitivity analysis which showed the most significant effect was clearly the aluminum
material replacement coefficient (S1), followed by the GHG intensity of imported aluminum
(S9). In S9, the most favorable case for aluminum approaches net zero GHG emissions near
the end of the study period. However, the cumulative GHG emissions of the aluminum option
do not come close to matching the GHG emissions savings in the AHSS option. In S1, for
the extreme low end of the range for aluminum MRC, the net cumulative GHG emissions of
the aluminum option reach a peak of over 60 million metric tons, before eventually reaching
the same level of net emissions as the AHSS option in Year 2046. At the extreme high end
of the aluminum MRC range, in the same year (2046), the aluminum option exhibits higher
net cumulative GHG emissions than the AHSS baseline option by over 500 million metric
tons.
The analyses for open vs closed loop recycling (S3) and mean vehicle lifetime (S10) showed
that these factors have a much smaller effect on AHSS than aluminum. This is due to the
27 | P a g e
comparatively small GHG intensity of AHSS, as well as its ability to be readily created from
and recycled into other steel products and sectors through the existing steel recycling
system.
As additional information, the following tornado diagrams show the effect of individual
variables on the net cumulative GHG emissions results. These are based on a specific point
in time, specifically Year 2038, which represents the point of the peak cumulative GHG
emissions in the aluminum option base case.
28 | P a g e
6. INTERPRETATION
6.1 Key Findings
The base case scenario in this study consists of two parts, for which separate Model runs
were conducted. The first Model run assessed the consequences of a significant increase in
the use of aluminum in vehicle body and closure panels, and the second Model run assessed
the consequences of a similar increase in the use of AHSS in vehicle body and closure
panels. The results of these two Model runs are shown on a single graph. (See Section 5.1)
The results show that the aluminum option exhibits a peak cumulative increase in GHG
emissions of about 209 million metric tons CO2eq, with this peak occurring at approximately
Year 2038. (Note that this increase is relative to a baseline of “no lightweighting”.) The
AHSS option results in an immediate and continuous decrease in cumulative GHG emissions.
In fact, because AHSS is lighter than the mild steel it is replacing and has a similar GHG
intensity, the use of AHSS results in a net GHG benefit even without considering the
associated use phase savings. The cumulative GHG emissions savings in the AHSS case
reach a level of over 123 million metric tons CO2eq in 2038, and a maximum value of over
261 million metric tons CO2eq in 2053. At Year 2038, the net difference in GHG emissions
between the aluminum and AHSS options is approximately 332 million metric tons CO2eq,
and this net difference continues to grow throughout the study period until it reaches over
411 million metric tons CO2eq by 2053.
It is important to note that there is a peak followed by a downward trend in the aluminum
results only because of the steady-state aluminum production assumption described in
Section 3.2. This is a conservative assumption due to the lack of automotive aluminum
production projections following 2028. If aluminum production continues to increase instead
of leveling off to a steady-state after this time, the baseline aluminum results would continue
to increase instead of peaking and declining even as the use phase and recycling benefits of
the lightweighted vehicles are realized.
Sensitivity analyses of several key parameters show that while varying some key input
parameters, especially the aluminum material replacement coefficient, does have an effect
on the gap in emissions between the aluminum and AHSS results, in no case do they change
the overall study conclusions. This means, the fundamental finding that a large-scale shift to
the use of increasing amounts of aluminum for automotive body and closure lightweighting in
North America results in a significant increase in GHG emissions followed by a gradual
decline, whereas the use of AHSS to lightweight the same automotive body and closure parts
results in an immediate and sustained decrease in GHG emissions.
6.2 Assumptions and Limitations
The following is a list of key assumptions and limitations that apply to this study:
• Only global warming potential (GHG emissions) are assessed, instead of a more
complete set of LCIA indicators. Therefore, trade-offs amongst other potential
environmental impacts were not evaluated.
29 | P a g e
• Values for many variables are estimated for future years. The details of the
assumptions inherent in these estimates are described in detail in the report. (See
Section 4.1.)
• Secondary data sources are used to establish specific values for several relevant
variables, including global average steel production data from 2006-2009 to reflect
North American production conditions, North American aluminum primary and
secondary production data from 2011, and global average primary aluminum ingot
production data from 2010 to represent North American aluminum imports in lieu of
country-specific data.
• The recycling of aluminum scrap from auto part stamping and forming (prompt scrap)
and from vehicles once they reach end-of-life was modeled as a closed loop recycling
process as a conservative case. It remains unknown whether the infrastructure will
exist to effectively separate and recover most or all aluminum prompt scrap and
aluminum scrap from automotive disassembly and shredding processes during the
study period.
• In LCA, the recycling of steel is often modeled as a closed loop process. However, due
to the Model calculations and structure, the recycling of steel scrap is modeled as an
open loop to avoid artificial increases in EAF steel production used in automotive
applications. Steel scrap from one product application or sector is recycled into many
other product applications depending on the demand for steel scrap. Furthermore,
steel scrap is recycled as an input to all integrated steel production in North America.
EAF steel production for automotive applications is projected to increase; however, this
increase is not dependent on the amount of prompt or end-of-life steel scrap recovered
from vehicles.
• There is a degree of uncertainty associated with some input parameters due to limited
published research and data availability.
• The production of aluminum for North American automotive applications is assumed to

reach a steady-state following 2028. This is a simplifying assumption of this study due
to the lack of aluminum industry projections regarding production trends after this
time.
• The key parameter assumptions of this study have been tested by evaluating various
scenarios in the sensitivity analysis (see Section 4.5), which have shown the results to
be robust. Note that uncertainty in background datasets has not been addressed
specifically aside from the variation of the foreground parameters, such as for the GHG
intensity of steel and aluminum production. Key assumptions and limitations have
been carefully considered during the interpretation phase of this study and in
developing conclusions.
30 | P a g e
6.3 Conclusions
Based on the goal and scope of the study and considering the study limitations described
above, the following conclusions can be drawn:
1. A significant increase in the use of aluminum to lightweight the vehicles in this study,
based on the aluminum industry’s own projections, results in a significant cumulative
increase in GHG emissions, even when use phase savings and GHG benefits due to
recycling are considered. In the base case, this increase reached a peak of 209 million
metric tons of GHG emissions in 2038.
2. The increase in GHG emissions for the aluminum base case is not offset by use phase
savings (due to improved fuel efficiency) and GHG credits for recycling aluminum scrap
until the Year 2084, more than 65 years in the future.
3. Alternately, if the body and closure components of the vehicles in this study were
instead lightweighted with AHSS, the result is an immediate and continuous decrease
in cumulative GHG emissions. The cumulative GHG emissions savings reaches a level
of over 123 million metric tons CO2eq in 2038, and a maximum value of over 261
million metric tons CO2eq in 2053.
4. Based on the scope of this study, and when the aluminum and AHSS lightweighting
scenarios are considered together, the use of AHSS for lightweighting results in a
dramatic and sustained decrease in GHG emissions versus the use of aluminum for
lightweighting of the same vehicles. At Year 2038 (where the aluminum cumulative
GHG results peak), the net difference in GHG emissions between lightweighting with
aluminum versus AHSS is approximately 332 million metric tons CO2eq, which grows
throughout the study period to a maximum of over 411 million metric tons CO2eq in
2053.
31 | P a g e
APPENDIX A: CONSEQUENTIAL LIFE CYCLE ASSESSMENT MODEL USER GUIDE AND
PROJECT REPORT
Consequential Life Cycle Greenhouse Gas Assessment
of Replacing Steel with Aluminum
in North American Vehicle Production
Methodology Report
Prepared by Roland Geyer, Dipl. Phys., Ph.D.

Associate Professor
Donald Bren School of Environmental Science and Management
University of California at Santa Barbara, CA, USA
On behalf of SMDI – Steel Market Development Institute
17 November 2016
TABLE OF CONTENTS
1 INTRODUCTION ..............................................................................................................................................1
1.1 GENERAL ASPECTS ........................................................................................................................................1

1.2 GOAL OF THE STUDY .....................................................................................................................................1
1.3 SCOPE OF THE STUDY ....................................................................................................................................2
1.4 LIFE CYCLE INVENTORY ANALYSIS ...............................................................................................................3
1.5 LIFE CYCLE IMPACT ASSESSMENT .................................................................................................................5
1.6 LIFE CYCLE INTERPRETATION .......................................................................................................................6
2 OVERALL DESCRIPTION OF THE CLCA MODEL ..................................................................................7
3 DESCRIPTION OF THE INDIVIDUAL SPREADSHEETS ....................................................................... 13
3.1 FLEET COMPOSITION ................................................................................................................................... 13

3.2 VEHICLE USE............................................................................................................................................... 15
3.3 MATERIAL PRODUCTION ............................................................................................................................. 18
3.4 SCRAP AT PRODUCTION ............................................................................................................................... 22
3.5 SCRAP AT END-OF-LIFE ............................................................................................................................... 24
4 APPENDIX A: CONSEQUENTIAL SYSTEM EXPANSION ..................................................................... 27
5 APPENDIX B: DEFAULT INPUT DATA ..................................................................................................... 31
5.1 RESULTS & DATA INPUT SPREADSHEET ....................................................................................................... 31

5.2 FLEET COMPOSITION .................................................................................................................................. 34
5.3 VEHICLE USE .............................................................................................................................................. 36
6 APPENDIX C: DEFAULT RESULTS AND INTERPRETATION ............................................................. 37
7 REFERENCES.................................................................................................................................................. 41
Frequent Acronyms
AISI American Iron and Steel Institute
ALCA Attributional life cycle assessment
BEV Pure battery electric vehicle
CLCA Consequential life cycle assessment
CO2eq Carbon dioxide equivalent
EAF Electric arc furnace
GHG Greenhouse gas
HEV Gasoline-based hybrid electric vehicle
ICEV-D Diesel-based pure internal combustion engine vehicle
ICEV-G Gasoline-based pure internal combustion engine vehicle
ISO International Organization for Standardization
LDV Light duty vehicle
MJ Mega Joule
PHEV Plug-in hybrid electric vehicle
SMDI Steel Market Development Institute
1 Introduction
1.1 General aspects

The consequential life cycle assessment (CLCA) reported here was commissioned by the Steel
Market Development Institute (SMDI), a business unit of the American Iron and Steel Institute
(AISI). It was executed by Dr. Roland Geyer and Joe Palazzo from the Bren School of Environ-
mental Science and Management at the University of California, Santa Barbara (UCSB). This
report was completed on February 16, 2016. The CLCA was conducted according to the re-
quirements of ISO 14044:2006. The sole selected impact category is climate change, so strictly
speaking this study is a consequential life cycle greenhouse gas assessment. However, for sim-
plicity the acronym CLCA will be used in the remainder of this report. It is also important to
keep in mind that ISO 14044 was written with attributional life cycle assessment (ALCA) in
mind and therefore some disambiguation or interpretation is necessary when it is applied to
CLCA.
1.2 Goal of the study

SMDI is interested in understanding the greenhouse gas (GHG) implications of a potential large-
scale shift of closure and body part material from steel to aluminum in North American produc-
tion of light duty vehicles (LDV). The intended application of this study is communication with
internal and external stakeholders. The target audience includes all internal and external stake-
holders with an interest in understanding the GHG implications of large-scale automotive mate-
rial substitution for reasons of vehicle mass reduction. The developed methodology could poten-
tially be used to support comparative assertions intended to be disclosed to the public after suc-
cessful critical review by a panel of interested parties. Due to its sole focus on climate change
impacts, the study does not investigate any potential trade-offs across different impact categories.
The CLCA model is completely parameterized and the user is encouraged to conduct extensive
scenario analysis by varying parameters of interest. The parameterization of the model is moti-
vated by the desire to completely separate computational structure and input data, which is not
customarily done in LCA studies (Geyer 2008). The Excel-based CLCA model is populated with
1
a complete set of default input data which generate a default result for illustration purposes. Such
a model design makes it possible to review the computational structure of the CLCA separately
from the default input data and the default model results. Previous experience has shown that
agreement over computational structure is significantly easier to reach than agreement over cor-
rect or appropriate input data values and parameter choices.
1.3 Scope of the study

The functional unit of the study consists of the transportation services (total vehicle miles trav-
eled) provided between 2012 and 2050 by all light duty vehicles produced in North America be-
tween 2012 and 2050. In other words, only vehicles produced between 2012 and 2050 are con-
sidered, and vehicle use or end-of-life (eol) management after 2050 is excluded. This definition
implies that not all included vehicles reach the end of their lives within the study period.
The alternative would be to stop considering vehicle production at a given year, but include ve-
hicle use until all of the included vehicles reach end of life. This creates annual GHG emissions
that are somewhat artificial and difficult to interpret, since vehicle production, use, and eol goes
on concurrently. Therefore, the method explained in the first paragraph was chosen. Defining the
functional unit in such a way means that the total amount of vehicle miles traveled within the
study period is not simply the total number of vehicles produced multiplied by their average life-
time vehicle mileage. Instead, the total amount of vehicle miles traveled within the study period
depends in a nontrivial way on a variety of input parameter choices, such as lifetime vehicle
mileage (in km), average vehicle lifetime (in years), and trend in vehicle production (i.e. annual
vehicle production for each year).
The reference flow of the study consists of two separate parts. The first is a time series of annual
North American light vehicle production, 𝑁𝑁(𝑇𝑇), given in cars per year from 2012 to 2050. For
each year the total production number is broken into five different power train types, 𝑁𝑁𝑖𝑖 (𝑇𝑇),
which are gasoline- and diesel-powered internal combustion vehicles (ICVs), standard and plug-
in hybrid electric vehicles (HEVs), and pure battery electric vehicles (BEVs).
The second part of the reference flow is a time series of annual amount of additional aluminum
body and closure parts, 𝑇𝑇𝑇𝑇(𝑇𝑇), used instead of steel body and closure parts in North American
light vehicle production. For each year the total amount is broken into aluminum sheet, extru-
2
sions, and castings, 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑇𝑇). The total amount is further converted into average amount per ve-
hicle, as a function of powertrain type and vehicle class composition of each power train type.
The modelling methodology does not restrict the selection of input data for 𝑁𝑁(𝑇𝑇) and 𝑇𝑇𝑇𝑇(𝑇𝑇) in
any way. However, the default input data uses forecasts from Ducker Worldwide (Ducker 2014,
2015) for the years 2012 to 2025 and is then kept level between 2026 and 2050. Keeping 𝑁𝑁(𝑇𝑇)
and 𝑇𝑇𝑇𝑇(𝑇𝑇) level for another 25 years has several advantages. First, it is a conservative modeling
approach that allows for the use phase savings to make up for the GHG increases during vehicle
production. Second, it creates a steady state once the last cars of production year 2025 have
reached end of life. This enables the model to calculate a cross-over time, i.e. the time at which
use phase savings have completely made up for the GHG increase during vehicle production and
cumulative GHG emissions have reached zero.
The system boundary is selected in order to capture the main GHG emission changes caused by
the material substitution described in the reference flow. The aim of a CLCA is to account for all
significant changes caused by the reference flow rather than to account for every significant part
of a product life cycle, which is the typical objective of an attributional LCA. This means that all
life cycle stages or processes that do not experience significant GHG emission changes due to
the material substitution are omitted. Examples are the production of tires, vehicle fluids, non-
structural materials, as well as vehicle assembly. The cut-off criterion of this study is to capture
at least 95% of all GHG emission changes caused by the studied replacement of steel closures
and body parts with equivalent aluminum closures and body parts. Equivalent means that the ma-
terial substitution does not affect the technical specifications and safety rating of the vehicles.
The system boundary that results from the cut-off criterion is described in Section 2.
1.4 Life cycle inventory analysis

The scope of the study results in the following set of unit processes:
• Cradle-to-gate production of primary aluminum and steel ingots
o Used in vehicle production
o Displaced due to production and end-of-life scrap recycling
• Scrap-to-gate production of secondary aluminum and steel ingots
o Used in vehicle production
o From automotive production and end-of-life scrap
3
• Gate-to-gate aluminum rolling, extrusion, and casting, steel rolling and casting
• Cradle-to-gate gasoline, diesel, and electricity production
• In-vehicle gasoline and diesel combustion
Since the model results are particularly sensitive to the GHG intensity of primary aluminum in-
got production, this data is modeled as a time series, so that temporal trends in primary alumi-
num production can be included.
The default input data has been collected from a variety of secondary data sources, which are
provided in Section 5 (Default input data) and Section 7 (References). Of particular importance
were previous (attributional) LCAs (ALCAs) of automotive material substitution conducted for
the World Steel Association by the practitioner of this study. These previous studies have a much
larger scope then this CLCA, i.e. they include many more processes, such as production of tires,
batteries, vehicle fluids, and non-structural materials, and life cycle stages, such as vehicle as-
sembly. Version 4 of the ALCA model was used to compare the GHG results of the original
scope with those of the reduced scope used in the CLCA model. The comparison shows that the
reduced scope captures well above 95% of the original scope.
CLCA does not require allocation procedures since all significant changes in process activities
are included through system expansion. In this study, system expansion is required when the
changes in scrap inputs to metal production and scrap outputs from material forming and vehicle
end-of-life management are recycled in an open loop (see Figure 1). The model equations are
detailed in Section 4 (Appendix A: Consequential system expansion). The two critical model pa-
rameters are ∝, which quantifies the response of the scrap market to scrap flow changes, and 𝛽𝛽,
which quantifies the response of the material market to changes in open-loop (external) scrap
recycling. Choosing 𝛼𝛼 = 0 models a scrap market response that is equivalent to the so-called re-
cycled content method in ALCA. Choosing 𝛼𝛼 = 1 and 𝛽𝛽 = 1 models scrap and material market
responses that are equivalent to the so-called avoided burden method in ALCA.
Central to CLCA are the causal relationships used to model and quantify the consequences of the
initial change described by the reference flow. Important relationships in this study are
• The primary and secondary mass reductions caused by replacing steel with aluminum
• The fuel economy improvements caused by vehicle mass reduction
4
• The scrap markets responses caused by changes in scrap supply and demand
• The changes in primary aluminum and steel production caused by changes in secondary
aluminum and steel production
The equations used to model these relationships can be found in Section 3.
1.5 Life cycle impact assessment

This CLCA considers only one impact category, which is climate change. It is thus a life cycle
greenhouse gas assessment, rather than a traditional LCA which considers a range of impact cat-
egories. However, the goal of this study is to assess the GHG implications of a potential large-
scale shift of closure and body part material from steel to aluminum in North American produc-
tion of light duty vehicles (LDV). In other words, the goal is to identify and quantify GHG emis-
sion trade-offs across different consequences of automotive material substitution, not trade-offs
across different impact categories. The exclusion of other impact categories therefore simply re-
flects the goal and scope of the CLCA. In fact, the study could be readily extended to include
additional impact categories if this was desired at some later point.
In the CLCA model, unit processes are described directly in kg CO2eq per unit output, i.e. the
elementary flows have already been classified, characterized and summed up to the indicator re-
sult. Due to the linear nature of unit process scaling and impact assessment it does not matter in
which order those two calculations are executed. To ensure completeness of the impact assess-
ment all process inventories should at least include carbon dioxide, methane, nitrous oxide, sul-
fur hexafluoride, and perfluorinated compounds (PFCs).
Current 100-year global warming potentials (GWPs) from the Intergovernmental Panel on Cli-
mate Change (IPCC) have been built into the model for characterization. This is the case for the
GHG intensities used in the default dataset. Characterization models and GWPs can be found in
the Assessment Reports of the IPCC. Different characterization methods, such as GWP20 or
GWP500 can be applied by entering the applicable emission intensity values. In its current itera-
tion, the model does not track GHGs individually. Therefore, the following limitations apply: i)
it is not possible to quantify the contributions of individual GHGs to the model results; and ii) it
5
is not possible to test time-dependent characterization factors, such as those used in dynamic
characterization models. GWP is a so-called midpoint indicator and model results are therefore
relative expressions and do not predict impacts on category endpoints, the exceeding of thresh-
olds, safety margins or risks.
1.6 Life cycle interpretation

Section 6 presents the results of the CLCA when the default input data are used and interprets
these default results. However, the focus of the model is to generate different plausible scenarios
rather than one authoritative solution to the research question. The results presented in Section 6
should therefore been seen as an illustration of the model, the model output, and the presentation
of the model output. It is important to keep this in mind while reading Section 6.
6
2 Overall description of the CLCA model
The aim of this consequential life cycle assessment (CLCA) is to identify and quantify the main
greenhouse gas (GHG) emission consequences of a shift from steel-based to aluminum-based
body and closure parts in the fleet of light duty vehicles produced in North America between
2012 and 2050. Four distinct sets of changes within and outside of the vehicle life cycles have
been identified as significant and were therefore modeled. They are changes in
1) the production of the steel and aluminum used in the modeled body and closure parts,
and the secondary mass savings
2) the fuel economy of the mass-reduced vehicles,
3) the generation and use of steel and aluminum scrap from material forming processes, and
4) the generation and use of steel and aluminum scrap from vehicle end-of-life management.
Figure 1: System boundaries of the Consequential Life Cycle Assessment
Figure 1 shows a high-level process flow diagram of the resulting inventory model. In terms of
unit processes, the assessment focuses on how the change in automotive material composition
impacts primary and secondary steel and aluminum production, vehicle fuel economy, and fuel
7
production and use. The direct GHG emissions changes of other unit processes, such as forming
and joining, are estimated to be at least one order of magnitude smaller than the modeled ones.
The first consequence of the studied change in automotive material composition is a change in
the amount of steel and aluminum produced for and consumed in North American light vehicle
production. All other things remaining equal, this means a reduction in steel production and
forming and an increase in aluminum production and forming. These changes in material produc-
tion and forming and the resulting changes in GHG emissions, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡), are
modeled on the spreadsheet ‘Material production’. The recycled content of the aluminum body
and closure parts can be set exogenously or is calculated endogenously by assuming closed-loop
recycling of their production or end-of-life scrap. Figure 2 shows the data flow between the Ex-
cel spreadsheets for the calculation of the GHG emission changes from material production.
Figure 2: Data flow for calculating changes GHG emissions from material production,
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡)
The second consequence of replacing automotive steel with aluminum, and the motivation be-
hind it, is a reduction in vehicle mass which results in an increase in fuel economy, i.e. a de-
crease in vehicle energy demand per mile. A variety of factors determine the relationship be-
tween vehicle mass reduction and fuel economy. The two most important ones are the powertrain
type of the vehicle and whether the powertrain is downsized after the vehicle has been mass-
reduced. Both are modeled explicitly in this CLCA, which therefore contains a breakdown of the
vehicle fleet into powertrain segments. The power train segmentation of the vehicle fleet is mod-
8
eled on the spreadsheet ‘Fleet composition’. The vehicle mass reductions, fuel economy im-
provements, and the GHG emission reductions, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡), from the resulting gasoline, diesel,
and electricity savings are modeled on the spreadsheet ‘Vehicle use’. This spreadsheet also con-
tains a vehicle lifetime model, which accounts for the fact that the life of a vehicle is distributed
around a mean value. Figure 3 shows the data flow between the Excel spreadsheets for the calcu-
lation of the GHG emission savings during vehicle use. Figure 4 shows the vehicle lifetime dis-
tribution model.
Figure 3: Data flow for calculating use phase GHG emission savings 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡)
Figure 4: Fraction of vehicles still in use after t years of driving, 𝐹𝐹𝐹𝐹𝐹𝐹(𝑡𝑡), is modeled as one mi-
nus a cumulative log-normal distribution (example for mean 12 years and standard deviation 2
years)
9
Changes in automotive material production and forming also cause changes in the amount of
scrap consumed during automotive material production and generated during material forming,
such as rolling, extruding, stamping, and casting. For the added aluminum the CLCA offers two
different ways to assess the GHG implications of these scrap input and output changes. One way
models aluminum production scrap recycling in a closed loop, i.e. all production scrap from
aluminum body and closure forming is recycled back into body and closure parts. The scrap flow
changes take place entirely within the vehicle life cycles, which is illustrated by the dashed red
line in Figure 1. The other way models the recycled content of aluminum body and closure parts
and the recycling of aluminum production scrap in an open loop. In other words, aluminum scrap
used in body and closure production comes from an external scrap market and the production
scrap goes back into it. For the removed steel, this is the only method used in the model. In
CLCA, such changes in scrap flows across the initial system boundaries require system expan-
sion in order to include significant GHG emission changes outside of the vehicle life cycles,
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡). The processes included through system expansion are exter-
nal automotive aluminum and steel scrap collection and recycling and external primary alumi-
num and steel production. The expanded system is shown in Figure 1. The model calculations
can be found on the spreadsheet ‘Scrap at production’. Figure 5 shows the data flow between the
Excel spreadsheets for calculating 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) caused by open loop recy-
cling of production scrap.
Figure 5: Data flow for calculating the GHG implications of production scrap generation and use
(open loop recycling only), 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡)
10
Finally, changing the material composition of the North American light vehicle fleet also chang-
es the composition of end-of-life vehicle scrap. Again, end-of-life aluminum scrap recycling can
be modeled in an open and a closed loop. End-of-life steel scrap recycling is always modeled in
an open loop. Another system expansion is required when end-of-life recycling is modeled in an
open loop and the changes in end-of-life scrap flows cross the initial system boundaries. The
processes included through system expansion are external automotive aluminum and steel scrap
collection and recycling and external primary aluminum and steel production (as shown in Fig-
ure 1). The model calculations can be found on the spreadsheet ‘Scrap at end-of-life’. Figure 7
shows the data flow between the Excel spreadsheets when the GHG emission implications,
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡), are calculated through open loop end-of-life scrap recycling.
Figure 7 illustrates the time delay between automotive material substitution and the generation of
automotive end-of-life scrap. The time delay is due to the vehicle lifetime and modeled using the
vehicle lifetime distribution 𝐹𝐹𝐹𝐹𝐹𝐹(𝑡𝑡) shown in Figure 4.
Figure 6: Data flow for calculating the GHG implications of end-of-life scrap generation and use
(open loop recycling only), 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡)
11
Figure 7: Illustration of the time delay between automotive material substitution and changes in
end-of-life (eol) scrap generation (all in million kg)
12
3 Description of the individual spreadsheets
3.1 Fleet composition

For each year of the modelling period (2012-2050), this spreadsheet calculates the power train
composition of the light duty vehicles assumed to be produced in North America during that
year. It also calculates, for each production year T and power train type i, the average amount of
body and closure parts made from aluminum, 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇).
The resulting output table contains 5 power train types i, which are Gasoline ICV, Diesel ICV,
Standard HEV, Plug-in HEV, and BEV. For each power train type i and production year T is lists
• 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇): Average amount of body and closure parts made from aluminum (in kg per vehi-
cle).
• 𝑇𝑇𝑇𝑇𝑖𝑖 (𝑇𝑇): Total amount of body and closure parts made from aluminum (in kg). This is the
average amount multiplied by the total number of vehicles of that power train type pro-
duced during that year, 𝑇𝑇𝑇𝑇𝑖𝑖 (𝑇𝑇) = 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) ∙ 𝑁𝑁𝑖𝑖 (𝑇𝑇).
• 𝑃𝑃𝑃𝑃𝑖𝑖 (𝑇𝑇): Share of the power train type i as % of the total number of vehicles produced
during production year T. This is the number of vehicles of that power train type divided
by the total number of vehicle produced during year T.
A significant amount of input data is required to calculate the outputs described above. Below is
a comprehensive table:
• Total amount of aluminum body and closure parts in light duty vehicles produced each
year (in kg), 𝑇𝑇𝑇𝑇(𝑇𝑇). This data is also broken down into sheet, extrusions, and castings for
other modelling purposes, 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑇𝑇) with ∑𝑙𝑙 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑇𝑇) = 𝑇𝑇𝑇𝑇(𝑇𝑇).
• Total number of light duty vehicles assumed to be produced each year, 𝑁𝑁(𝑇𝑇) in #.
• For each year, the share of annually produced vehicles that are ICVs, 𝑃𝑃𝑃𝑃𝐼𝐼𝐼𝐼𝐼𝐼 (𝑇𝑇).
• For each year, the share of annually produced ICVs that are Gasoline ICVs.
• For each year, the share of annually produced vehicles that are HEVs, 𝑃𝑃𝑃𝑃𝐻𝐻𝐻𝐻𝐻𝐻 (𝑇𝑇).
• For each year, the share of annually produced HEVs that are Standard HEVs.
13
• The share of each vehicle class as % of the total number of vehicles produced in 2015,
𝑗𝑗
𝑉𝑉𝑉𝑉𝑎𝑎𝑎𝑎𝑎𝑎 with j being the vehicle class.
• For each vehicle class j, the average amount of aluminum in vehicles produced in 2015
(in lbs).
𝑗𝑗
• The vehicle class composition of HEVs produced in 2014, 𝑉𝑉𝑉𝑉𝐻𝐻𝐻𝐻𝐻𝐻 . Assumed to be con-
stant over the modelling period and the same for Standard and Plug-in HEVs.
𝑗𝑗
• The vehicle class composition of BEVs produced in 2014, 𝑉𝑉𝑉𝑉𝐵𝐵𝐵𝐵𝐵𝐵 . Assumed to be con-
stant over the modelling period.
Eight different vehicle classes are considered. They are A/B, C, D, E, MPV, SUV, VAN, PUP.
Calculating the average amount of body and closure aluminum per vehicle for each power train
type is complicated by the fact that each power train type has a different composition of vehicle
classes and vehicle classes differ in the amount of aluminum they contain.
The first step is to calculate 𝐴𝐴𝐴𝐴 𝑎𝑎𝑎𝑎𝑎𝑎 (𝑇𝑇), the body and closure aluminum added to each average ve-
hicle in production year T. This is done by dividing the total amount of body and closure alumi-
num added in year T by the total number of vehicles produced in year T. The results are in cells
T55:T93. The next step is to express the amount of aluminum per 2015 vehicle for each vehicle
class (in lbs/vehicle) relative to the amount of aluminum per 2015 vehicle across all vehicle clas-
ses (in lbs/vehicle), which was 398 pounds. The results are in cells L53:S53 and denote the
amount of aluminum per vehicle class as percent of average amount of aluminum across all ve-
hicles. Multiplying the added body and closure aluminum per vehicle and production year,
𝐴𝐴𝐴𝐴 𝑎𝑎𝑎𝑎𝑎𝑎 (𝑇𝑇), with those ratios yields the added body and closure aluminum per vehicle and produc-
tion year for each vehicle class, 𝐴𝐴𝐴𝐴𝑗𝑗 (𝑇𝑇). This data is stored in cells L55:S93. On more interme-
diate step needed to calculate the added body and closure aluminum per vehicle for each power
train type and production year is to calculate the vehicle class composition for ICVs:
𝑗𝑗 𝑗𝑗 𝑗𝑗 𝑗𝑗
𝑃𝑃𝑃𝑃𝐼𝐼𝐼𝐼𝐼𝐼 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝐼𝐼𝐼𝐼𝐼𝐼 + 𝑃𝑃𝑃𝑃𝐻𝐻𝐻𝐻𝐻𝐻 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝐻𝐻𝐻𝐻𝐻𝐻 + 𝑃𝑃𝑃𝑃𝐵𝐵𝐵𝐵𝐵𝐵 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝐵𝐵𝐵𝐵𝐵𝐵 = 𝑉𝑉𝑉𝑉𝑎𝑎𝑎𝑎𝑎𝑎
𝑗𝑗 𝑗𝑗 𝑗𝑗
𝑗𝑗 𝑉𝑉𝑉𝑉 − 𝑃𝑃𝑃𝑃𝐻𝐻𝐻𝐻𝐻𝐻 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝐻𝐻𝐻𝐻𝐻𝐻 − 𝑃𝑃𝑃𝑃𝐵𝐵𝐵𝐵𝐵𝐵 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝐵𝐵𝐵𝐵𝐵𝐵
⇒ 𝑉𝑉𝑉𝑉𝐼𝐼𝐼𝐼𝐼𝐼 (𝑇𝑇) = 𝑎𝑎𝑎𝑎𝑎𝑎
𝑃𝑃𝑃𝑃𝐼𝐼𝐼𝐼𝐼𝐼 (𝑇𝑇)
with j being the vehicle class.
14
Since 𝑃𝑃𝑃𝑃𝑖𝑖 (𝑇𝑇) varies with each production year this calculation is repeated for each vehicle class j
and each production year T. The results are stored in cells L110:S148.
Finally, the added body and closure aluminum per vehicle for each power train and production
year, 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇), can be calculated:
𝑗𝑗
𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) = � 𝐴𝐴𝐴𝐴𝑗𝑗 (𝑇𝑇) ∙ 𝑉𝑉𝑉𝑉𝑖𝑖 (𝑇𝑇)
𝑗𝑗
Note that the calculated values are the same for Gasoline and Diesel ICVs, i.e. 𝐴𝐴𝐴𝐴𝐼𝐼𝐼𝐼𝐼𝐼−𝐺𝐺 (𝑇𝑇) =
𝐴𝐴𝐴𝐴𝐼𝐼𝐼𝐼𝐼𝐼−𝐷𝐷 (𝑇𝑇) = 𝐴𝐴𝐶𝐶𝐼𝐼𝐼𝐼𝐼𝐼 (𝑇𝑇), and for Standard and Plug-in Hybrids, i.e. 𝐴𝐴𝐴𝐴𝐻𝐻𝐻𝐻𝐻𝐻−𝑆𝑆 (𝑇𝑇) =
𝑗𝑗
𝐴𝐴𝐴𝐴𝐻𝐻𝐻𝐻𝐻𝐻−𝑃𝑃 (𝑇𝑇) = 𝐴𝐴𝐴𝐴𝐻𝐻𝐻𝐻𝐻𝐻 (𝑇𝑇), since the vehicle class shares 𝑉𝑉𝑉𝑉𝑖𝑖 are assumed to be identical for the
two ICV types and the two HEV types. The total amount of added body and closure aluminum as
a function of power train and production year, 𝑇𝑇𝑇𝑇𝑖𝑖 (𝑇𝑇) is calculated as follows:
𝑇𝑇𝑇𝑇𝑖𝑖 (𝑇𝑇) = 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) ∙ 𝑁𝑁𝑖𝑖 (𝑇𝑇) = 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) ∙ 𝑃𝑃𝑃𝑃𝑖𝑖 (𝑇𝑇) ∙ 𝑁𝑁(𝑇𝑇)
Variable Description Location of data

𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) Average amount of body and closure parts made B5:B43, E5:E43, H5:H43,
from aluminum for powertrain type i and produc- K5:K43, N5:N43
tion year T (in kg/car)
𝑇𝑇𝑇𝑇𝑖𝑖 (𝑇𝑇) Total amount of body and closure parts made C5:C43, F5:F43, I5:I43,
from aluminum for powertrain type i and produc- L5:L43, O5:O43
tion year T (in kg)
𝑃𝑃𝑃𝑃𝑖𝑖 (𝑇𝑇) Share of the power train type as % of the total D5:D43, G5:G43, J5:J43,
number of vehicle produced during production M5:M43, P5:P43
year T
Table 1: Output data from the spreadsheet ‘Fleet composition’
3.2 Vehicle use

For each calendar year t of the modelling period (2012-2050), this spreadsheet calculates the to-
tal amount of GHG savings (in million kgCO2eq) that result from driving the mass-reduced ve-
hicle fleet modeled on the spreadsheet ‘Fleet composition’. These total GHG savings per calen-
dar year are reported in cells AO232:AO269. The total GHG savings in each calendar year t are a
function of the age composition of the fleet during year t, i.e. how many vehicles of each produc-
tion year T were in use. Each production year is characterized by the total number of vehicle
produced, 𝑁𝑁(𝑇𝑇), and the total amount of aluminum closures and body parts, 𝑇𝑇𝑇𝑇(𝑇𝑇), added to
15
those vehicles. For each calendar year t, the total use phase GHG reductions are thus calculated
as the sum of use phase GHG reductions from the vehicles of each production year T. The table
of use phase GHG savings for each year of driving t and each year of vehicle production T is
given in cells B232:AN269.
Use phase savings are calculated separately for each power train type, but the calculation process
is identical. The starting point is the added amount of aluminum body and closure parts per vehi-
cle, 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇), which is given in kg per vehicle. The first step is to calculate the resulting mass re-
ductions per vehicle, ∆𝑀𝑀𝑖𝑖 (𝑇𝑇), which are given in kg per vehicle and are calculated as
(𝑘𝑘 − 1)
∆𝑀𝑀𝑖𝑖 (𝑇𝑇) = 𝐴𝐴𝐴𝐴𝑖𝑖 (𝑇𝑇) ∙ (1 + 𝑠𝑠)
𝑘𝑘
With k being the material replacement coefficient of aluminum relative to steel (in kg alumi-
num/kg steel) and s being the secondary mass savings (in kg secondary mass savings/kg primary
mass savings). The next step is to calculate the life time fuel and electricity savings per car, 𝐹𝐹𝐹𝐹𝑖𝑖
(in liters per car) and 𝐸𝐸𝐸𝐸𝑖𝑖 (in MJ per car), according to the following equations:
𝐹𝐹𝐹𝐹𝑖𝑖 (𝑇𝑇) = (1 − 𝐸𝐸𝐸𝐸) ∙ ∆𝑀𝑀𝑖𝑖 (𝑇𝑇) ∙ ∆𝐹𝐹𝑖𝑖 (𝑇𝑇) ∙ 𝑇𝑇𝑇𝑇 ∙ 0.0001
𝐸𝐸𝐸𝐸𝑖𝑖 (𝑇𝑇) = 𝐸𝐸𝐸𝐸 ∙ ∆𝑀𝑀𝑖𝑖 (𝑡𝑡) ∙ ∆𝐸𝐸𝑖𝑖 (𝑇𝑇) ∙ 𝑇𝑇𝑇𝑇 ∙ 0.0001
with EL being the share of life time driving powered by plug electricity (in %), ∆𝐹𝐹𝑖𝑖 (𝑇𝑇) the fuel
savings per mass savings (in liters per 100km driven and 100kg mass reduction), ∆𝐸𝐸𝑖𝑖 (𝑇𝑇) the
electricity savings per mass savings (in MJ per 100km driven and 100kg mass reduction), and
TM the assumed vehicle life (in km). Life time fuel and electricity savings per vehicle are con-
verted into lifetime GHG emissions savings per vehicle, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖 (𝑇𝑇), according to this equation:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖 (𝑇𝑇) = 𝐹𝐹𝐹𝐹𝑖𝑖 (𝑇𝑇) ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑓𝑓 +𝐸𝐸𝐸𝐸𝑖𝑖 (𝑇𝑇) ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑒𝑒
with 𝐺𝐺𝐺𝐺𝐺𝐺𝑓𝑓 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑒𝑒 being the GHG intensities of the fuel and the electricity, in kg CO2eq per
liter and kg CO2eq per MJ respectively. The final step in calculating the life time use phase GHG
savings per power train type and production year is to multiply the life time use phase savings
per car with the total number of vehicles of the given power train i being produced in each given
production year T, i.e. 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖 (𝑇𝑇) ∙ 𝑁𝑁𝑖𝑖 (𝑇𝑇).
The calculations above require various additional input parameters, such as the material re-
placement coefficient k, the secondary mass savings coefficient s, the plug electricity share EL,
16
the life time vehicle driving TM, the GHG intensities of fuel and electricity 𝐺𝐺𝐺𝐺𝐺𝐺𝑓𝑓 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑒𝑒 , and
finally the fuel and electricity savings per mass savings ∆𝐹𝐹𝑖𝑖 (𝑇𝑇) and ∆𝐸𝐸𝑖𝑖 (𝑇𝑇). The last two pa-
rameters can be modeled as time dependent but are currently assumed to be constant over time,
i.e. ∆𝐹𝐹𝑖𝑖 (𝑇𝑇) = ∆𝐹𝐹𝑖𝑖 and ∆𝐸𝐸𝑖𝑖 (𝑇𝑇) = ∆𝐸𝐸𝑖𝑖 . For each power train type the energy savings per mass sav-
ings are calculated from a set of input parameters. Energy savings per mass savings are signifi-
cantly higher in the case that the power train of the vehicle is resized, i.e. optimized to the new,
reduced vehicle mass. However, this is not always feasible or cost-effective. Also, fuel consump-
tion models show that energy savings per mass savings vary across power train types and vehicle
classes. For this reason, they are calculated as follows for each power train type i:
𝑗𝑗 𝑗𝑗
∆𝐹𝐹𝑖𝑖 = 𝑅𝑅𝑅𝑅 ∙ 𝐴𝐴𝐴𝐴𝐴𝐴�∆𝐹𝐹𝑖𝑖,𝑟𝑟𝑟𝑟 � + (1 − 𝑅𝑅𝑅𝑅) ∙ 𝐴𝐴𝐴𝐴𝐴𝐴�∆𝐹𝐹𝑖𝑖,𝑛𝑛𝑛𝑛−𝑟𝑟𝑟𝑟 �
𝑗𝑗 𝑗𝑗
∆𝐸𝐸𝑖𝑖 = 𝑅𝑅𝑅𝑅 ∙ 𝐴𝐴𝐴𝐴𝐴𝐴�∆𝐸𝐸𝑖𝑖,𝑟𝑟𝑟𝑟 � + (1 − 𝑅𝑅𝑅𝑅) ∙ 𝐴𝐴𝐴𝐴𝐴𝐴�∆𝐸𝐸𝑖𝑖,𝑛𝑛𝑛𝑛−𝑟𝑟𝑟𝑟 �
where RE is the fraction of power train resizing benefit that mass-reduced vehicles can realize on
average, superscript j denotes the vehicle class, and subscripts re and no-re stand for resizing and
no-resizing, respectively. The average is calculated over the input values for the different vehicle
classes (the yellow cells on the spreadsheet in columns J to V). Different numbers of vehicle
classes may be used to characterize different power trains.
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑇𝑇) are the lifetime GHG use phase savings from all vehicles produced in year T and cal-
culated as follows:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑇𝑇) = � 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖 (𝑇𝑇) ∙ 𝑁𝑁𝑖𝑖 (𝑇𝑇)

𝑖𝑖
These values need to be converted into the GHG use phase savings that occur during each calen-
dar year of the modeling period. The first step is to convert the lifetime savings into annual sav-
ings according to the following equation:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑇𝑇)
𝐹𝐹𝐹𝐹𝐹𝐹(𝑡𝑡) ∙ 𝑓𝑓𝑓𝑓𝑓𝑓 𝑇𝑇 < 𝑡𝑡
𝐴𝐴𝐴𝐴(𝑇𝑇, 𝑡𝑡) = � 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙
0 𝑓𝑓𝑓𝑓𝑓𝑓 𝑇𝑇 ≥ 𝑡𝑡
where lifetime is the average lifetime of each vehicle (assumed to be constant across time and all
vehicles), and FIU(t) is the fraction of vehicles still in use after t years of driving. FIU(t) is de-
fined as 𝐹𝐹𝐹𝐹𝐹𝐹(𝑡𝑡) = 1 − 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙(𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙, 𝑆𝑆𝑆𝑆), with lifetime as the mean and an additional
17
parameter SD, the standard deviation. The total GHG use phase savings in each calendar year is
now calculated as the sum of all annual savings across all production years:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) = � 𝐴𝐴𝐴𝐴(𝑇𝑇, 𝑡𝑡)

𝑇𝑇
In the model, vehicles are being produced every year of the modeling period (2012-2050). This
means that not all vehicles will reach the end of their lives and as a result the use phase savings
accruing during the modeling period are smaller than the sum of the lifetime savings of all cars
produced during the modeling period, i.e. the following inequality holds:
� 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) < � 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑇𝑇)

𝑡𝑡 𝑇𝑇

𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡)Total GHG use phase savings per calendar year t AO232:AO269
(in million kg CO2eq)
Table 2: Output data from spreadsheet ‘Vehicle use”
3.3 Material production

The aim of this spreadsheet is to calculate all changes in GHG emissions from material produc-
tion, both for the added aluminum and the removed steel. The spreadsheet calculates and tallies
all direct emission changes, i.e. it reflects to what extent the added aluminum and removed steel
come from primary or secondary production. In other words, the results on this spreadsheet re-
flect the recycled content of the added and removed material. The implications of changes in
scrap input and output are calculated on two separate, dedicated spreadsheets.
The starting point for calculating the changes in production GHGs from the added aluminum are
the total amount of aluminum body and closure parts added to light duty vehicles produced each
year (in million kg), broken down into sheet, extrusions, and castings. The three time series are
denoted by 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡), with subscript l standing for sheet, extrusions, and castings. The first calcu-
lation step is to convert the aluminum contained in the vehicles as body and closure parts into
𝑝𝑝𝑝𝑝
shipped primary and secondary aluminum, denoted by 𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) and 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡), where superscript
pa stands for primary aluminum and sa for secondary aluminum. The calculations are as follows:
18
𝑝𝑝𝑝𝑝 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑆𝑆𝑀𝑀𝑙𝑙 (𝑡𝑡) = �1 − 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡)� ∙
𝛾𝛾𝑙𝑙𝑎𝑎
𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) = 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡) ∙
where 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡) is the recycled (secondary) content of aluminum type l in production year t, and
𝛾𝛾𝑙𝑙𝑎𝑎 is the manufacturing yield of aluminum type l. The results of these calculations are also in
million kg and stored in cells B6:G44. The next step is to calculate the resulting production GHG
emissions for each aluminum type by multiplying the shipped material quantities with the GHG
intensities of aluminum production:
𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙 (𝑡𝑡) = 𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑎𝑎 �
𝑠𝑠𝑠𝑠
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) = 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 + 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑎𝑎 �
𝑝𝑝𝑝𝑝 𝑠𝑠𝑠𝑠
where 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 are the GHG intensities of primary and secondary aluminum in-
got production and 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑎𝑎 the GHG intensities of aluminum ingot rolling, extruding, and casting.
The final step is to sum over all aluminum types l in order to calculate the GHG emissions that
result from increases in aluminum production and forming in any given production year:
𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑎𝑎 (𝑡𝑡) = � �𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙 (𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡)�
𝑙𝑙
The changes in GHG emissions due to increased aluminum production 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑎𝑎 (𝑡𝑡) are stored in
cells R6:R44. Calculating the changes in production GHGs from the removed steel has the same
computational structure, once the total amount of removed steel has been determined. The latter
is complicated somewhat by the fact that it needs to account for both primary and secondary
mass savings and the different steel types l, which are flat, long, and cast. The total amount of
steel type l removed from all vehicles produced in year t is calculated as follows:
𝑇𝑇𝑇𝑇(𝑡𝑡) 𝑇𝑇𝑇𝑇(𝑡𝑡)
𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡) = − ∙ 𝑝𝑝𝑝𝑝𝑙𝑙 − (1 − 𝑘𝑘) ∙ 𝑠𝑠 ∙ 𝑠𝑠𝑠𝑠𝑙𝑙
𝑘𝑘 𝑘𝑘
where 𝑇𝑇𝑇𝑇(𝑡𝑡) is the total amount of aluminum body and closure parts added to the vehicles pro-
duced in year t, k is the material replacement coefficient of aluminum relative to steel (in kg
aluminum/kg steel), s is the secondary mass savings coefficient (in kg secondary mass sav-
ings/kg primary mass savings), and 𝑝𝑝𝑝𝑝𝑙𝑙 and 𝑠𝑠𝑓𝑓𝑙𝑙 are the fractions of primary and secondary mass
savings that are of steel type l. Note that 𝑇𝑇𝑇𝑇(𝑡𝑡) = ∑𝑙𝑙 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡) and is one of the central data inputs
into the model. The subsequent calculation steps are the same as for the added aluminum. First,
19
𝑝𝑝𝑝𝑝
the amount of no longer shipped primary and secondary steel, 𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) and 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡), is calculat-
ed:
𝑝𝑝𝑝𝑝 𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) = �1 − 𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 (𝑡𝑡)� ∙
𝛾𝛾𝑙𝑙𝑠𝑠
𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) = 𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 (𝑡𝑡) ∙
where 𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 (𝑡𝑡) is the electric arc furnace (EAF) content of steel type l in production year t, and 𝛾𝛾𝑙𝑙𝑠𝑠
is the manufacturing yield of steel type l. The next step is to calculate the avoided production
GHG emissions for each steel type by multiplying the no longer shipped material quantities with
the GHG intensities of steel production:
𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙 (𝑡𝑡) = 𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 + 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠 �
𝑠𝑠𝑠𝑠
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) = 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 + 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠 �
where 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 are the GHG intensities of primary and secondary steel ingot
production and 𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠 the GHG intensities of steel ingot rolling and casting. The final step is to
add over all steel types l in order to calculate the GHG emissions that result from reductions in
steel production and forming in any given production year:
𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) = � �𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙 (𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡)�
𝑙𝑙
The changes in GHG emissions due to reduced steel production 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) are stored in cells
R52:R90.
There are three different ways in which recycled content is modeled for aluminum. The first
models recycling as an open loop, which means that scrap inputs and outputs are independent
from each other and modeled exogenously. In this case the recycled content of aluminum sheet,
extrusions, and castings is given through time-dependent input values, which means that recycled
content could be different every year. In the default setting the values are assumed to be constant
over time and are set to zero for the recycled content of sheet and extrusions, and 0.85 for the
recycled content of castings. This open loop methodology is the only way available to model the
EAF fraction for flat, long, and cast steel. In the default setting the EAF fractions are assumed to
be constant over time and are set to 5% for flat, 85% for long, and 100% for cast steel. As with
all input variables, the values can be changed by the model user. The other two approaches avail-
20
able for aluminum model recycling as a closed loop, i.e. the scrap inputs into production of alu-
minum body and closure parts come from scrap generated within the modeled North American
light duty vehicle life cycles. In option one of the closed-loop model, only production scrap is
used in a closed loop, i.e. for aluminum body and closure production; the scrap at vehicle end-of-
life is still recycled externally, i.e. in an open loop. The calculations for shipped primary and
secondary aluminum are now as follows:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) =
𝑠𝑠𝑟𝑟𝑎𝑎
𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) = − 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡)
where 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡) is the amount of prompt scrap of aluminum type l generated and collected dur-
ing calendar/production year t, and 𝑠𝑠𝑟𝑟𝑎𝑎 is the amount of scrap used to produce one kg of recycled
(secondary) aluminum ingot. In option two of the closed-loop model, production and end-of-life
scrap is used in a closed loop, i.e. for aluminum body and closure production. The calculations
for shipped primary and secondary aluminum change to the following:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡) + 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) =
𝑠𝑠𝑟𝑟𝑎𝑎
𝑆𝑆𝑆𝑆𝑙𝑙 (𝑡𝑡) = − 𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡)
where 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑙𝑙𝑎𝑎 is the amount of end-of-life (eol) scrap of aluminum type l generated and collected
during calendar/production year t.

𝑎𝑎 (𝑡𝑡) GHG emission changes from increases in alumi- R6:R44
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺
num production and forming in calendar year t
(in million kg CO2eq)
𝑠𝑠 (𝑡𝑡) GHG emission changes from reductions in steel R52:R90
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺
production and forming in calendar year t (in
million kg CO2eq)
Table 3: Output data from spreadsheet ‘Material production’
21
3.4 Scrap at production
Changing the material composition of North American light duty vehicles produced between
2012 and 2050 changes the quantities and types of scrap that are generated during vehicle pro-
duction. The aim of this spreadsheet is to calculate the GHG implications of changes in the use
and generation of scrap during material production and forming. In rows 2 to 45 this is done for
aluminum, and in rows 48 to 90 it is done for steel.
For aluminum, the starting point is the total amount of aluminum body and closure parts in light
duty vehicles produced each year (in million kg), broken down into sheet, extrusions, and cast-
ings, i.e. 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡), with subscript l standing for sheet, extrusions, and castings. These values are
turned into amounts of generated and collected production scrap as follows:
(1 − 𝛾𝛾𝑙𝑙𝑎𝑎 )
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡) = 𝑐𝑐𝑝𝑝𝑝𝑝𝑝𝑝
𝑎𝑎
∙ ∙ 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑎𝑎
where 𝑐𝑐𝑝𝑝𝑝𝑝𝑝𝑝 is the production scrap collection rate for aluminum and 𝛾𝛾𝑙𝑙𝑎𝑎 is the forming yield of
aluminum type l.
The next step depends on which recycling model is chosen for aluminum production scrap. If
closed-loop recycling is selected, the values for collected production scrap, 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡), are for-
warded to the ‘Material production’ spreadsheet and used to calculate the recycled content of the
aluminum body and closure parts in North American light vehicle production (see Section 3.3).
If open-loop recycling is selected, the first step is to calculate the scrap input into vehicle produc-
tion according to this equation:
𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑎𝑎 (𝑡𝑡) = �𝑠𝑠𝑝𝑝𝑎𝑎 ∙ �1 − 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡)� + 𝑠𝑠𝑟𝑟𝑎𝑎 ∙ 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡)� ∙
where 𝑠𝑠𝑝𝑝𝑎𝑎 the scrap input into primary aluminum production (in kg/kg), 𝑠𝑠𝑟𝑟𝑎𝑎 the scrap input into
secondary (recycled) aluminum production (in kg/kg), and 𝑅𝑅𝑅𝑅𝑙𝑙𝑎𝑎 (𝑡𝑡) is the exogenously given re-
cycled (secondary) content of aluminum type l used in body and closure parts in production year
t. The second step is to calculate the net change in external secondary aluminum production
caused by the net scrap flow into or out of vehicle production according to the following equa-
tion:
22
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡) − 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑎𝑎 (𝑡𝑡)
𝛼𝛼𝑎𝑎 ∙
𝑠𝑠𝑟𝑟𝑎𝑎 − 𝛽𝛽𝑎𝑎 ∙ 𝑠𝑠𝑝𝑝𝑎𝑎
where 𝛼𝛼𝑎𝑎 models the response of the aluminum scrap market to a change in automotive alumi-
num scrap generation and use, and 𝛽𝛽𝑎𝑎 models the response of external primary aluminum pro-
duction to a change in external secondary aluminum production. A detailed explanation of those
two parameters and the equation above can be found in Section 4. The penultimate step of the
open-loop recycling model is to calculate the GHG implications of the net change in external
secondary aluminum production:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑎𝑎 (𝑡𝑡) − 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑎𝑎 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑎𝑎 (𝑡𝑡) = 𝛼𝛼𝑎𝑎 ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑎𝑎 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
In the final step, the total external GHG implications due to changes in aluminum scrap use and
generation during material production and forming are calculated by summing over aluminum
sheet, extrusions, and castings:
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) = � 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑎𝑎 (𝑡𝑡)

𝑙𝑙
For steel, the calculations are identical to the open-loop recycling calculations for aluminum.
Here, the starting point is the amount of steel removed from the vehicles, 𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡), which has
been calculated in the ‘Material production’ spreadsheet. These values are converted into
amounts of production scrap no longer generated and collected, 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑠𝑠 (𝑡𝑡), and amounts of steel
scrap no longer used into vehicle production, 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠 (𝑡𝑡):
(1 − 𝛾𝛾𝑙𝑙𝑠𝑠 )
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑠𝑠 (𝑡𝑡) = 𝑠𝑠
𝑐𝑐𝑝𝑝𝑝𝑝𝑝𝑝 ∙ ∙ 𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡)
𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑡𝑡)
𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠 (𝑡𝑡) = �𝑠𝑠𝑝𝑝𝑠𝑠 ∙ (1 − 𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 ) + 𝑠𝑠𝑟𝑟𝑠𝑠 ∙ 𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 � ∙
𝑠𝑠
where 𝑐𝑐𝑝𝑝𝑝𝑝𝑝𝑝 is the collection rate for steel production scrap, 𝛾𝛾𝑙𝑙𝑠𝑠 is the forming yield of steel type l,
𝑅𝑅𝑅𝑅𝑙𝑙𝑠𝑠 is the secondary content of steel type l, 𝑠𝑠𝑝𝑝𝑠𝑠 is the scrap input into primary steel production (in
kg/kg), and 𝑠𝑠𝑟𝑟𝑠𝑠 is the scrap input into secondary steel production (in kg/kg). The net change in
external secondary steel production is now calculated as follows:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑠𝑠 (𝑡𝑡) − 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠 (𝑡𝑡)
𝛼𝛼𝑠𝑠 ∙
𝑠𝑠𝑟𝑟𝑠𝑠 − 𝛽𝛽𝑠𝑠 ∙ 𝑠𝑠𝑝𝑝𝑠𝑠
where 𝛼𝛼𝑠𝑠 models the response of the steel scrap market to a change in automotive steel scrap
generation and use, and 𝛽𝛽𝑠𝑠 models the response of external primary steel production to a change
23
in external secondary steel production. As in the case of aluminum production scrap, the last two
steps are to calculate and aggregate the GHG implications of the net change in external second-
ary steel production:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑠𝑠 (𝑡𝑡) − 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑠𝑠 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠 (𝑡𝑡) = 𝛼𝛼𝑠𝑠 ∙ 𝑠𝑠 𝑠𝑠 �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑠𝑠 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
𝑠𝑠𝑟𝑟 − 𝛽𝛽𝑠𝑠 ∙ 𝑠𝑠𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) = � 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠 (𝑡𝑡)

𝑙𝑙

𝑎𝑎 (𝑡𝑡) GHG changes due to changes in external use R6:R44
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺
of aluminum production scrap in calendar year
t (in million kg CO2eq)
𝑠𝑠 (𝑡𝑡) GHG changes due to changes in external use R51:R89
of steel production scrap in calendar year t (in
million kg CO2eq)
Table 4: Output data from spreadsheet ‘Scrap at production”
3.5 Scrap at end-of-life

Changing the material composition of the vehicles produced between 2012 and 2050, means that
the quantities of end-of-life aluminum and steel scrap change when these vehicles reach the end
of their lives. The aim of this spreadsheet is to calculate the GHG implications of these changes
in automotive end-of-life scrap generation, collection, and recycling.
The main data input into this spreadsheet is the total amount of aluminum body and closure parts
added to the vehicles produced in year T, 𝑇𝑇𝑇𝑇(𝑇𝑇) = ∑𝑙𝑙 𝑀𝑀𝑀𝑀𝑙𝑙𝑎𝑎 (𝑇𝑇), and the amount of steel removed
from all vehicles produced in year T, 𝑇𝑇𝑇𝑇(𝑇𝑇) = ∑𝑙𝑙 𝑀𝑀𝑀𝑀𝑙𝑙𝑠𝑠 (𝑇𝑇). The first step is to calculate the
changes in end-of-life scrap generated and collected in calendar year t, taking into account the
lifetime distribution of the vehicles. For aluminum and steel the calculations are identical and as
follows:
𝑡𝑡−1
𝑎𝑎 𝑎𝑎
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑎𝑎 (𝑡𝑡) = 𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 ∙ 𝑠𝑠𝑠𝑠𝑟𝑟𝑒𝑒𝑒𝑒𝑒𝑒 ∙� 𝑇𝑇𝑇𝑇(𝑇𝑇) ∙ �𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡 − 1) − 𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡)�
𝑇𝑇=2012
𝑡𝑡−1
𝑠𝑠 𝑠𝑠
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑠𝑠 (𝑡𝑡) = 𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 ∙ 𝑠𝑠𝑠𝑠𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒 ∙� 𝑇𝑇𝑇𝑇(𝑇𝑇) ∙ �𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡 − 1) − 𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡)�
𝑇𝑇=2012
24
𝑎𝑎 𝑠𝑠 𝑎𝑎
where 𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 and 𝑐𝑐𝑒𝑒𝑒𝑒𝑒𝑒 are the end-of-life collection rates for aluminum and steel scrap, 𝑠𝑠𝑠𝑠𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒 and
𝑠𝑠
𝑠𝑠𝑠𝑠𝑠𝑠𝑒𝑒𝑒𝑒𝑒𝑒 are the shredder separation rates for end-of-life aluminum and steel scrap. 𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡) is
the fraction of vehicles still in use after 𝑇𝑇 − 𝑡𝑡 years of driving, and 𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡 − 1) is the frac-
tion of vehicles still in use after 𝑇𝑇 − 𝑡𝑡 − 1 years of driving (see also Section 3.2). Therefore,
�𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡 − 1) − 𝐹𝐹𝐹𝐹𝐹𝐹(𝑇𝑇 − 𝑡𝑡)� is the fraction of vehicles that reach end of life during calendar
year 𝑇𝑇 − 𝑡𝑡.
If closed-loop recycling is selected for the end-of-life aluminum scrap, 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑎𝑎 (𝑡𝑡) is broken out
into sheet, extrusions, and castings, 𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝑙𝑙𝑎𝑎 (𝑡𝑡), forwarded to the ‘Material production’ spread-
sheet, and used to calculate the recycled content of the aluminum body and closure parts in North
American light vehicle production (see Section 3.3).
If open-loop recycling is selected for end-of-life aluminum scrap, 𝐸𝐸𝐸𝐸𝐸𝐸𝑆𝑆 𝑎𝑎 (𝑡𝑡) is recycled external-
ly and has the following GHG implications:
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑎𝑎 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) = 𝛼𝛼𝑎𝑎 ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑎𝑎 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
where 𝛼𝛼𝑎𝑎 models the response of the aluminum scrap market to a change in automotive alumi-
num scrap generation and use, and 𝛽𝛽𝑎𝑎 models the response of external primary aluminum pro-
duction to a change in external secondary aluminum production. The open-loop model used to
calculate the GHG consequences of changes in end-of-life scrap flows is the same as the one
used for changes in prompt scrap flows. A detailed explanation of the two central model parame-
ters, 𝛼𝛼 and 𝛽𝛽, can be found in Section 4. Other parameters used in the calculation above are the
GHG intensities of primary and secondary aluminum ingot production (in kgCO2eq/kg),
𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 , and the scrap input into primary and secondary (recycled) aluminum
production (in kg/kg), 𝑠𝑠𝑝𝑝𝑎𝑎 and 𝑠𝑠𝑟𝑟𝑎𝑎 .
The changes in end-of-life steel scrap are always modeled in an open loop. The GHG implica-
tions of reducing the amount of end end-of-life steel scrap are calculated as follows:
𝑠𝑠 (𝑡𝑡)
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑠𝑠 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 = 𝛼𝛼𝑠𝑠 ∙ 𝑠𝑠 𝑠𝑠 �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑠𝑠 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
𝑠𝑠𝑟𝑟 − 𝛽𝛽𝑠𝑠 ∙ 𝑠𝑠𝑝𝑝
where 𝛼𝛼𝑠𝑠 models the response of the steel scrap market to a change in automotive steel scrap
generation and use, and 𝛽𝛽𝑠𝑠 models the response of external primary steel production to a change
25
in external secondary steel production. Other parameters used in the calculation above are the
𝑝𝑝𝑝𝑝
GHG intensities of primary and secondary steel ingot production (in kgCO2eq/kg), 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 and
𝑠𝑠𝑠𝑠
𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 , and the scrap input into primary and secondary (recycled) steel production (in kg/kg),
𝑠𝑠𝑝𝑝𝑎𝑎 and 𝑠𝑠𝑟𝑟𝑎𝑎 .

𝑎𝑎 (𝑡𝑡) GHG changes due to changes in external use O5:O43
of end-of-life aluminum scrap in calendar year
t (in million kg CO2eq)
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) GHG changes due to changes in external use M5:M43
of end-of-life steel scrap in calendar year t (in
million kg CO2eq)
Table 5: Output data from spreadsheet ‘Scrap at end-of-life”
26
4 Appendix A: Consequential system expansion
The reference flow of this CLCA is an increasing replacement of steel with aluminum in body
and closure parts used in North American light vehicle production between 2012 and 2050. Such
a change in automotive material composition causes changes in scrap flows into and out of the
vehicle life cycles. If automotive scrap recycling is modeled in a closed loop, the changes in
scrap input to automotive material production are equal to the changes in automotive production
or end-of-life scrap generation, and there is thus no need to expand the system boundaries be-
yond the modeled vehicle life cycles. If production or end-of-life scrap recycling are modeled in
an open loop, however, system expansion is required to describe the consequences of these scrap
flow changes.
Specifically, the system boundaries are expanded to include the impact of automotive scrap flow
changes on the external scrap market and the impact of the external scrap market changes on ex-
ternal secondary and primary steel and aluminum production. External here means scrap recy-
cling and metal production outside of the modeled vehicle life cycles. Flows of automotive scrap
out of or into the scrap market change at two separate times in the life cycles of the modeled ve-
hicles; during vehicle production and during vehicle end-of-life management (scrapping). Figure
8 shows the system expansion during vehicle production. Note that scrap flow changes can be
positive or negative. The arrows in Figure 8 indicate positive flow changes.
27
Figure 8: Consequential system expansion for scrap input to and output from automotive materi-
al production and forming
In the open loop recycling scenario, changes in automotive material composition cause changes
in scrap input to material production, 𝑆𝑆𝑆𝑆𝑆𝑆, and scrap output from material forming, 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃. Since
it is assumed that scrap stocks are constant in the long run, the scrap market can respond to the
net change in scrap flow, 𝑆𝑆𝑆𝑆𝑆𝑆 − 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃, by changing external scrap collection or changing exter-
nal scrap consumption, i.e. recycling. The parameter 𝛼𝛼 ∈ [0; 1] quantifies the impact the net
change 𝑆𝑆𝑆𝑆𝑆𝑆 − 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 has on external scrap collection. For example, an increase in scrap supply by
one unit increases external scrap collection by (𝛼𝛼 − 1) units.
The impact of 𝑆𝑆𝑆𝑆𝑆𝑆 − 𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 on external recycling is slightly more complicated if primary material
production also consumes production or end-of-life scrap, as is the case with steel. This creates
an additional scrap flow, which creates a feedback loop between the material market and the
scrap market as shown in Figure 8 and Figure 9. In general, a change in external recycling im-
pacts the production of the equivalent primary material since they compete with each other on
the material market. The response of primary production to changes in external recycling (sec-
ondary production) is quantified by the parameter 𝛽𝛽 ∈ [0; 1]. For example, an increase in exter-
nal secondary production by one unit decreases external primary production by 𝛽𝛽 units.
28
The assumption that all scrap flow changes balance, i.e. stocks are constant in the long run,
means that the interim parameter 𝑌𝑌 can be calculated by balancing all scrap flows, as shown in
the equation below:
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 + (𝛼𝛼 − 1)(𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑆𝑆𝑆𝑆𝑆𝑆) + 𝑠𝑠𝑝𝑝 𝛽𝛽𝛽𝛽(𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑆𝑆𝑆𝑆𝑆𝑆) = 𝑆𝑆𝑆𝑆𝑆𝑆 + 𝑠𝑠𝑟𝑟 𝑌𝑌(𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃 − 𝑆𝑆𝑆𝑆𝑆𝑆)
𝛼𝛼
⇒ 𝑌𝑌 =
𝑠𝑠𝑟𝑟 −𝑠𝑠𝑝𝑝 𝛽𝛽
where 𝑠𝑠𝑝𝑝 is the scrap input into primary production (in kg/kg) and 𝑠𝑠𝑟𝑟 the scrap input into second-
ary (recycled) production (in kg/kg). In the case of aluminum, 𝑠𝑠𝑝𝑝𝑎𝑎 = 0, which simplifies the sys-
tem expansion. Figure 9 shows the consequential system expansion model during vehicle end-of-
life management, i.e. scrapping. The modeling approach is identical, and the scrap balance equa-
tion therefore yields the identical result for the interim parameter 𝑌𝑌:
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 + (𝛼𝛼 − 1)𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 + 𝑠𝑠𝑝𝑝 𝛽𝛽𝛽𝛽𝛽𝛽𝛽𝛽𝛽𝛽𝛽𝛽 = 𝑠𝑠𝑟𝑟 𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌𝑌
𝛼𝛼
⇒ 𝑌𝑌 =
𝑠𝑠𝑟𝑟 −𝑠𝑠𝑝𝑝 𝛽𝛽
Figure 9: Consequential system expansion for scrap output from vehicle end-of-life management
The final step in the system expansion is to calculate the changes in GHG emissions that are
caused by the changes in external secondary and primary production. For the changes in scrap
input to and output from automotive material production and forming they are calculated as fol-
lows:
29
𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑃𝑙𝑙𝑥𝑥 (𝑡𝑡) − 𝑆𝑆𝑆𝑆𝑆𝑆𝑙𝑙𝑥𝑥 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑙𝑙𝑠𝑠𝑠𝑠 (𝑡𝑡) = 𝛼𝛼𝑠𝑠 ∙ �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑥𝑥 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
𝑠𝑠𝑟𝑟𝑥𝑥 − 𝛽𝛽𝑥𝑥 ∙ 𝑠𝑠𝑝𝑝𝑥𝑥
𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
where x indicates the material and l the material type. 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 and 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 are the GHG in-
tensities of primary and secondary ingot production. The equation above is used on the spread-
sheet ‘Scrap at production’ if open loop recycling is chosen for production scrap (see Section
3.4). For the changes in end-of-life scrap generation the GHG implications are calculated as fol-
lows:
𝐸𝐸𝐸𝐸𝐸𝐸𝐸𝐸 𝑥𝑥 (𝑡𝑡) 𝑠𝑠𝑠𝑠 𝑝𝑝𝑝𝑝
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑥𝑥 (𝑡𝑡) = 𝛼𝛼𝑥𝑥 ∙ 𝑥𝑥 𝑥𝑥 �𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 − 𝛽𝛽𝑥𝑥 ∙ 𝐺𝐺𝐺𝐺𝐺𝐺𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖𝑖 (𝑡𝑡)�
𝑠𝑠𝑟𝑟 − 𝛽𝛽𝑥𝑥 ∙ 𝑠𝑠𝑝𝑝
The equation above is used on the spreadsheet ‘Scrap at end-of-life’ if open loop recycling is
chosen for end-of-life scrap (see Section 3.5). If 𝛼𝛼 is set to zero, the two expressions above are
also zero. This is the CLCA equivalent to the recycled content approach in ALCA, which ignores
scrap inputs to and outputs from the studied product system. If 𝛼𝛼 and 𝛽𝛽 are both set to one, the
two expressions above are identical to the avoided burden approach in ALCA, which also goes
under the names of ‘closed loop approximation method’, ‘0/100 method’, ‘end-of-life approach’,
‘substitution method’, and occasionally even simply ‘system expansion’. Using the parameters 𝛼𝛼
and 𝛽𝛽 allows the user of the CLCA model to study the impact of different scrap and material
market dynamics in great detail. Unfortunately, there is currently no research available that al-
lows for a scientifically sound selection of parameter values.
30
5 Appendix B: Default input data
This section contains the default input data. While the default input data has been carefully se-
lected, its main purpose is to illustrate how the model works rather than to generate definitive
results. Model users are highly encouraged to use their own input data and also study the sensi-
tivity of the model results to changes in input data values. Firm conclusions about the GHG im-
pacts of a wide-spread diffusion of aluminum into North American light duty vehicle production
can only be done in studies where the representativeness of the default data and parameters are
well established and adequate range and sensitivity analyses are conducted.
5.1 Results & data input spreadsheet

Table 1: General input data
Parameter Value Source
Material replacement coefficient (aluminum to steel) 0.65 SMDI 2016
Flat steel fraction of total replaced steel 0.9 Geyer 2013
Secondary mass savings coefficient 0.2 SMDI 2016
Flat share of secondary mass savings 0.4 Geyer 2013
Long share of secondary mass savings 0.3 Geyer 2013
Table 2: Aluminum recycling parameters

Scrap input to primary production 0 IAI 2007
Scrap input to secondary production 1.048 SMDI 2016
Prompt scrap collection rate 0.99 Geyer 2013
EOL scrap collection rate 0.97 Geyer 2013
Shredder separation rate 0.9 Geyer 2013
Alpha 0.9 Geyer 2013, 2015
Beta 1 Geyer 2013, 2015
Table 3: Steel recycling parameters

Scrap input to primary production 0.209 SMDI 2016
Scrap input to secondary production 1.05 WSA 2010
Prompt scrap collection rate 0.99 Geyer 2013
EOL scrap collection rate 0.97 Geyer 2013
Shredder separation rate 0.98 Geyer 2013
Alpha 0.9 Geyer 2013, 2015
Beta 1 Geyer 2013, 2015
31
Table 4: GHG intensities of aluminum production and forming (in kgCO2eq/kg output)
Process GHG intensity Source
Primary ingot (North America) 8.937 TAA 2013
(cradle-to-gate)
Secondary ingot 1.23 SMDI 2016
(cradle-to-gate)
Rolled aluminum 0.589 thinkstep 2015 (PE, EU 27)
(ingot-to-gate, aluminum rolling)
Extruded aluminum 0.689 thinkstep 2015 (PE, EU 27)
(ingot-to-gate, aluminum extrusion)
Cast aluminum 0.590 thinkstep 2015 (PE, DE)
(ingot-to-gate, aluminum casting)
Table 5: GHG intensities of steel production and forming (in kgCO2eq/kg output)
Process GHG intensity Source
BF/BOF slab 1.870 WSA 2010
EAF slab (slab-to-gate) 0.399 WSA 2010
Flat steel (slab-to-gate) 0.485 WSA 2010
Long steel (gate-to-gate, steel rolling) 0.290 WSA 2010
Cast steel (gate-to-gate, steel casting) 0.135 WSA 2010
Table 6: Vehicle use phase parameters

Parameter Value Unit Source
NCV of gasoline 32.27 MJ/liter thinkstep 2015
(PE, EU27)
GHG intensity of gasoline production 15.50 gCO2eq/MJ thinkstep 2015
(cradle-to-gate, at gas station)) (PE, EU27)
GHG intensity of gasoline combustion 72 gCO2eq/MJ thinkstep 2015
(gate-to-gate, combustion in light duty (PE, GLO)
vehicle)
NCV of diesel 36.00 MJ/liter thinkstep 2015
(PE, EU27)
GHG intensity of diesel production 7.74 gCO2eq/MJ thinkstep 2015
(cradle-to-gate, at gas station)) (PE, EU27)
GHG intensity of diesel combustion 75 gCO2eq/MJ thinkstep 2015
(gate-to-gate, combustion in light duty (PE, GLO)
vehicle)
GHG intensity of electricity production 0.278 kgCO2eq/MJ thinkstep 2015
(cradle-to-gate, U.S. average, at consum- (PE, US)
er)
Share of plug electricity as energy source 0.5 % Geyer 2013
Vehicle lifetime driving 245,000 km SMDI 2016
Mean vehicle lifetime 12 years SMDI 2016
Standard deviation of vehicle lifetime 2 years Estimate
32
Table 7: Forming yields
Material Yield Source
Aluminum, Sheet 0.52 Geyer 2014
Aluminum, Extrusion 0.8 Geyer 2013
Aluminum, Castings 0.8 Geyer 2013
Steel, Flat 0.55 Geyer 2014
Steel, Long 0.75 Geyer 2013
Steel, Castings 0.8 Geyer 2013
Table 8: Primary aluminum ingot production GHGs

(Sources: Hao 2015, Thinkstep 2015, IAI 2014, SMDI 2016)
Imported Al
Year % Imported
kgCO2/kg
2012 16.59 0.20
2013 16.45 0.24
2014 16.30 0.28
2015 16.15 0.32
2016 16.01 0.36
2017 15.86 0.40
2018 15.72 0.44
2019 15.58 0.48
2020 15.44 0.52
2021 15.30 0.56
2022 15.17 0.60
2023 15.03 0.64
2024 14.90 0.68
2025 14.76 0.72
2026-2050 14.76 0.72
33
5.2 Fleet Composition
Table 9: Light duty vehicle production forecast (Source: Ducker 2015 p. 55)
Year # of Vehicles
2012 16,181,282
2013 16,200,000
2014 16,752,614
2015 17,500,000
2016 17,323,946
2017 17,609,612
2018 17,900,000
2019 18,180,944
2020 18,200,000
2021 18,752,277
2022 19,000,000
2023 19,323,609
2024 19,609,275
2025-2050 20,000,000
The resulting cumulative light duty vehicle production between 2012 and 2015 is therefore
752,533,558, i.e. just over 750 million cars.
Table 10: Powertrain type inputs (Source: Ducker 2015 p. 21 & 22)
Year ICV Gasoline Hybrid Standard
Share Share of Share Share of Hy-
ICV brid
2012 0.982 0.932 0.018 0.500
2013 0.976 0.931 0.024 0.500
2014 0.972 0.935 0.026 0.500
2015 0.965 0.936 0.034 0.500
2016 0.952 0.932 0.046 0.500
2017 0.947 0.931 0.050 0.500
2018 0.938 0.932 0.057 0.500
2019 0.924 0.931 0.070 0.500
2020 0.913 0.934 0.081 0.500
2021 0.909 0.932 0.084 0.500
2022 0.900 0.931 0.092 0.500
2023 0.892 0.932 0.100 0.500
2024 0.883 0.931 0.108 0.500
2025 0.874 0.934 0.110 0.500
2026-2050 0.870 0.934 0.110 0.500
34
Table 11: Projected amount of aluminum body & closure parts in North American light vehicle
production (Source: Ducker 2014 p. 20)
Year Sheet (million kg) Extrusions (million kg) Castings (million kg)
2012 81.9 0.0 0.0
2013 100.3 7.3 0.0
2014 200.5 12.7 1.8
2015 420.9 20.0 18.2
2016 592.2 32.8 23.7
2017 681.6 38.2 27.3
2018 738.1 41.9 29.1
2019 878.4 56.4 29.1
2020 1,078.8 69.2 45.5
2021 1,228.2 80.1 45.5
2022 1,303.0 81.9 65.5
2023 1,419.7 87.4 74.6
2024 1,640.1 111.0 78.3
2025 1,707.6 112.9 85.6
2026 1,755 114 91
2027 1,778 115 93
2028-2050 1,790 115 95
Table 12: Vehicle class data

Class lbs of aluminum NA LDV fleet share HEV share BEV share
Source Ducker 2014 p. 9 ANL 2014
A/B 251.60 0.03 0.09 0.18
C 273.90 0.17 0.35 0.51
D 363.30 0.21 0.42 0.31
E 546.90 0.03 0.11 0.00
MPV 396.50 0.04 0.00 0.00
SUV 410.30 0.33 0.02 0.01
VAN 273.20 0.02 0.00 0.00
PUP 548.90 0.17 0.00 0.00
35
5.3 Vehicle Use
Table 13: Energy and Electricity savings per mass savings with and without powertrain resizing
(Source: FKA 2011)
No powertrain resizing Powertrain Resizing
(MJ/100km100kg) (MJ/100km100kg)
PT type Fuel Compact Midsize SUV Compact Midsize SUV
ICEV-G Gasoline 3.586 2.981 3.406 8.034 10.374 9.331
ICEV-D Diesel 3.405 2.739 3.415 6.421 8.584 7.260
HEV-S Gasoline 2.994 3.329 2.927 4.998 6.324 6.409
HEV-P Gasoline 3.856 4.146 n/a 4.291 4.964 n/a
HEV-P Electricity 1.434 1.349 n/a 1.516 1.490 n/a
BEV Electricity 1.381 1.515 n/a 1.407 1.495 n/a
Note: Values could be varied over time, i.e. entered as time series
36
6 Appendix C: Default results and interpretation
While the default input data has been carefully selected, its main purpose is to illustrate how the
model works rather than to generate definitive results. Model results are very sensitive to some
input parameters and less so to others. Sensitivity analysis of the model is outside of the scope of
this study, which is exclusively about computational structure and model methodology. Firm
conclusions about the GHG impacts of a wide-spread diffusion of aluminum into North Ameri-
can light duty vehicle production can only be done in studies where the representativeness of the
default data and parameters are well established and adequate range and sensitivity analyses are
done.
The main objective of this report section is to explain how the results are displayed on the
spreadsheet ‘Results and data input’ and should thus be interpreted. The example chosen for this
task are the results generated by the default input data. As explained in Section 1.6, the focus of
the CLCA model is to generate plausible scenarios rather than one authoritative solution of the
research question. This focus reflects the significant uncertainties inherent in consequential anal-
ysis in general and in particular the one reported here. For every year of the study period (2012-
2050) the seven GHG emission changes introduced in Section 2 and explained in Section 3 are
reported:
Variable name GHG emission change due to change in Location

𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑎𝑎 (𝑡𝑡) automotive aluminum production and forming B5-B43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) automotive steel production and forming C5-C43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) fuel combustion and fuel and drive electricity production D5-D43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) aluminum production scrap generation and open-loop recycling E5-E43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) steel production scrap generation and open-loop recycling F5-F43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑎𝑎 (𝑡𝑡) aluminum end-of-life scrap generation and open-loop recycling G5-G43
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑠𝑠 (𝑡𝑡) steel scrap end-of-life generation and open-loop recycling H5-H43
Table 6: For every year of the study period (2012-2050) these seven GHG changes are reported
on the ‘Results & data input’ spreadsheet
37
A positive GHG emission value means that GHG emissions increase due to the modeled chang-
es, a negative value indicates a GHG emission decrease. In addition to the results explained in
Table 6, the annual net GHG emission changes are calculated as
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) = ∑𝑚𝑚=𝑎𝑎,𝑠𝑠 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝑚𝑚 (𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑚𝑚 (𝑡𝑡) + 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺 𝑚𝑚 (𝑡𝑡) and displayed
in cells J5-J43. Finally, for each year the cumulative net GHG emission changes, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ) =
∑𝜏𝜏𝑡𝑡=2012 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡), are calculated by summing over all previous calendar years and displayed in
cells K5-K43.
All time series of all results are also visualized in two figures, which are shown for default data
input results on the next page. Figure 10 shows the annual and cumulative GHG emission chang-
es, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) and 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ). Figure 11 shows the individual annual GHG emission changes of
material production, vehicle use, and scrap generation and recycling. The X-axis in Figures 10
and 11 can be interpreted as a “base case” signifying no shift of closure and body part materials
from steel to aluminum. This case would result in no GHG emission changes.
Figure 10: Annual and cumulative GHG emission changes (in million kg CO2eq), 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) and
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ), for the default input data
The cumulative GHG emission changes 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ) provide the highest level overview of the re-
sults. Figure 10 shows a steady increase in GHG emissions until they reach a peak of 161 million
38
metric tons of CO2eq in the year 2033 and a steady decrease thereafter. In 2050, the cumulative
GHG emission increase is still 81 million metric tons of CO2eq, but the annual GHG emission
change 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(𝑡𝑡) has reached a negative steady state of 5.8 million metric tons of CO2eq/year,
which means that 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ) will eventually reach zero and go negative. The year in which
𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ) crosses the X-axis is calculated in cell I46. For the default data input results the year is
2064.
Figure 11: Annual GHG emission changes (in million kg CO2eq) due to changes in material pro-
duction and forming, vehicle use, and production (prompt) and end-of-life (eol) scrap generation
and use, for the default input data
Figure 11 shows how the overall result from Figure 10 comes about. It shows that the growth in
GHG emissions until 2033 is caused by automotive aluminum production, which significantly
outweighs the simultaneous GHG emission reductions from production aluminum scrap recy-
cling and the avoided steel production. Figure 11 also shows that the GHG emission reductions
from vehicle fuel economy improvements and end-of-life scrap generation and recycling take
39
place with a significant time delay. Only when the annual GHG emission reductions from end-
of-life aluminum scrap recycling exceed 12 million metric tons of CO2eq/year in 2034 does the
annual net change become negative. Relatively soon thereafter, the system reaches a steady state
in which the GHG increases of automotive aluminum production and forming are outweighed by
the combined effect of avoided steel production, fuel economy improvements, and production
and end-of-life aluminum scrap recycling. Figure 11 shows that those four effects have the same
order of magnitude (12-18 million metric tons of CO2eq/year) and that the effects of production
and end-of-life aluminum scrap recycling are somewhat larger than that of fuel economy im-
provements. The effect of avoided steel production is the smallest of the four.
The large GHG impact of aluminum scrap generation and use explains why the model results are
so sensitive to the aluminum recycling parameters of the CLCA model, in particular 𝛼𝛼 and 𝛽𝛽 of
the consequential system expansion (see Section 4). In the default setting, aluminum scrap recy-
cling is modeled as an open loop with 𝛼𝛼 = 0.9 and 𝛽𝛽 = 1. Choosing closed-loop recycling for
production scrap reduces the crossover year to 2053, choosing it for production and end-of-life
scrap to 2048. However, if aluminum scrap recycling is modeled as an open loop with parameter
settings 𝛼𝛼 = 0.8 and 𝛽𝛽 = 0.9, 𝐺𝐺𝐺𝐺𝐺𝐺𝐺𝐺(τ) would reach 265 million metric tons of CO2eq by 2050
and continue to increase in perpetuity.
40
7 References
ANL (2014) Light Duty Electric Drive Vehicles Monthly Sales Updates. Argonne National
Laboratory (ANL), Published online 1 Oct. 2015. Accessed 8 Oct. 2015, available at
http://www.anl.gov/energy-systems/project/light-duty-electric-drive-vehicles-monthly-
sales-updates
Ducker (2014) 2015 North American light vehicle aluminum content study, Executive Summary,
June 2014, Ducker Worldwide, Troy, MI, on behalf of Drive Aluminum/The Aluminum
Association, Arlington, VA. Available at http://www.drivealuminum.org/research-
resources/PDF/Research/2014/2014-ducker-report
Ducker (2015) AISI Materials Content Analysis, Final Report, January 2015, Ducker
Worldwide, Troy, MI, on behalf of Steel Market Development Institute, Pittsburgh, PA.
EAA (2000) Environmental Profile Report for the European Aluminium Industry, European Al-
uminium Association, Brussels, Belgium.
FKA (2011) Determination of weight influence on the energy consumption according to US

regulations, Report 105620, April 2011, Forschungsgesellschaft Kraftfahrwesen (FKA),
Aachen, Germany, on behalf of WorldAutoSteel, Middletown, OH.
Geyer R (2008) Parametric Assessment of Climate Change Impacts of Automotive Material Sub-
stitution, Environmental Science and Technology, 42(18) 6973-6979.
Geyer R (2013) User Guide for Version 4 of the WorldAutoSteel Energy and GHG Model,
September 2013, on behalf of the World Steel Association, Brussels, Belgium.
Geyer R (2014) 14 April 2014 Data Update of Version 4 of the WorldAutoSteel Energy and
GHG Model, September 2013, on behalf of the World Steel Association, Brussels,
Belgium.
Geyer R (2015) Evaluation of life cycle assessment (LCA) end-of-life recycling methodologies
for steel products produced in North America, March 2015, on behalf of the Steel Market
Development Institute (SMDI), Pittsburgh, PA.
41
Hao H, Geng Y, Hang W (2015) GHG emissions from primary aluminum production in China:
Regional disparity and policy implications, Applied Energy, published online, available
at http://dx.doi.org/10.1016/j.apenergy.2015.05.056
IAI (2007) Life Cycle Assessment of Aluminium: Inventory Data For The Primary Aluminium
Industry, Year 2005 Update, International Aluminium Institute (IAI), London, UK.
IAI (2014) Primary aluminium production statistics, International Aluminium Institute (IAI),
London, UK. Accessed on 4/23/2014 at http://www.world-aluminium.org/statistics/.
SMDI (2016) The Importance of the Production Phase in Vehicle Life Cycle GHG Emissions,
Steel Market Development Institute (SMDI) Technical White Paper, May 2016, available
online at http://www.autosteel.org.
TAA (2013) The Environmental Footprint of Semi-Finished Aluminum Products in North Amer-
ica, December 2013, The Aluminum Association (TAA), Arlington, VA.
thinkstep (2015) GaBi Professional Database, DB Version 6.115, Service Pack 28.
WSA (2010) Data provided by World Steel Association (WSA), Brussels, Belgium.
42
35 Bracebridge Road
Newton, Massachusetts 02459
Critical Review by Independent Third Party
For this study, the Critical Review Panel was charged with providing comments on the consequential life
cycle assessment (CLCA) of replacing steel with aluminum in North American vehicles, as well as a
second round of comments to determine whether the first round of comments was adequately
addressed by AISI and Dr. Roland Geyer, Bren School of Environmental Science and Management,
University of California, Santa Barbara.
The scope of the CLCA includes the greenhouse gas (GHG) consequences of replacing conventional steel
with aluminum in North American light duty vehicles for vehicle mass reduction. The principal
investigators prepared an Excel-based model for this purpose that contained the following GHG
consequences:
• Changes in material production emissions due to primary and secondary mass reductions
• Changes in GHG emissions due to changes in prompt scrap recycling
• Changes in vehicle use phase emissions due to vehicle mass reduction
• Changes in GHG emissions due to changes in end-of-life (EOL) scrap recycling
In accordance with ISO/TS 14071:2014, an expert panel was assembled that included proficiency in the
applicable ISO standards; the LCA methodology, both attribution and consequential techniques; the
critical review practice; scientific disciplines relevant to the impact categories used in the study;
automotive, steel, and aluminum industry expertise; and the language used in the study.
The panel of reviewers of the study included:
Thomas Gloria, PhD (Chair)

Program Director, Sustainability, Harvard University
Managing Director, Industrial Ecology Consultants
Newton, Massachusetts, US
Alissa Kendall, PhD

Associate Professor, Civil and Environmental Engineering
University of California Davis
Davis, California, US
John Sullivan, PhD

Research Associate II, University of Michigan
Principal, EcoSpherica LLC
Ann Arbor, Michigan, US
Review Process
The scope of this ISO 14040:2006 and 14044: 2006 critical review was limited to the methodology,
comprehensiveness, and functionality of the study. The default input parameters built into the model by
the principal investigators were outside the scope of this review. While it is the ultimate intent of AISI to
select credible input parameters, develop a baseline scenario, conduct sensitivity assessments, and
publish the findings, this will be handled by a separate, subsequent review process and was outside of
the scope of the review.
35 Bracebridge Road
Newton, Massachusetts 02459
Review Results
Based on the objectives set forth to review this study, the Critical Review Panel concludes that the study
conforms to the LCA requirements of ISO 14040:2006 and 14044: 2006 standards. As determined by the
second round of review, all responses by AISI and Dr. Roland Geyer to the first round of comments
submitted by the Critical Review Panel have been adequately addressed.
Respectfully,
Thomas P. Gloria, Ph.D.
22 November 2016
Newton, Massachusetts, US
APPENDIX B: CRITICAL REVIEW FOR THIS STUDY
35 Bracebridge Road
Neewton, Massachusettts 02459

Critical Reeview by Pan nel of Externaal Experts
The Critical Review Pan nel was chargged with revie ewing and co mmenting on n the consequ uential life cycle
assessment (CLCA) study titled “Con nsequential LLife Cycle Greeenhouse Gass Study of Au utomotive
Lightweigghting with Advanced High h Strength Stteel (AHSS) an nd Aluminum m”. The study was conductted
by the Steeel Recycling Institute, a bu usiness unit o
of the Americcan Iron and SSteel Institutee, using the
spreadshe eet model described in the e UC Santa Baarbara reportt Consequentiial Life Cycle G Greenhouse G Gas
Assessmeent of Replacin ng Steel with Aluminum in n North Ameriican Vehicle PProduction ‐ Methodologyy
Report de eveloped by D Dr. Roland Geyer, dated No ovember 17, 2016. The go oal of the stud
dy was to ideentify
and quanttify the main GHG emissio ons consequences of a signnificant increaase in the usee of aluminum m for
vehicle boody and closu ure parts, andd to compare these conseqquences to the GHG emissions
conseque ences of the uuse of AHSS fo or the same p parts, as part oof an overall vehicle fleet light weightin ng
strategy. AISI retained d this panel to
o conduct a rigorous assesssment of thee study scope,, assumptionss,
inputs, annd results. The e following is the final reviiew statemennt by the exteernal review p panel based o on
the July 30th, 2018 Fin nal Report verrsion.

Panel Me embers
Thhomas Gloriaa, Ph.D. (Pane el Chair), Industrial Ecologyy Consultantss
William O. Col
W linge, Ph.D., Independently contracted by AISI
Trroy A. Hottle,, Ph.D., Indeppendently con ntracted by A ISI

Critical Reeview Tasks & & Objectives
Reviewerss were asked to review ressponses to th he comments and accompaanying revisio ons of the LCA A
report until adequatelyy addressed b by AISI.

Per Internnational Organization of Sttandardization (ISO) 140444:2006(E) Envvironmental m management – Life
cycle asseessment – Req quirements an nd guideliness, the critical rreview processs included thhe following
objectivess to ensure co onformance w with applicable standards::
‐ The m methods used d to carry out the LCA were e consistent w with the appliicable internaational standaards
‐ The m methods used d to carry out the LCA were e scientificallyy and techniccally valid
‐ The data used werre appropriate e and reasonable in relatioon to the goaal of the studyy
‐ The in nterpretations reflected th he limitationss identified annd the goal off the study, and
‐ The sttudy report w was transpare ent and consisstent.

Review Results
The overaall review wass conducted in an equitablle and constr uctive manneer. All commeents were
addressed d, and all opeen issues resolved. There w were no disse nting opinion ns held by thee reviewers or the
commissio oner upon fin nalization of tthe review. P
Please note thhat the review wer panel’s sttatements of
conformaance are as individual expe erts.

Respectfu ully,

Thomas P P. Gloria, Critical Review Paanel Chair

15 Augustt 2018
Newton, M
Massachusettts

SRI SMDI Auto CLCA Study Final Report

Uploaded by

Copyright:

Available Formats

SRI SMDI Auto CLCA Study Final Report

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

SRI SMDI Auto CLCA Study Final Report

Uploaded by

Copyright:

Available Formats

TABLE OF CONTENTS

2. GOAL OF THE STUDY 4

3. SCOPE OF THE STUDY 5

4. LIFE CYCLE INVENTORY ANALYSIS 7

5. LIFE CYCLE IMPACT ASSESSMENT (LCIA) RESULTS 17

APPENDIX A – Consequential Life Cycle Assessment Model User Guide and

These consequences were assessed using a spreadsheet model (hereinafter referred to as

Four sets of consequences are modeled in the study:

Source: U.S. EPA, 2018. 3

Report. November 2016. pg. 47.

Lightweighting.” November 7, 2017. http://www.steelsustainability.org/automotive/auto-ghg/.

7(3):588-595, 2014, https://doi.org/10.4271/2014-01-1004.

vehicle mass reduction,” Transportation Research Part D 56 (2017) 241–257. http://dx.doi.org/10.1016/j.trd.2017.08.010.

Organization for Standardization. 2006.

3.1 Product System Description

3.2 Functional Unit

3.3 System Boundary

3.4 Cut-off Criteria

4.1 Data Collection and Sources

4.1.1 Material Inputs

Furthermore, according to the Aluminum Association’s energy production accounting

4.1.2 Material Replacement Coefficients

Available at: https://minerals.usgs.gov/minerals/pubs/commodity/aluminum/.

implications,” Applied Energy, 2015. Published online: http://dx.doi.org/10.1016/j.apenergy.2015.05.056.

http://www.aluminum.org/getting-trade-right. Accessed October 2017.

Assessment Report.” December 2013.

vehicle mass reduction,” Transportation Research Part D 56 (2017) 241–257. http://dx.doi.org/10.1016/j.trd.2017.08.010.

4.1.3 Secondary Mass Savings

Table 3. Composition of Secondary Mass Savings

4.1.4 Lifetime Driving Distance

for WorldAutoSteel. January 2013.

Vehicle Lightweighting.” Environmental Science & Technology. 2012, 46, 2893−2901.

4.1.5 Fuel Reduction Values and Power Train Resizing

4.1.6 Recycling Rates

Table 4. Recycling Rates for Aluminum and Steel

4.2 CLCA Model Calculation Procedures

4.4 Data Validation and Data Quality Assessment

4.4.1 Time-related Coverage

4.4.2 Geographical Coverage

4.4.3 Technology Coverage

Completeness addresses the percentage of flow that is measured or estimated. As described

Consistency entails a qualitative assessment of whether the study methodology is applied

4.4.9 Sources of Data

This criterion simply requires an assessment of data uncertainty. Data uncertainty is

4.5 Sensitivity Analysis

5.1 Base Case Results

Physical Science Basis.” 2013.

Figure S2: AHSS Greenhouse Gas Intensity Increased by 25%

Figure S6: Growth Rate of Imported Aluminum Varied from 0% to 6%

Figure S10: Mean Vehicle Lifetime Varied from 10 Years to 16 Years

6.1 Key Findings

6.2 Assumptions and Limitations

• The production of aluminum for North American automotive applications is assumed to

Prepared by Roland Geyer, Dipl. Phys., Ph.D.

University of California at Santa Barbara, CA, USA

On behalf of SMDI – Steel Market Development Institute

1.1 GENERAL ASPECTS ........................................................................................................................................1

2 OVERALL DESCRIPTION OF THE CLCA MODEL ..................................................................................7

3 DESCRIPTION OF THE INDIVIDUAL SPREADSHEETS ....................................................................... 13