A NEW PARADIGM FOR AUTOMOTIVE MASS BENCHMARKING

A NEW PARADIGM FOR AUTOMOTIVE MASS BENCHMARKING September 2015 WorldAuto 2015 1

Table of contents 1.0 Introduction 1 1.1 Study Scope 2 1.1.1 Mass Benchmarking Methodology 2 1.1.2 Vehicle Types Included 5 1.1.3 Cost References 6 1.1.4 Life Cycle Assessment 6 2.0 Automotive Mass Benchmarking Key Findings 7 2.1 Mass efficiency of today s Designs Vary Drastically Average and Non-Optimized vs. Designs 7 2.1.1 Subsystem Examples 7 2.1.1.1 Front Door 7 2.1.1.2 Front Bumper 9 2.1.1.3 Body Structure 11 2.1.2 All Subsystems Comparison Table 13 2.2 When Compared to an Design, the Mass Savings Gap with Aluminium Significantly Reduces 14 2.2.2 Subsystem Examples 14 2.2.2.1 Front Door 14 2.2.2.2 Front Bumper 16 2.2.2.3 Body Structure 18 2.2.3 Summary of Subsystems 20 2.3 Mass Savings Achieved at the Component Level Often Not Realized at the System Level 21 2.3.1 Subsystem Examples 21 2.3.1.1 Front Door 21 2.3.1.2 Front Bumper 22 2.3.2 Summary of Subsystems 23 2.4 There Is A Narrow Margin In Vehicle Curb Weights Between Vehicles Using Stee Body Structures And Aluminium Body Structures 24 Page WorldAuto 2015 i

2.5 There is Yet Untapped Mass Savings Potential for 25 2.5.1 FutureVehicle Body Structure 25 2.5.2 Member Company Examples 25 3.0 The Cost of Lightweighting 26 3.1 Subsystem Examples 26 3.1.1 Body Structure 26 3.1.2 Door Structure 26 3.1.3 Hood Structure 26 3.2 Relationship Between Cost and Fuel Efficiency 27 3.3 Summary 28 4.0 Life Cycle Assessment 29 4.1 Model Parameters 29 4.1.1 Body Structure Mass 29 4.1.2 Bill of Materials (BOM) Calculations 30 4.2 Body Structure Results 31 4.2.1 Total Life Cycle GHG Emissions 31 4.2.2 Body Structure GHG Emissions by Life Cycle Phase 31 4.3 Other Systems 34 4.3.1 Subsystem Masses 34 4.3.2 LCA Results 35 5.0 Conclusions 36 5.1 Final Observations on the Power of Statistical Benchmarking as a Tool for Mass Efficiency 37 Annotations 38 Appendix 1 Model Equations 39 WorldAuto 2015 ii

1.0 Introduction The increasing demand for automotive fuel efficiency and mass reduction has resulted in increased use of alternative materials in the design of various components that have historically been produced using steel. Mass benchmarking often is done with a one-at-a-time approach. A reference vehicle is selected, the vehicle is disassembled, parts weighed and analyzed, and then the data used to set mass targets for a vehicle under design. However, a benchmarking study results, commissioned by WorldAuto and the Market Development Institute (SMDI) and conducted by EDAG Int l, Inc., point to a powerful new statistics-based benchmarking methodology. Rather than considering a single vehicle or a small set of vehicles, this method looks at a very large sample of vehicles over a range of sizes and segments. From this larger population, assumptions on mass drivers and their influence on real vehicles may be tested via statistical methods. An automotive designer can then look at subsystems which are much lighter than the average vehicle, and therefore set subsystem targets on a more accurate basis than that which is being accomplished in the industry today. The findings of the study, which used the methodology to study automotive closures, were surprising and put single-point mass studies into context with the reality of what is achieved in current vehicles. The findings can be summarized in the following statements: 1. Mass efficiency of today s steel designs vary drastically. When comparing some steel components of the same size and function, there are drastic differences in the range of weights in what should be similar mass subsystems. In fact, some steel designs of the same size, performance and similar segment are nearly twice as heavy as others. There is a great deal of untapped opportunity in current production vehicles for mass reduction, even with the technology and materials already in use. 2. When compared to an efficient steel design, the mass savings gap with aluminium significantly reduces. Statistical evaluation conducted in the study identified those components that are most mass efficient. When the efficient steel designs were compared to the aluminium designs in the study sample, the mass savings achieved was greatly reduced, and in some cases, resulted in a mass increase for aluminium compared to steel. 3. Mass savings achieved at the component level are often not realized at the system level. While this study did not investigate the reason for this loss of mass advantage, it is clear that in nearly every component reviewed, the mass savings that was achieved at the component structure level was lost along the design development process. Is the structure mass savings causing the addition of other, heavier features to the system? Are other elements of the system not downsized to adjust to the lighter weight? Is the goal for mass savings simply to allow inclusion of other heavier features? These are unanswered questions about which we hope the results of this study will spur further discussion. 4. There is a narrow margin in vehicle curb weights (CVW) between vehicles with steel structures and those with aluminium structures. The data show that while the state-of-the-art of the aluminum body structures in the database reduce vehicle curb weight by 9.3% compared to average steel structures, current efficient steel structures reduce vehicle curb weight by 6.5% compared to average steel structures, closing the gap with aluminium to just 2.8%. 5. There is yet untapped mass savings potential for steel. Current applications do not take advantage of the full portfolio of steel grades with the newest GigaPascal strength steels that also increase ductility and flexibility. industry application demonstrations, which use a broader portfolio of steel grades than seen in today s production designs, have demonstrated greater mass reduction potential than currently achieved. WorldAuto 2015 1

Two other important dimensions are also addressed in this study costs and environmental performance. Cost: Reference examples of subsystem cost comparisons taken from other studies were used to illustrate the cost of the lightweighting achieved. Environmental performance: A separate Life Cycle Assessment LCA) was conducted on several key components to compare emissions savings achieved. The following report details the findings and brings into sharp focus the gap between current lightweighting efforts and their impact on vehicle mass and environmental efficiency and the need to include this statistical methodology to develop design mass targets. 1.1 Study Scope The study investigated mass efficiency for the vehicle subsystems shown in Table 1.1-1. Table 1.1-1: Vehicle Subsystems Included in the Study a. Front door frame f. Wheels b. Hood g. Instrument panel beam c. Hatchback h. Front bumper beam d. Deck lid i. Rear bumper beam e. Lift gate j. Body structure The study included the following scope of work for each of the identified subsystems. A. The mass statistics including mean, standard deviation, minimum and maximum. B. Compare and contrast the trend-lines of the material usage and mass efficiency of these more recent vehicle systems to those of the 2010 report. C. Update the insight and conclusions into the mass drivers of each subsystem. D. Identify the subsystems with exceptional mass efficiency, and identify the effect that material selection has on this efficiency. The mass data was taken from the A2Mac1 European and North American tear-down databases. The method employed looks at a large sample of vehicles (240) over a range of sizes, segments and markets. From this larger population the subsystems which are much lighter mass efficient than the average vehicle were determined, after adjusting for the system attributes. 1.1.1 Mass Benchmarking Methodology The methodology followed in this study is consistent with a Society of Automotive Engineers paper published in 2015 by Dr. Don Malen, University of Michigan, and Jason Hughes, A2Mac1. 1 Figure 1.1-1 provides a summary of the steps in the process. WorldAuto 2015 2

Figure 1.1-1: Steps in the Statistical Benchmarking Process as Detailed in the Referenced SAE Paper The analysis described in this WorldAuto report used the Regression models identified in the SAE paper to examine the database and draw conclusions. This Statistical regression analysis (Regression Process in the diagram) was used to fit both a linear equation and a power equation to the various components attributes. The linear model was first determined retaining the statistically significant attributes, and then the same retained attributes were used for the power model. The generic forms of the model equations used were as follows. The specific equation used for each component is provided in Appendix 1. Linear models were of the form: ḿ Where ḿ is the subsystem nominal mass for the full set of vehicles considered. Some of the attributes were continuous variables (gross vehicle weight, for example), some were categorical attributes (construction A, construction B, or construction C), and others were a binary attribute (side air bag: yes/no). After determining significant attributes using the linear model, a power model was fit of the form: ḿ and was used for comparing mass performance of competitive material designs. To determine a model for efficient designs, the power model equation was modified by varying the number (defined as n value) of standard errors below the mean, keeping at least three data points below the curve: ḿ Where: ḿ EFF = predicted mass for mass efficient designs r = standard error factor determined by the regression n = number of standard errors below the mean for which at least three samples were observed WorldAuto 2015 3

Figure 1.1-2, a plot of the actual mass value versus the primary mass driver, shows the average and efficient model equations for each material. Figure 1.1-2: Example Plot Actual Mass Value vs. Primary Mass Driver Mass efficient components were identified by calculating the estimated mass using the linear model equation (using separate equations for steel and for aluminium) and by plotting the actual measured mass versus the estimated mass for each component, as show in Figure 1.1-3. The solid line represents cases where actual and estimated masses are equal (i.e. average ). Those door frames above the solid line (where actual is greater than estimated) are heavier than average whereas those below the solid line represent door frames lighter than the average Subsystems which are much lighter than the average vehicle (those that fall at least 1 standard deviation below the curve) are considered to be mass efficient. Figure 1.1-3: Example Plot Actual Measured Mass vs. Estimated Component Mass Functional Equivalence / Performance Criteria: Within this mass benchmarking methodology, there may be questions about whether mass efficient components meet functional market performance requirements and therefore are not lighter weight because of compromises in performance. A separate WorldAuto 2015 4

door study sponsored by SMDI and conducted by EDAG 3 verified that door structures, as one prime example, meet market performance criteria and that performance differences are not generally an explanation for differences in mass efficiency. Additionally, this statistical methodology minimizes the effect of a few outlier points in relation to the bulk of the data set points. 1.1.2 Vehicle Types Included As of July 2012 the A2Mac1 European database contained 172 vehicles and the North American database contains 24 vehicles. The number of vehicles per production year is shown in Figure 1.1-4. The types of vehicles in the total data sample are shown in Figures 1.1-5. For the new data set, the vehicle destination market was determined to be 42 to North/South America, 64 to Europe/United Kingdom, and 4 to Pacific Rim/Asia. These destination market figures do not reflect the distribution of vehicle production locations. The combined A2Mac1 database contains a broad representation of global automotive manufacturers. NUMBER OF VEHICLES 40 35 30 25 20 15 10 5 0 2000200120022003200420052006200720082009201020112012 VEHICLE PRODUCTION YEAR Figure 1.1-4: Number of Vehicles per Production Year in Total Data Sample NUMBER OF VEHICLES 100 90 80 70 60 50 40 30 20 10 0 Van SUV Sedan Hatchback Light Duty Truck VEHICLE TYPES Figure 1.1-5: Vehicles Types in Total Data Sample WorldAuto 2015 5

1.1.3 Cost References This report focuses on mass benchmarking. In order to provide direct cost comparison data for the subsystems studied, hardware tear-down and engineering analysis cost study of representative subsystems would be necessary (but, was not part of this study s work scope). However, for the purpose of this report, subsystem cost comparisons relevant to this Auto Mass Benchmarking study were taken from other studies as references. These reference examples illustrate (Section 3.0) typical cost comparison data. 1.1.4 Life Cycle Assessment Life Cycle Assessment (LCA) is a methodology that considers a vehicle s entire life cycle, from the manufacturing phase (including material production and vehicle assembly) through the use phase (including production and combustion of fuel) to the end of life (EOL) phase (including end of life disposal and recycling). Current automotive emissions regulations around the world are aimed at reducing Greenhouse Gas (GHG) emissions of automobiles, but focus only on tailpipe emissions, which are only a portion of the actual lifecycle impact of an automobile (Figure 1.1-6). Emphasis on the tailpipe alone may have the unintended consequence of increasing GHG emissions during the vehicle life. For example, many automakers, in order to comply with increasingly stringent tailpipe emissions regulations, are turning to low- density materials in an effort to reduce mass. By reducing the mass of a vehicle, it is possible to reduce the fuel consumption and, Figure 1.1-6: Sources of GHG Emissions In A Vehicle's Life Cycle consequently, the tailpipe emissions. However, many of these materials can have impacts in the other life cycle phases that outweigh any advantage that may be gained in the use phase. This means that, contrary to the stated objective of reducing the GHG emissions of automobiles, tailpipe-only regulations may have the unintended consequence of actually increasing the GHG impact. This is why WorldAuto is participating in the development of LCA tools and methodology and encouraging the use of LCA in the formulation and implementation of automotive emissions regulations. Section 4.0 contains an LCA study based on the EDAG A2Mac1 benchmarking study results. As with the FutureVehicle (FSV) program 2, the LCA work was conducted for key subsystems (body structure, front bumper, rear bumper, wheels, hatchback, hood, front door) using the University of California at Santa Barbara (UCSB), Bren School of Environmental Management s Automotive Materials Greenhouse Gas (GHG) Comparison Model 4, whose methodology has been peer-reviewed and approved according to ISO 14040:2006. Specific parameters and methodology are outlined in Section 4.0. WorldAuto 2015 6

2.0 Automotive Mass Benchmarking Key Study Findings 2.1 Mass efficiency of today s Designs Vary Drastically Average and Non-Optimized vs. Designs As the data was investigated during this study, it became clear that all steel components are not created equally. For instance, there are many levels of efficiency in the sample among the steel doors, with some doors of similar segment class and performance weighing twice as much as others. There are similar examples among some of the other components, which indicate that there is much room for improvement and optimization among the current production steel component structures. Following are examples comparing the average versus efficient mass as well as the non-optimized mass (defined as the approximate heaviest example considering the same mass drivers) of the steel component structures in the data. 2.1.1 Subsystem Examples 2.1.1.1 Front Door The front door system is the complete door assembly inclusive of intrusion beam, glass, linkages, trim, lock and all mounting hardware. The door frame consists of the door frame welded assembly: door inner and outer, window frame, door beam, and all welded reinforcements. For all the doors in the sample (219), the dimensions shown in Figure 2.1-1 were recorded. Table 2.1-1 provides example data. Table 2.1-1: Example of Data Collected for the Front Door Front Door Frame Material: 1= 2= Alum 3=Mag 4= Others Figure 2.1-1: Front Door Dimensions Area (length x beltline height + top triangular) (m 2 ) Window Regulator Type: 1= Cable 2= Linkage 3= Print Window Frame Type: 1= Rear of A 2= Overlapping 3= Frameless Front Door Frame (kg) Front Door Total (kg) Front Door Glass (kg) Length (mm) Height (mm) Tumble home (mm) Hinge Span (mm) 12.91 1 26.40 3.16 1167.00 1163.0 1.081 245.0 388.0 3 1 16.27 1 32.79 3.11 1248.0 1100.0 1.373 254.0 447.0 3 1 WorldAuto 2015 7

The statistical data for the door frame weights and the number of vehicles are shown in Table 2.1-2. Table 2.1-2: Front Door Frame Weights Door Frame Material Nr Mass (kg) Mean St Dev Min Max 203 17.01 2.70 9.85 27.20 Aluminium 16 11.71 1.77 9.23 15.54 A significant mass driver for the door frame was found to be the side view area, the material of construction and, to a lesser degree, the window regulator type. Figure 2.1-2 illustrates average door mass and efficient door mass plus one of the heaviest (non-optimized) doors for the specified area of 1.1m^2. Table 2.1-3 shows that the non-optimized door is +23% heavier than the average steel door. The efficient steel door, among the lightest according to the regression analysis, saves -21% mass compared to the overall average and a full -35% over the non-optimized example. Figure 2.1-2: Door Frame Comparison Average vs. and Heaviest (Non-Optimized) Door with 1.1 m^2 area Table 2.1-3 Front Door Frame Material Comparisons Comparison Average to Designs Frame Mass (kg) Average 16.3 Mass Difference Non-Optimized 19.9 + 23% 12.9-21% For door with 1.1 m^2 area WorldAuto 2015 8

2.1.1.2 Front Bumper The front bumper system is the complete bumper assembly inclusive of fascia, energy absorbing foam and all mounting hardware. The bumper beam consists of the bumper beam welded assembly, beam and crash cans and all welded on reinforcements, Figure 2.1-3. For all the bumper systems in the sample (222) the data shown in Table 2.1-4 was recorded. Figure 2.1-3: Front Bumper System Table 2.1-4: Example of Collected Data for the Front Bumper Beam Front Bumper Mass Total (kg) Front Bumper Beam Mass (kg) Beam Mat: 1 = Stee 2=Alum 3=Mag 4=Plasti c Manufacturing Method: 1=Extrusion 2=Stamped 3=Roll Form 4=Hydroform 5=Hot Stamp Crush Cans: 1=Yes 2=No Bumper Length (mm) Distance Between Rails (mm) CVW Curb Vehicle Weight(kg) Vehicle Type: 1=Utility: pick up & SUV, 2=Passeng er: sedan, Hatchback & Others Destination Marke: 1 = North America 2 = Europe - UK 3 = Pacific Rim - Asia 15.36 3.70 2 1 1 1204.0 1026.0 1656 2 1 17.19 4.07 2 1 1 1180.0 1000.0 1328 2 2 The statistical data for the beam weights and the number of vehicles versus beam weights are shown in Table 2.1-5. Table 2.1-5: Example of Data Collected for the Front Bumper Beam Weights Front Bumper Beam Nr Mass (kg) Mean St Dev Min Max 150 7.31 3.08 2.26 28.19 Aluminium 72 4.89 1.51 1.45 9.23 A significant mass driver for the front bumper beam was the vehicle curb weight (CVW), construction material steel or aluminium, and width of the mounting rails. A notable example of the need for optimization of current steel production components can be seen with the front bumper, though the data is similar for the rear bumper as well. The non-optimized steel bumper (approximated in Figure 2.1-4) is nearly double (+89%) the weight of the average bumper in the same CVW of 1500 kg (Table 2-1-6). The steel efficient bumper saves -47% mass over the average. WorldAuto 2015 9

Figure 2.1-4: Front Bumper Frame Comparison Average vs. and Heaviest (Non-Optimized) Bumper CVW 1500 kg, 982 mm Table 2.1-6 Front Bumper Frame Material Comparison Comparison Average, and Non- Optimized Bumper Mass Mass (kg) Difference Average 6.3 Non-Optimized 11.1 + 89% 3.3-47% For CVW 1500kg, Rail Width 982mm WorldAuto 2015 10

2.1.1.3 Body Structure The body structure analyzed is the complete body assembly inclusive of paint, sealer and the engine cradle as shown in Figure 2.1-5. Figure 2.1-5: Body Structure and Engine Cradle Of the 240 vehicles, suitable data for the body structure was available for 217 vehicles, 207 constructed from steel, four from aluminium and six mainly steel but with some panels made from aluminium. The statistical data for the structure weights and the number of vehicles are shown in Table 2-1.7. Table 2.1-7: Body Structure Weights Body Structure Material Nr Mass (kg) Mean St Dev Min Max 207 334.9 72.3 159.5 665.9 Aluminium 4 247.3 68.8 144.4 289.1 /Aluminium 6 389.6 64.5 301.6 465.2 In order to normalize the mass of the body structure to its size, for all the body structures in the sample (217) the data shown in Table 2.1-8 was collected. Table 2.1-8: Example of Collected Data for Body Structure Body Structure + Engine Cradle Mass Total (kg) Body Structure Mass (kg) Body Material: 1 = 2=Alum 3=Mag 4=Plastic 5= - Alloy GVW (Gross Vehicle Weight) kg Vehicle Type: 1=Utility: pick up & SUV 2=Passenger: sedan, Hatchback & Others Destination Market: 1 = North America 2 = Europe - UK 3 = Pacific Rim - Asia Drive Config: 1 = FWD 2 = RWD 3 = AWD 4 = 4WD Plan View Area (m^2) 366.8 349.4 5 2259 2 2 3 9.496 356.4 339.0 5 2123 2 2 1 9.211 Significant mass drivers identified for the body structure were gross vehicle weight (GVW), the plan view area of the vehicle (length x width) and the material used for construction. The average curves are plotted for incremental plan view areas since both GVW and plan view area are mass drivers for body structures. WorldAuto 2015 11

Figure 2.1-6 and Table 2.1-9 provide this comparison for body structures. The smaller difference among the average, non-optimized and steel efficient may be indicative of the industry s efforts to reduce vehicle weight and what is still left to be done. Included in this figure and table is the fully optimized, lightweight FSV, which includes a high AHSS body structure content (95%). FSV and other concept projects (see examples in Section 2.5) note opportunities for more mass reduction with advanced steels and steel technologies. Figure 2.1-6: Body Structure Comparison Average vs. and Heaviest (Non-Optimized) Body Structure with GVW 1928 kg, Area 8.75 m 2 Table 2.1-9: Body Structure Material Comparison Comparison Average to Designs Structure Mass (kg) Average 348.3 Mass Difference Non-Optimized 385.0 + 10% 305.7-12% FSV AHSS 286.6-18% GVW 1928kg & Plan Area 8.75 m 2 WorldAuto 2015 12

2.1.2 All Subsystems Comparison Table Table 2.1-10 provides the data for mid-size vehicle steel componentry currently on the road today, comparing the average steel designs to the most efficient in each mass driver. It provides an overview of the opportunities that still exist to employ AHSS and design optimization to obtain greater, more cost effective mass reduction. Table 2.1-10: Mass saving potential - Average vs. - Structure Only Mid-Size Vehicle Estimated Mass Normalized Designs Average Vehicle Subsystem Mass Driver Values Difference kg kg % Front Doors Area = 1.1 m 2 16.3 12.9-21% Hood Area = 1.68 m 2 13.5 10.4-23% Hatchback Area = 1.23 m 2 & 444mm depth 11.4 8.7-24% Decklid Area = 1.34 m 2 10.7 8.9-17% Liftgate Area = 2.11 m 2 15.3 12.2-20% Wheel Rim Dia = 415mm & Width = 184mm 9.2 7.2-22% Front Bumper CVW 1500kg, Rail Width 982mm 6.3 3.3-47% Rear Bumper CVW 1500kg 6.2 2.7-57% Body Structure GVW 1928kg & Plan Area 8.75 m 2 348.3 305.7-12% Body Structure FSV-AHSS GVW 1928kg & Plan Area 8.75 m 2 286.6-18% Note: Structure mass based on power regression equations WorldAuto 2015 13

2.2 When Compared to an Design, the Mass Savings Gap with Aluminium Significantly Reduces Aluminium material substitution for steel is often credited with as much as 40% mass savings. This could be true if the comparison is made with non-optimized, inefficient steel designs. What is the reality of mass savings achieved when comparisons are made to the spectrum of vehicles in this database? The following subsystem examples address that question. 2.2.2 Subsystem Examples 2.2.2.1 Front Door Figure 2.2-1 illustrates the relationship of the door frame mass to side view area. Appendix 1.A. provides the equations used to identify the expected average for steel and aluminium door frames, as well as the mass estimates for the efficient door frame and corresponding n values. Figure 2.2-1: Front Door Frame Mass vs. Area of Door Frame (Power Estimation Models) Table 2.2-1 provides the estimated masses for a door frame with a side view area of 1.1m 2. The results indicate the average aluminium door frame is 30% lighter than the average steel. However, when the steel design is an optimized, fully efficient one, the mass savings potential for aluminium is reduced to 22% (2.8 kg per door), assuming that the aluminium door is also an efficient design. However, the average door among the aluminium designs, those whose weight represents the more typical masses seen in production today, the mass savings is reduced to 1.6 kg per door or just 12%. Table 2.2-1 Front Door Frame Material Comparisons Comparison Average to Aluminium Average Designs Comparison to Aluminium Designs Frame Mass Reduction Frame Mass Reduction Mass (kg) % Delta (kg) Mass (kg) % Delta (kg) Average 16.3 12.9 Alum Average 11.3-30% 5.0 Alum 10.1-22% 2.8 For door with 1.1 m^2 area Alum Average 11.3-12% 1.6 WorldAuto 2015 14

Figure 2.2-2 is a parity plot of actual vs. estimated masses for each door frame. The solid line represents cases where actual and estimated masses are equal (i.e. average ). Those door frames above the solid line (where actual is greater than estimated) are heavier than average whereas those below the solid line represent door frames lighter than the average. Highlighted on this graph in green are the steel door frames which were determined to be most mass efficient based on the linear regression model. OEM Data Figure 2.2-2: Front Door Frame Mass Designs (Linear Estimation Models) As an example, included in Figures 2.2-3 and 2.2-4 are graphs highlighting the doors for one particular, but unidentified, OEM. From the charts, it can be seen that some of this OEM s doors are efficient and others are not. This type of data can provide a unique comparison for OEMs to identify the outliers in efficiency compared to other current production vehicles. Figure 2.2-3: OEM Example, Mass vs. Area Comparison Figure 2.2-4: OEM Example, Estimated vs. Actual Comparison WorldAuto 2015 15

2.2.2.2 Front Bumper Figure 2.2-5 illustrates the relationship of the bumper beam mass to the vehicle curb weight (CVW). Appendix 1.G. provides the front bumper beam equations. Figure 2.2-5: Front Bumper Mass vs. Area (Power estimation models) Table 2.2-2 shows the estimated masses for the bumper beam for CVW of 1500 kg and rail mounting width of 982 mm. The results indicate a mass saving of 33% for the average values for aluminium versus steel. The mass potential of aluminium is reduced to 11% (0.4 kg per beam) when comparing the efficient designs of both materials. When compared to a steel efficient design, an average aluminium bumper increases component mass by +26% (+0.9 kg). Table 2.2-2: Front Bumper Material Comparison Comparison Average to Aluminium Average Designs Material Option Frame Mass (kg) Mass Reduction Comparison to Aluminium Designs Mass Frame Reduction Material Option Mass (kg) % Delta % Delta Average 6.3 3.3 Alum Average 4.2-33% 2.1 Alum 3.0-11% 0.4 Alum Average 4.2 + 26% + 0.9 For CVW 1500 kg, Average Rail Width 982 mm WorldAuto 2015 16

Figure 2.2-6 is a parity plot of the actual versus estimated masses for front bumper beams. Bumper beams above the solid line are heavier than average whereas those below the solid line represent bumper beams lighter than the average. Highlighted on this graph in green are steel front bumpers beams which are exceptionally lighter. OEM Data Figure 2-2.6: Front Bumper Beam Mass Designs (Linear estimation model) Figures 2.2-7 and 2.2-8 are graphs indicating the front bumpers for another unidentified, OEM. Note again that there are those examples within the same OEM brand that are considered efficient designs and those that are not. Figure 2.2-7: OEM Example, Mass vs. Curb Mass Comparison Figure 2.2-8: OEM Example, Estimated vs. Actual Comparison WorldAuto 2015 17

2.2.2.3 Body Structure Figure 2.2-9 illustrates the relationship of the body structure mass to Gross Vehicle Weight (GVW). The average curves are plotted for incremental plan view areas since both GVW and plan view area are mass drivers for body structures. Figure 2.2-9: Body Structure Mass vs. GVW and Area (Power estimation models) Table 2.2-3 shows the estimated masses for a typical mid-size sedan body structure with GVW of 1928 kg and plan view area of 8.75 m². The results indicate a mass saving of 31% for the average values for aluminium versus steel. The mass potential of aluminium is reduced to 16% when comparing the efficient designs of aluminium with FSV AHSS structures. Due to the limited data points for aluminium body structures, no differentiation was made between average and efficient aluminium structures. Table 2.2-3: Body Structure Material Comparison Body Material Options Body & Engine Cradle Mass (kg) Average 348.3 Mass Reduction % Delta (kg) - Alum 329.9-5% 18.4 305.7-12% 42.5 FSV AHSS 286.6-18% 61.7 Aluminium 240.7-31% 107.5 Compared to FSV -16% 45.9 Plan View Area: 8.75 m^2 GVW: 1928 kg (Mid Size Sedan) WorldAuto 2015 18

Figure 2.2-10 is a parity plot of the actual versus estimated masses for body structures. The solid line represents cases where actual and estimated masses are equal (i.e. average ). Body structures above the solid line are heavier than average whereas. The body structures below the estimated equals actual line represent body structures lighter than the average. Highlighted on this graph are green circle points for steel body structures which are exceptionally lighter `(efficient). The green triangular points on the graph are the FSV body structure points. The red square points identify aluminium body structures. OEM Data Figure 2.2-10: Body Structure Mass Designs (Linear estimation models) Figures 2.2-11 and 2.2-12 include graphs indicating the body structures for one particular, but unidentified, OEM. Figure 2.2-11: OEM Example, Mass vs. GVW Comparison Figure 2.2-12: OEM Example, Estimated vs. Actual Comparison WorldAuto 2015 19

2.2.3 Summary of Subsystems The summary of results in Tables 2.2-4 and 2.2-5 shows that the mass saving potential of aluminium is significantly reduced when compared with efficient steel designs. Table 2.2-4: Mass saving potential - Average vs. Average Aluminium - Structure Only Vehicle Subsystem Mid-Size Vehicle Mass Driver Values Estimated Mass Normalized Designs Average Average Difference Aluminium kg kg % Front Doors Area = 1.1 m 2 16.3 11.3-30% Hood Area = 1.68 m 2 13.5 8.1-40% Hatchback Area = 1.23 m 2 & 444mm depth 11.4 8.1-29% Decklid Area = 1.34 m 2 10.7 8.1-24% Liftgate Area = 2.11 m 2 15.3 10.2-33% Wheel Rim Dia = 415mm & Width = 184mm 9.2 8.3-10% Front Bumper CVW 1500kg, Rail Width 982mm 6.3 4.2-33% Rear Bumper CVW 1500kg 6.2 4.3-31% Body Structure GVW 1928kg & Plan Area 8.75 m 2 348.3 240.7-31% Note: Structure mass based on power regression equations Table 2.2-5: Mass saving potential - vs. Aluminium - Structure Only Vehicle Subsystem Mid-Size Vehicle Mass Driver Values Estimated Mass Normalized Designs Difference Aluminium kg kg % Front Doors Area = 1.1 m 2 12.9 10.1-22% Hood Area = 1.68 m 2 10.4 6.9-34% Hatchback Area = 1.23 m 2 & 444mm depth 8.7 7.6-13% Decklid Area = 1.34 m 2 8.9 7.8-11% Liftgate Area = 2.11 m 2 12.2 10.2-16% Wheel Rim Dia = 415mm & Width = 184mm 7.2 6.4-12% Front Bumper CVW 1500kg, Rail Width 982mm 3.3 3-11% Rear Bumper CVW 1500kg 2.7 2.1-22% Body Structure GVW 1928kg & Plan Area 8.75 m 2 305.7 240.7* -21% Body Structure FSV-AHSS GVW 1928kg & Plan Area 8.75 m 2 286.6 240.7* -16% Note: Structure mass based on power regression equations * Due to limited data points for aluminium body structures, no differentiation was made between average and efficient aluminium structures. WorldAuto 2015 20

2.3 Mass Savings Achieved at the Component Level Often Not Realized at the System Level The following subsystem examples and summary table provide insight on the mass reduction achieved at the component structure level, e.g., a door frame assembly, compared to the mass of the total system assembly. In many of the systems studied the mass savings achieved at the component structure level are not seen in the mass of the total system. One possible reason for this is that the lighter component structures are in premium vehicles and the mass discrepancy in the total system is a result of additional content. This possibility was investigated and no such correlation was identified. Section 2.1.1 provides the details on the composition of the total assembly for each example subsystem following. 2.3.1 Subsystem Examples 2.3.1.1 Front Door Figure 2.3-1 shows the steel and aluminium door frames mass differences. Figure 2.3-2 plots the total system mass of the same doors. Note that the doors with lighter frames are in parity with the heavier frame doors once they have been assembled into the total door system and any mass reduction achieved is negated at the system level. Premium vehicle classes are flagged in the parity charts and tables to illustrate that the average masses are not inflated by this traditionally heavier vehicle class. Figure 2.3-1: Door Frame Mass Comparison Figure 2.3-2: Total Door System Mass Comparison Table 2.3-1 Door Frame vs. Total System Mass Comparison Data Frame Mass (kg) Mass Reduction Total Mass Reduction System % Delta (kg) Mass (kg) % Delta (kg) Avg 16.3 - - Avg 32.4 - - Prem 15.6-4% -0.7 Prem 33.5 3% 1.1 Alum Avg 11.3-30% -5.0 Alum Avg 32.5 0% 0.1 Alum Prem 11.1-32% -5.2 Alum Prem 33.1 2% 0.7 For door with 1.1 m^2 area For door with 1.1 m^2 area WorldAuto 2015 21

2.3.1.2 Front Bumper Figure 2.3-3 shows the differences in steel and aluminium bumper frames Figure 2.3-4 plots the total system mass of the same bumpers. Table 2.3-2 shows that average aluminium bumper beam frames are typically 33% ( 2.5kg) lower mass compared with the average steel beams. However at the system level, the mass reduction is reduced to 21%. Figure 2.3-3: Bumper Frame Mass Comparison Figure 2.3-4: Total Bumper System Mass Comparison Table 2.3-2 Front Bumper Beam vs. Total System Mass Vehicle Subsystem Beam Only Average Designs Aluminum Total System Average Designs Aluminum % Delta % Delta Front Bumper 6.3 4.2-33% -2.1 15.3 15.9 +4% +.6 WorldAuto 2015 22

2.3.2 Summary of Subsystems As shown in Table 2.3-3, most of the systems in the database do not fully benefit, or show no benefit, from the mass savings achieved at the component level. Though identifying the cause is not a part of the scope of this study, it is a reasonable deduction that the efforts made to reduce component mass are being lost along the vehicle design chain. Table 2.3-3: Mass saving potential - vs. Aluminium Structure & Total System Vehicle Sub System Component Mass Savings & Total System Mass Savings Component (Structure) only Avg. Avg. Alum Diff Avg. Total System Avg. Alum kg kg % kg kg % Front Doors 16.3 11.3-30% 32.4 32.5 0% Hood 13.5 8.1-40% 17.2 11.4-34% Hatchback 11.4 8.1-29% 21.7 20.1-7% Decklid 10.7 8.1-24% ** ** ** Liftgate 15.3 10.2-33% 31.3 28.6-9% Wheel Rim 9.2 8.3-10% * * * Front Bumper 6.3 4.2-33% 15.3 15.9 +4% Rear Bumper 6.2 4.3-31% 13.7 10.8-21% Body Structure 348.3 240.7-31% * * * * Structure and System are same Diff ** Sufficient statistical data not available WorldAuto 2015 23

2.4 There Is A Narrow Margin In Vehicle Curb Weights Between Vehicles Using Body Structures And Aluminium Body Structures The data from Table 2.4-1 can be used to develop an approximate analysis of the total vehicle curb weight change that can be achieved if most main subsystem mass reductions, as shown, were accomplished. Table 2.4-1 assumes subsystem mass reductions for a typical mid-size passenger car and compares the accumulated mass reduction to the original curb weight, which for this example is assumed to be 1500 kg. The data show that while the state-of-the-art of the aluminum body structures in the database reduce vehicle curb weight by 9.3% compared to average steel structures, current efficient steel structures reduce vehicle curb weight by 6.5% compared to average steel structures, closing the gap with aluminium to just 2.8%. Table 2.4-1: Curb Weight Reduction - Average vs. Design Total System Vehicle Subsystem Average Aluminium kg kg Diff kg Diff Front Doors x 4 129.6 130.0 +0.3 110.8-18.9 Hood 17.2 11.4-5.8 13.6-3.6 Decklid 17.7 16.6-1.1 16.3-1.4 Wheel Rim x 4 36.8 25.5-11.4 29.0-7.9 Lower Control Arms x 2 (McPh) 16.3 11.5-4.6 11.2-5.0 Other Front Suspension (McPh) 34.1 27.4-6.7 27.4-6.7 Rear Suspension (twist beam) * 47.6 47.6 0.0 39.5-8.1 Front Bumper 15.3 15.9 +0.6 14.8-0.5 Rear Bumper 13.7 10.8-2.9 10.6-3.1 Body Structure 348.3 240.7-107.6 305.7-42.6 All Other Systems (no change) 823.4 823.4 0.0 823.4 0.0 Vehicle Curb Weight 1500 1361-139 1402-98 Curb Wt. % Reduction: 9.3% 6.5% * Twist beam available only in steel Note that the above 2.8% difference between steel and aluminum can be reduced even further with continued adaptation of steel advances (see next section). For example, if an FSV-type body structure (286.6 kg for mid-size vehicle) would be utilized in the above table, the gap could potentially be reduced to less than 2%. Curb Weight reduction impact on fuel consumption: The Table 2.4-1 margin of 2.8% between an efficient steel vehicle and aluminium vehicle results in a very small change in fuel consumption (fuel economy). Even assuming an average 30 miles per gallon (7.8 liters per 100 kilometers) vehicle with a perfectly optimized powertrain, the largest gain that could be expected with this small mass difference is only about 0.6 miles per gallon (0.15 liters per 100 kilometers) a very small gain in exchange for higher costs and poorer life cycle environmental impact. WorldAuto 2015 24

2.5 There Is Yet Untapped Mass Savings Potential For The steel efficient designs in the current production vehicle data set provide evidence that there is, indeed, more potential for steel to reduce structural mass. Design optimization techniques combined with AHSS can provide cost effective mass reduction, as particularly noted when comparing aluminium average and efficient designs to steel efficient designs. Above and beyond what automakers themselves are demonstrating in efficient designs, WorldAuto members, both collectively and individually, have invested sizably in research and development programs that demonstrate examples of automotive applications that can meet the challenges of mass reduction without compromise to performance, and using manufacturing facilities and infrastructures already in place. Following are examples: 2.5.1 FutureVehicle Body Structure Extensive use of a broad portfolio of AHSS grades, coupled with engineering design optimization, enables a robust body structure that is feasible to produce and achieves 5-star crash performance against all global crash standards, while also exceeding mass reduction targets. A lower weight, mass-efficient body creates opportunities for downsizing subsystems, including the powertrain, and promotes reductions in overall vehicle mass. As an example, the body structure mass achievement for the FSV Battery Electric Vehicle (BEV) variant is 177 kg, including the battery tray (unique to a BEV product). This compares quite favorably with the 2010 VW Polo, at 231 kg, which was recognized as Car of the Year in Europe as a result of its mass-efficient design, but carries a lighter ICE gasoline powertrain. Please note Table 2.5-1 for benchmarking comparisons. Table 2.5-1: FSV benchmarking Vehicle Class Powertrain Curb Weight Body Mass Fuel Consumption (NEDC) FSV BEV B+ BEV 958 kg 177 kg 2.4 l/100 km (equiv) ULSAB - AVC C ICE-G 933 kg 202 kg 4.4 l/100 km VW Polo B ICE-G 1067 kg 231 kg 5.7 l/100 km 2.5.2 Member Company Examples 5 FSV s steel portfolio represented those grades that would be commercially available within the 2017-2025 timeframe. company projects, however, demonstrate mass reduction and design efficiency using AHSS available on the market today and achieve significant mass reduction. An AHSS steel body structure designed within the packaging confines of an existing production vehicle achieved 25% mass reduction with steel having average tensile strengths of 900 MPa. An AHSS steel body structure for an electric vehicle achieved 26% mass reduction through the use of 65% (of total body weight) AHSS, which includes 45% Ultra High-Strength (UHSS) grades, combined with cost effective steel technologies. An advanced door design that uses AHSS DP 300/500 0.55mm steel achieves 11% mass reduction, material cost neutral. A door design using a combination of existing AHSS and UHSS to achieve 27% mass reduction; and another that stretches to technologies available 2017 and beyond that achieves 34% mass reduction. These are just a few examples of steel industry efforts to help automakers achieve the necessary mass reduction to meet new tailpipe regulations with steel that are both cost effective and implementable today. WorldAuto 2015 25

3.0 The Cost of Lightweighting A cost comparison study from 2001 6 shows that the cost of an aluminium body structure compared with a steel structure is typically a $600 increase. This is also a rule of thumb used in the industry by some body design engineers. A study by the National Highway Transportation Safety Administration (NHTSA) 7, based on 2011 figures, showed a cost increase of $720 for the aluminium body structure and a cost increase of $147 for the AHSS design. 3.1 Subsystem Examples 3.1.1 Body Structure Table 3.1-1 shows the body structure material type and mass reduction potential based on the benchmark data from this study. Comparing the aluminium structure to the FSV AHSS shows a mass saving of 45.9 kg (286.6kg - 240.7kg). This mass saving will be achieved at an additional cost of $573 ($720 - $147), equivalent to $12.49 per kg saved, a relatively high premium to pay for mass reduction. 3.1.2 Door Structure Table 3.1-1: Body Structure Material Comparison Body & Engine Cradle Mass Reduction Body Material Options Mass (kg) % Delta (kg) Average 348.3 - Alum 329.9-5% 18.4 305.7-12% 42.5 FSV - AHSS 286.6-18% 61.7 Aluminium 240.7-31% 107.5 Compared to FSV -16% 45.9 Plan View Area: 8.75 m^2 GVW: 1928 kg (Mid Size Sedan) The EDAG NHTSA study showed a cost increase of $24.80 for the aluminium door structure and a cost increase of $5.12 for the AHSS design. Table 3.1-2 shows the door structure material type and mass reduction potential based on the benchmark data from this study. Comparing the aluminium structure to the AHSS shows an additional mass saving of 2.82 kg (12.92kg 10.10kg). This mass saving will be achieved at an additional cost of $19.68 ($24.80 - $5.12), equivalent to $6.98 per kg saved. 3.1.3 Hood Structure Table 3.1-2: Front Door Frame Material Comparison Door Frame Mass Mass Reduction Material Option (kg) % Delta (kg) Average 16.3 Alum Average 11.3-30% 5.0 12.9 Alum 10.1-22% 2.8 For door with 1.1 m^2 area According to the NHTSA study, a cost increase of $21.26 could be expected for the aluminium hood structure and a cost increase of $4.74 for the AHSS design. Table 3.1-3 show the hood structure material WorldAuto 2015 26

type and mass reduction potential based on the benchmark data from this study. Comparing the aluminium structure to the steel efficient (AHSS) shows an additional mass saving of 3.52 kg (10.42kg 6.90kg). This mass saving will be achieved at an additional cost of $16.52 ($21.26 - $4.74), equivalent to $4.69 per kg saved. Table 3.1-3: Hood Frame Material Comparison Hood Material Option Frame Mass (kg) Average 13.5 Mass Reduction % Delta (kg) Alum Average 8.1-40% 5.4 10.4 Alum 6.9-34% 3.5 For Hood with 1.68 m² area 3.2 Relationship Between Cost and Fuel Efficiency The NHTSA study sought to identify the maximum feasible weight reduction possible using lightweighting technologies which will be available in the years 2017 to 2025, to achieve fuel reduction savings standards as set by the Administration, while keeping costs to within 10%, plus or minus, over the baseline vehicle. Constraints of the study required the same vehicle footprint, styling and utility as the baseline vehicle (2011 Honda Accord) in passenger space, options, and luggage and towing capability. The study investigated four alternative lightweighting structural material concepts: 1) An AHSS-intensive design, which reduced mass by 19% compared to the baseline vehicle; 2) A Light Weight Vehicle Design which combined an AHSS body structure with aluminium closures, reduced mass by 22%; 3) an aluminiumintensive design, which reduced mass by 25%; and 4) a multi-material design (aluminum, magnesium and carbon fiber reinforced plastics) that reduced mass by 28%. Relative to the 2011 baseline, all scenarios reduce curb weight mass from 20% to 30% and increase fuel economy from the baseline 27 mpg to approximately 32 mpg (Figure 3.2-1). The study does not include enhanced powertrain efficiencies nor does it account for the reduced mass savings at the systems level shown by some aluminum designs, which would reduce both the curb mass savings and the fuel economy gains. A further comparison between the alternative advanced material solutions in Figure 3.2-1 shows that the alternatives to AHSS save the consumer about one fuel fill-up per year. This fuel saving cannot offset the increase in manufacturing costs associated with these materials over the average six-year ownership of a vehicle, per a recent R. L. Polk 8 study. WorldAuto 2015 27

Concept Design AHSS Intensive LWV Alum Intensive Alum/Mag/CFRP Material Applications Structural Material Content Conventional AHSS Aluminum Magnesium Carbon Fiber Figure 3.2-1: NHTSA Vehicle Results Source: NHTSA Report 7 AHSS Body AHSS Closures AHSS Chassis AHSS Body Alum Closures Alum Chassis Alum Body Alum Closures Alum Chassis CFRP Body Alum/Mag Closures Alum Chassis Curb Weight (kg) 1197 kg 1149 kg 1108 kg 1060 kg % Curb Weight Reduction - 4% 7% 11% Manufacturing Cost Premium - $210 $820 $2,680 Fuel Economy (mpg) 31.4 31.6 31.8 32.0 Reduced Fuel Fill-up/year - 0.4 1.1 1.5 Years to Recover Manufacturing Cost in Fuel Saving(@$4.00/gal) - 8.6 11.5 28.6 Increased LCA CO 2 (Tonnes) - 0.4 1.0 4.4 3.3 Summary Based on previous analysis conducted by EDAG, it was determined that the cost of lightweighting for the three subsystems reviewed could amount to more than US$12.00/kg of mass reduction. A recent Ricardo study 9 based on a survey of European OEMs indicated that there is an EU 5.00/kg limit to what OEMs are currently willing to pay for mass reduction. Further, the NHTSA study indicated that the cost of lightweighting with aluminium or a multi-material design would be US$820 and US$2,680 per vehicle (body structure + closures + chassis), respectively, and would require as many as 28 years to recover manufacturing costs in fuel savings. At current prices, use of aluminium for lightweighting is cost prohibitive for some applications. This report shows that benchmarking is an important tool in analyzing the value of competitive material lightweighting, considering their effectiveness to achieve actual mass reduction at the subsystem level, and when compared to steel efficient designs. WorldAuto 2015 28

4.0 Life Cycle Assessment Auto Mass Benchmarking study, to investigate the life cycle greenhouse gas (GHG) impact of three principal material usage categories in the body structure subsystems represented in the A2Mac1 data: 1. Average steel design Using regression methodology, EDAG developed a power model to determine an estimated mass of each subsystem based on the influence of a primary mass driver. In the case of the body structure, the primary mass driver is gross vehicle weight (GVW). 2. steel design The model developed for the average steel designs was iteratively manipulated until it was representative of the 17 most efficient steel designs. 3. aluminium design A power model of the most efficient aluminium designs was developed in the same way as for the efficient steel designs. These three categories were further compared with a fourth category, developed from body structure designs taken from the FSV program. FSV is a clean-sheet vehicle architecture that offers mass-efficient, steel-intensive solutions to automotive lightweighting challenges. The estimation of life cycle GHG emissions was conducted using the UCSB Automotive Materials GHG Comparison Model 4. The UCSB Model was designed to quantify the energy and GHG impacts of automotive material substitution on a total vehicle life cycle basis, under a broad range of conditions and in a completely transparent fashion, and has been peer-reviewed. 4.1 Model Parameters 4.1.1 Body structure mass Body structure masses from each of the four categories were applied to two vehicle classes as defined by the NHTSA 7 study: Passenger Car/Light (PC/L) curb weight 907-1134 kg Passenger Car/Compact (PC/C) curb weight 1134-1360 kg Mass drivers for vehicles in the A2Mac1 data that fit these categories were averaged and used in the appropriate models to generate an estimated average body structure mass for each category and class. The FSV design masses were averaged into the two NHTSA classes as follows: FSV1 (2 A-B class designs) PC/L FSV2 (2 C-D class designs) PC/C Table 4.1-1 provides the resulting body structure average masses: Table 4.1-1: Average Mass of Body Structures in kg NHTSA Class # in A2Mac1 GVW (kg) Average Aluminum FSV PC/L 11 1487 250.1 219.5 172.9 195.9 PC/C 16 1714 295.1 259.0 204.0 216.7 WorldAuto 2015 29

4.1.2 Bill of Materials (BOM) Calculations The bill of materials (BOM) for each design was calculated using the average curb mass of each category and class (Table 4.1-2). The UCSB model contains default values for each material as a percentage of curb mass. These defaults include a distribution in the body structure of 90% flat/10% long for steel designs, and 70%flat/30% extruded for aluminium designs. Table 4.1-2: Bill of Materials in kg Average Passenger Car/Light (PC/L) Alum FSV1 Passenger Car/Compact (PC/C) Average Alum Flat carbon steel 373 148 148 148 464 198 198 198 Long steel 140 115 115 115 174 144 144 144 Cast iron 93 93 93 93 116 116 116 116 Flat AHSS 0 198 0 176 0 233 0 195 Long AHSS 0 22 0 20 0 26 0 22 Rolled aluminum 9 9 130 9 12 12 154 12 Extruded aluminum 9 9 61 9 12 12 73 12 Cast aluminum 47 47 47 47 58 58 58 58 Other 350 350 350 350 414 414 414 414 Plastic 112 112 112 112 139 139 139 139 Rubber 28 28 28 28 35 35 35 35 Glass 28 28 28 28 35 35 35 35 Copper 19 19 19 19 23 23 23 23 Other 75 75 75 75 93 93 93 93 Fluids 29 29 29 29 29 29 29 29 Tires 60 60 60 60 60 60 60 60 Vehicle mass 1022 992 945 968 1249 1213 1158 1170 Other key parameters include: Recycling methodology in accordance with the Declaration by the Metals Industry on Recycling Principles 10, the avoided burden method, in which credit is given for producing material (scrap) that allows a downstream user to avoid production of primary material, was used. Power train - for purposes of determining the use phase impacts, a conventional gasoline powertrain has been assumed. Lifetime Driving Distance (LTDD) A2Mac1 database includes cars from all OEMs, and because automotive GHG modeling is very sensitive to this parameter, results were calculated using both European (150,000 km) and North American (250,000 km) averages for LTDD. Powertrain resizing because the model is also very sensitive to the decision whether or not to resize the powertrain to take full advantage of mass reduction, results have been calculated both with and without resizing. Secondary mass change as the mass differences involved in this study are relatively small (in all cases <100 kg), no secondary mass effects have been considered. Driving cycle the New European Driving Cycle (NEDC) was used. Fuel Consumption the UCSB model relies on baseline fuel consumption and weight elasticity values (WEV) developed by Forschungsgesellschaft Kraftfahrwesen mbh Aachen (fka) 11. For purposes of this case study, the baseline fuel consumption and WEV for the compact class (NEDC driving cycle) was used. This WEV equates to a fuel reduction value (FRV) of.102 l/100kg/100km FSV2 WorldAuto 2015 30

when the powertrain is not resized, and.282 l/100kg/100km when the powertrain is resized. The compact class baseline fuel consumption given by fka is 5.56 l/100km. 4.2 Body Structure Results 4.2.1 Total Life Cycle GHG Emissions The results (Table 4.2-1) show that, for all eight scenarios studied, the efficient steel design yields the lowest life cycle GHG emissions, with GHG savings over the baseline average steel designs of 193 to 798 kg CO2e. The FSV designs show the potential for an additional 118 to 721 kg GHG reduction. The possibility for unintended consequences is apparent from the results of the aluminium designs. In all cases, except the two that combine the 250,000 km driving distance with optimum powertrain resizing, the aluminium design, while producing the lowest use phase emissions, shows a net increase in GHG emissions over the existing average steel design. Even in the two scenarios for which the aluminium design shows life cycle GHG emissions lower than the baseline, it is clear that in order to minimize emissions, an efficient steel design is the right choice. Table 4.2-1: Relative GHG Emissions in kgco 2 e Body Structure No Resizing Average (baseline) Passenger Car/Light (PC/L) Alum FSV1 Average (baseline) Passenger Car/Compact (PC/C) Alum FSV2 150000 - -190 623-337 - -224 736-488 250000 - -277 403-492 - -327 476-711 With 150000 - -420 42-745 - -496 50-1,078 Resizing 250000 - -661-565 -1,172 - -779-667 -1,695 The results (Table 4.2-1) show that, for all eight scenarios studied, the efficient steel design yields the lowest life cycle GHG emissions, with GHG savings over the baseline average steel designs of 193 to 798 kg CO2e. The FSV designs show the potential for an additional 118 to 721 kg GHG reduction. The possibility for unintended consequences is apparent from the results of the aluminium designs. In all cases, except the two that combine the 250,000 km driving distance with optimum powertrain resizing, the aluminium design, while producing the lowest use phase emissions, shows a net increase in GHG emissions over the existing average steel design. Even in the two scenarios for which the aluminium design shows life cycle GHG emissions lower than the baseline, it is clear that in order to minimize emissions, an efficient steel design is the right choice. 4.2.2 Body Structure GHG Emissions by Life Cycle Phase GHG emissions for each combination of class and driving distance are displayed by life cycle phase in Figures 4.2-1 through 4.2-4. For all eight cases studied, the efficient steel and FSV designs show a consistent pattern: Lower production phase emissions due to the reduced amount of material required. Lower use phase emissions due to reduced mass of the vehicle. Slightly higher EOL emissions due to smaller credit for recycling because less material goes into the vehicle, less EOL scrap is available for downstream recycling. The effect of this pattern is, as outlined above, lower total emissions. The slightly higher EOL impact is outweighed by savings in the production and use phases. The two aluminium designs show a different, but still consistent, pattern: WorldAuto 2015 31

Significantly higher production phase emissions due to energy-intensive aluminium production. Lower use phase emissions due to reduced mass of the vehicle. Significantly lower EOL emissions due to larger credit for recycling recycling credit is based on the difference between primary and secondary material production, and for aluminium this difference is relatively high. The effect of this pattern is higher overall emissions, except for the two cases that assume both the 250,000 km LTDD and optimal resizing of the powertrain (Figures 4.2-1 and 4.2-2). The lower use and EOL phase emissions are outweighed by the increase in the production phase. Average (baseline) Without Powertrain Resizing Aluminium FSV1 Average (baseline) With Powertrain Resizing Aluminium FSV1 Production 0-101 2519-180 0-101 2519-180 Use 0-131 -330-232 0-361 -911-640 End of Life 0 42-1566 74 0 42-1566 74 Total 0-190 623-337 0-420 42-745 Figure 4.2-1: Relative GHG Emissions by driving distance- PC/L, 150,000 km WorldAuto 2015 32

Average (baseline) Without Powertrain Resizing Aluminium FSV1 Average (baseline) With Powertrain Resizing Aluminium Production 0-101 2519-180 0-101 2519-180 Use 0-218 -550-386 0-601 -1519-1066 End of Life 0 42-1566 74 0 42-1566 74 Total 0-277 403-492 0-661 -565-1172 Figure 4.2-2: Relative GHG Emissions by driving distance - PC/L, 250,000 km FSV1 Average (baseline) Without Powertrain Resizing Aluminium FSV1 Average (baseline) With Powertrain Resizing Aluminium FSV1 Production 0-120 2973-260 0-120 2973-260 Use 0-154 -390-335 0-425 -1075-925 End of Life 0 49-1848 107 0 49-1848 107 Total 0-224 736-488 0-496 50-1078 Figure 4.2-3: Relative GHG Emissions by driving distance- PC/C, 150,000 km WorldAuto 2015 33

Average (baseline) Without Powertrain Resizing Aluminium FSV1 Average (baseline) With Powertrain Resizing Aluminium FSV1 Production 0-120 2973-260 0-120 2973-260 Use 0-257 -649-559 0-709 -1792-1542 End of Life 0 49-1848 107 0 49-1848 107 Total 0-327 476-711 0-779 -667-1695 4.3 Other Systems Figure 4.2-4: Relative GHG Emissions by driving distance - PC/C, 250,000 km Six additional subsystems described in the A2Mac1 database were analyzed in the same manner: Front Bumper, Rear Bumper, Wheels (4 wheels), Hatchback, Hood, and Front Door (2 doors). Because of the relatively small mass of these subsystems, no powertrain resizing was considered. Only the three categories (average steel, efficient steel, efficient aluminum) from the A2Mac1 database were considered, since the comparison efficient steel vehicle, FSV, did not include designs for the other subsystems. 4.3.1 Subsystem Masses Table 4.3-1 : Subsystem Masses (kg) Average (baseline) Passenger Car/Light (PC/L) Aluminum Passenger Car/Compact (PC/C) Average (baseline) Aluminum Front Bumper 4.5 2.4 2.1 5.4 2.9 2.5 Rear Bumper 4.4 1.9 1.5 5.3 2.3 1.8 Wheels 35.1 27.5 24.2 41.7 32.7 28.8 Hatchback 11.4 8.7 7.6 12.2 9.2 8.1 Hood 8.8 6.8 4.5 11.3 8.8 5.8 Front Door 30.9 24.6 19.2 32.6 25.9 20.3 WorldAuto 2015 34

4.3.2 LCA Results The LCA results in Table 4.3-2 show a trend similar to that of the body structure analysis. In 14 of the 24 scenarios studied, the efficient aluminum design shows the unintended consequence of higher life cycle emissions than the baseline design. And, in all scenarios, the efficient aluminium design shows higher life cycle emissions than the efficient steel design. As in the case of the body structure, the most efficient steel design shows the lowest life cycle emissions in all cases. Table 4.3-2 : Relative Difference in Total Life Cycle GHG Emissions in kg CO 2 e - Subsystems Front Bumper Avg (baseline) Passenger Car/Light (PC/L) Aluminum Passenger Car/Compact (PC/C) Avg (baseline) Aluminum 150000 0-13 0 0-16 0 250000 0-19 -7 0-23 -8 Rear Bumper Wheels Hatchback Hood Front Door 150000 0-16 -8 0-19 -9 250000 0-23 -16 0-27 -19 150000 0-48 101 0-57 121 250000 0-69 71 0-82 84 150000 0-17 29 0-19 31 250000 0-25 18 0-27 19 150000 0-13 4 0-16 6 250000 0-18 -8 0-24 -10 150000 0-40 61 0-42 64 250000 0-58 27 0-61 29 WorldAuto 2015 35

5.0 Conclusions The challenges of reduced tailpipe emission standards are driving the industry to find ways to reduce vehicle mass. A study commissioned by WorldAuto and the Market Development Institute (SMDI) sought to analyze mass and material data for a variety of automotive components/subsystems from a database constructed and owned by A2Mac1, a global automotive benchmarking company, to understand the current state of lightweighting in production vehicles and identify further opportunities. This study helps identify gaps in the process of lightweighting vehicles by evaluating current production vehicles. Moreover, this study demonstrates the power of statistical benchmarking and the limitations of a one-off tear down of an efficient vehicle. Statistical benchmarking gives an entirely new perspective, a positive breakthrough in vehicle mass comparison and target setting. Using this methodology, an automotive designer can look at subsystems from a large body of vehicle data and identify those that are much lighter than the average vehicle. It provides a means to set subsystem targets on a more accurate basis than that which is being accomplished in the industry today. It creates a better road to holistic design and real progress in mass efficiency. The study revealed five key findings. 1. Today s steel designs vary drastically in efficiency. 2. When compared to an efficient steel design, the mass savings gap with aluminium significantly reduces. 3. Mass savings achieved at the component level are often not realized at the system level. 4. There is a narrow margin in curb weights between vehicles using efficient steel body structures and aluminium body structures. 5. Based on these four previous findings, it can be understood that there is yet untapped mass savings potential for steel. The data shown in this report for current production vehicles illustrates that there is disconnect between the level of lightweighting efforts in the automotive industry and the results actually achieved. It also shows that there is a great deal more opportunity to achieve this lightweighting with design optimization, advanced steels and steel technologies. Considering the vehicle curb weight savings gap of just 2.8% between aluminium and efficient steel designs, the cost incurred with aluminium materials is not justifiable. Most importantly, from a total life cycle perspective, the mass savings achieved by aluminium in current production vehicles is not resulting in a smaller vehicle emissions footprint overall. For all cases studied, the efficient steel designs show a consistent pattern of lower emissions in production, use and EOL, which results in lower total cycle emissions. The aluminium designs showed higher production phase emissions due to the energy-intensive manufacturing process, which is offset neither by the reduced use phase emissions nor the significantly lower EOL due to the larger recycling credit. It is important to consider the impact of any light weighting technology on the total mass of the system assembly. Each light weighting technology requires unique methods of assembly and localized mounting requirements for other components. Additional mass may need to be added to accommodate such WorldAuto 2015 36

requirements. Often additional consumer features are added to the product, offsetting the structure mass saving. Regulations based on tailpipe emissions drive the use of low density materials, such as aluminium, to achieve fleet average fuel consumption targets. However, as these challenges are faced during the vehicle design process, the value of the lightweighting technology must be properly weighed against cost and life cycle emissions. 5.1 Final Observations on the Power of Statistical Benchmarking as a Tool for Mass Efficiency As demonstrated by these study results, the following observations can be applied to the next generation of automotive design to ensure that future vehicles are safe, environmentally efficient, and remain affordable to manufacture: Statistical mass benchmarking is critical, as demonstrated in this study, to ensure that subsystem mass targets are accurately set. WorldAuto has made available, through its member companies, an Auto Mass Benchmarking Calculator that uses the same regression models and data employed in this study to enable quick comparisons of new or existing designs to the A2Mac1 vehicle database study results. Contact your member company representative for more information. Statistical mass benchmarking is also important to ensure that the value of the lightweighting option in terms of mass reduction validates the extra cost involved in using alternative materials. The question should be raised: Could a more efficient, holistic design provide the same or better mass reduction advantages, at less cost than an alternative material? It is crucial that mass reduction goals achieved at one place along the design chain are supported throughout the system design to ensure that the savings are maintained and optimized in the final subsystem. A Life Cycle Assessment should be conducted to ensure that compromises made to achieve lower tailpipe emissions do not result in a net increase in total life cycle emissions. Continuing on the track of focusing only on tailpipe emissions will result in an automotive sector whose efforts and expense to reduce vehicle environmental footprint may result in the unintended consequence of a net increase for future generations to address. WorldAuto members can make the full data results of this Auto Mass Benchmarking project available to customers, and cooperate with them to address mass efficiency in vehicle designs. As this project proceeds to further data updates and analysis in future phases, we will continue to make data and findings freely available at worldautosteel.org. WorldAuto 2015 37

Annotations 1 Malen, Donald E., Hughes, Jason, Mass Benchmarking Using Statistical Methods Applied to Automotive Closures, Paper No. 2015-01-0574, (April 2015) 2 FutureVehicle (May 2011, May 2013), available at www.worldautosteel.org 3 Davies, Jim; Singh, Harry, Automotive Front Door Benchmarking (Light Weight Door Structure), (2013) unpublished. 4 Geyer, Roland: The Example of Mild, Advanced High Strength and Aluminium in Body in White Applications Methodology Report (December 2007). The Methodology Report and a free download of the UCSB Automotive Greenhouse Gas Materials Comparison Model are available at http://www.worldautosteel.org/projects/vehicle-lca-study/assessments-of-automotive-material/ 5 Study Examples cited: a. U. S. Body-In-White (2013) b. Kim, Jaehyun; Lee, Hongwoo; Chung, Kyunghwan; Lee, Hyounyoung; Kang, Yeonsik; Nam, Jaebok, POSCO, Republic of Korea, A New Body Concept For Electric Vehicle: PBC-EV, FISITA Paper No. F2012-E04-020 (November, 2012) c. Hoffman, Oliver G, InCar The Innovative Solution Kit for the Automotive Industry (October 2009) d. ArcelorMittal, ArcelorMittal unveils new ultra lightweight car door solutions offering up to 34 percent weight savings over existing steel car doors (news release 25 June 2013) 6 Kelkar et al, Automobile Bodies: Can Aluminium Be an Economical Alternative to? (August 2001 Issue of JOM., 53 (8) (2001) pp. 28-32). 7 Singh, Harry, Mass Reduction for Light-Duty Vehicle Models Years 2017 2025 Final Report, NHTSA Report DOT HS 811666 (http://www.nhtsa.gov), U. S. Summary Analysis (August 2012,). 8 Polk, R.L. & Company, Length of U.S. Vehicle Ownership Hits Record High, (Feb 2012), https://www.polk.com/knowledge/polk_views/length_of_u.s._vehicle_ownership_hits_record_high 9 Ricardo AEA, Improving the understanding of the potential for weight reduction in cars and vans (May 2014) 10 AISI, et al., Declaration by the Metals Industry on Recycling Principles, International Journal of Life Cycle Assessment, 2006 11 fka, Wohlecker, Roland, et al., Determination of Weight Elasticity of Fuel Economy for Conventional ICE Vehicles, Hybrid Vehicles and Fuel Cell Vehicles, fka, Report 55510, June 2007. WorldAuto 2015 38

Appendix 1 Model Equations The table of results below (Table A1-1) summarizes the mass drivers for each of the subsystems and shows the power model predictive equations for the average and the mass efficient designs. The accuracy/quality of the predicted mass by these equations is indicated by the R² value calculated by the regression analysis. An R² value of 0.6 indicates that 60% variation in the data point values is accounted for by the chosen variables/attributes in the equation. For R² values less than 0.5 the mass efficient options identified should be further reviewed and engineering judgment should be applied to identify the designs for further study. Comparison of the R² values for all the systems is shown in Table A1-2. When interpreting the results one should also understand that the A2Mac1 data base could have measurement and recording errors as the vehicle teardown process involve manual part disassembly, weighing and recording of mass and material type data. Some of the data for material type was corrected in this study, especially for the doors, where a previous SMDI study had identified some discrepancies. Table A1-1: Model Summary of Subsystem Mass Drivers A. Front Door Linear model: door frame mass estimate 1.06 4.205 13.57 5.10 0.00 0.47 0.00 Power model: frame mass estimate.. ²..... R 2 = 0.56, Standard Error = 1.123 Power model: Mass estimate for steel efficient design.. ²...... Power model: Mass estimate for aluminum efficient design.. ²...... N Values: 1.12 Aluminium 1.0 B. Hood Linear model: hood frame mass estimate 7.11 8.75 6.09 0.00 Power model: frame mass estimate.... R 2 = 0.80, Standard Error = 1.177 Power model: Mass estimate for steel efficient design WorldAuto 2015 39

..... Power model: Mass estimate for aluminum efficient design..... N Values: 1.2 Aluminium 1.0 C. Hatchback Linear model: hatchback frame mass estimate 0.35 4.67, 0.0042, 3.36 0.00 Power model: frame mass estimate. ².,... R 2 = 0.50, Standard Error = 1.138 Power model: mass estimate for steel efficient design. ².,.... Power model: mass estimate for aluminum efficient design. ².,.... N Values: 1.1 Aluminium.09 D. Decklid Linear model: decklid frame mass estimate 2.22 4.429, 2.71 1.07 0.00 0.00 1.32 0.00 Power model: frame mass estimate. ²....... R 2 = 0.50, Standard Error = 1.15 Power Model: Mass estimate for steel efficient design. ²........ Power Model: Mass estimate for aluminum efficient design. ²........ N Values: 1.05 Aluminium.09 WorldAuto 2015 40

E. Liftgate Linear model: liftgate frame mass estimate 9.08 3.02,. Power model: frame mass estimate. ². R 2 = 0.15, Standard Error = 1.194 Power model: mass estimate for steel efficient design. ²... Power model: mass estimate for aluminum efficient design. ²... N Values: 1.05 Aluminium 1.25 F. Instrument Panel Beam Linear model: IP Beam mass estimate 0.0 6.45 0.0079 3.376 Power Model: Beam mass estimate.... R 2 = 0.36, Standard Error = 1.310 G. Front Bumper Beam Linear model: bumper beam mass estimate 2.59 9.07 0.0040, 0.00743, 0.00 1.27 0.86 0.00 0.00 Power Model: Bumper beam mass estimate......... R 2 = 0.48, Standard Error = 1.346 Power Model: Mass estimate for steel efficient design..,........ Power Model: Mass estimate for aluminum efficient design..,....... N Values: 1.40 Aluminium 1.05 WorldAuto 2015 41

H. Rear Bumper Beam Linear model: bumper beam mass estimate 2.00 3.07 0.0046, 1.19 0.00 0.00 Power Model: Bumper beam mass estimate...... R 2 = 0.44, Standard Error = 1.439 Power Model: Mass estimate for steel efficient design....... Power Model: Mass estimate for aluminum-plastic efficient design....... N Values: 1.60 Aluminium/plastic 1.40 I. Wheel Rim Linear model: wheel rim mass estimate 12.20 0.0369, 0.0276, 1.31 0.00 Power model: wheel mass estimate.,.... R 2 = 0.56, Standard Error = 1.158 Power Model: Mass estimate for steel efficient design.,..... Power Model: Mass estimate for aluminum efficient design.,..... N Values: 1.1 Aluminium 1.12 WorldAuto 2015 42

J. Body Structure Linear model: body structure mass estimate 0.00 154.5 0.098, 21.37 ² 110.9 126.7 Body structure and engine cradle mass estimate... ²... R 2 = 0.87, Standard Error = 1.085 Mass estimate for steel efficient design.. ².... Mass estimate for aluminum efficient design.. ².... Mass estimate for FSV - AHSS design.. ².... N Values: 1.05 Aluminium 0.875 FSV-AHSS 1.12 Table A1-2: Comparison of R² Values Vehicle Subsystem Power Model R² EDAG 2013 Malen 2010 Front Door 0.56 0.45 Hood 0.80 0.77 Hatchback 0.50 0.31 Decklid 0.50 0.72 Liftgate 0.15 0.41 Wheel Rim 0.56 0.38 Rear Bumper 0.48 0.31 Front Bumper 0.44 0.49 Body Structure 0.87 0.83 IP 0.36 0.24 WorldAuto 2015 43