material selection approach to evaluate material substitution for minimizing the life cycle environmental impact of vehicles

Weight reduction is commonly adopted in vehicle design as a means for energy and emissions savings. However, selection of lightweight materials is often focused on performance characteristics, which may lead to sub optimizations of life cycle environmental impact. Therefore systematic material selection processes are needed that integrate weight optimization and environmental life cycle assessment. This paper presents such an approach and its application to design of an automotive component. Materials from the metal, hybrid and polymer families were assessed, along with a novel self-reinforced composite material that is a potential lightweight alternative to non-recyclable composites. It was shown that materials offering the highest weight saving potential offer limited life cycle environmental benefit due to energy demanding manufacturing. Selection of the preferable alternative is not a straightforward process since results may be sensitive to critical but uncertain aspects of the life cycle. Such aspects need to be evaluated to determine the actual benefits of lightweight design and to base material selection on more informed choices.


Introduction
The transport sector accounts for about one third of total energy demand in Europe and is among the major contributors of greenhouse gas emissions [1]. Road vehicles have a considerable share in this. To meet increased regulations and customer requirements for reduced environmental impact, the automotive industry seeks solutions to improve the performance of their fleet especially during the dominant operation stage. Weight reduction is commonly adopted in vehicle design, resulting in both energy and emissions savings [2,3]. Weight savings can be realised through material substitution to lightweight materials, such as lighter metals, polymers and composites [4,5], through use of materials in a more weight efficient manner, for instance as a sandwich structure [6], or through a combination of the two. Lightweight design does, however, have drawbacks. Compared to steel, composite materials such as carbon fibre reinforced polymers (CFRPs) are costly and have higher energy demand during manufacturing [5]. Recycling of composites and sandwich structures is also costly and complicated; since composites consist of two different materials (matrix and reinforcement) and traditional sandwich structures often consist of three (face sheet-, core material and adhesive), separation of the materials is difficult. A relatively new class of composite materials, self-reinforced polymers (SrPs), also called single-polymer or all-polymer, could combine weight savings with high recyclability at end-of-life (EOL), since fibres and matrix are based on the same recyclable polymer [7,8]. They also exhibit good mechanical properties [9,10] and may hence be attractive for automotive applications. To assess whether there are any significant unexpected environmental trade-offs over the life cycle, these novel materials need to be assessed using a life cycle perspective. Such an assessment should consider physical, mechanical, and environmental characteristics in an integrated manner [11] and be adopted in standard material selection and vehicle design processes. Such assessments are rarely applied in practice, as vehicle designers primarily consider performance and acquisition cost of the material [12] while environmental performance is rarely prioritised [13,14]. If performed, environmental assessments are usually done for finalized products, not as an integrated part of the design process [13,15]. For the automotive industry to meet the challenge of reducing the climate and other environmental impacts, a shift in the material selection paradigm is needed. Eco-design guidelines, e.g. ISO TR 14062 [22], suggest that environmental assessment should be performed in parallel to the traditional product design process and before any final design decision is made. Hence, comprehensive frameworks are needed which ensure that both functional and environmental vehicle performance requirements are considered and balanced. Although rarely applied in practice, several such integrated models for evaluating and selecting materials can be found in literature [3,5,[16][17][18][19][20]. Simoes et. al. [3] and Witik et al. [5] suggest a thorough environmental and cost analysis. However, design parameters and requirements are not presented as part of a comprehensive material selection model, limiting those studies to an environmental and cost assessment of finished products. Only a few complete frameworks for material selection in a design context based on structural optimization and life cycle environmental assessment techniques have been published [16,18,20]. Some of these are prescriptive, for instance concerning the use of preselected environmental indicators or decision making functions. While this may simplify application, there is also a risk that it prevents integration in a company's established material selection process.

Aim and scope of the paper
An integrated material selection approach is presented in which structural weight minimization is combined with environmental life cycle assessment (LCA), with the aim to allow for systematic evaluation of material alternatives before any final design decision, and to reduce the risk for suboptimizations and shift of environmental burdens along the different life cycle stages of the vehicle.
It builds on previous frameworks and expands their scope especially concerning environmental performance of the materials. Efforts were made to capture critical aspects in vehicle design. The approach was tested on a vehicle design case study. SrP composites were compared to commonly used materials for a particular vehicle component, in order to show how materials for vehicle design can be systematically assessed, but also to provide specific results regarding the environmental performance of SrP composites.

Life cycle based material selection
The integrated material selection approach was aligned to traditional material selection frameworks that consider material properties and structural weight optimization to derive feasible and weight efficient design alternatives [21]. It adds evaluation of life cycle environmental impact of all feasible material alternatives at an early stage. It consists of five major steps; Definition of design target, Selection of material families and candidate materials, Weight minimization, Life cycle modelling and assessment, and Results analysis and material selection (Fel! Hittar inte referenskälla.). Selection of weight minimization models, environmental analysis tools and impact indicators are to be decided by the users. Possible ways to analyse and interpret the results are suggested. Although cost is an important parameter in material selection today, it is not considered in this study.

Definition of design target; setting requirements and constraints
Design targets are defined by a number of functional (fundamental properties) and non-functional requirements according to the intended application. These requirements need to be fulfilled in a limited design space. A functional requirement of a car roof for instance, is to protect passengers from outdoor conditions and from accidents. In this material selection approach, low environmental impact is also considered a functional requirement of the design target as suggested by Deutz et. al. [23], in order to "extend the definition of functional requirement" from considering only the performance of the product to its life cycle performance. Additionally, properties related to the intended application (vehicle) may also have an influence and restrict the design target. The type of

Selection of material families and candidate materials
Engineering materials can be classified into six families: metals, ceramics, glasses, polymers, elastomers and hybrids [21]. The functions and requirements defined by the design target constrain the selection of certain material families [24]. A front window of a vehicle for instance should be transparent which obviously excludes the use of metals. From the remaining families, a more specific list of properties and constraints (such as application temperature, specific strength or stiffness, manufacturing constraints, regulation requirements etc.) will lead to only a few representative alternatives or material candidates that fulfil the design target [13,18,24]. Those materials represent feasible design solutions for the design target and will be used in the consecutive stages of weight minimization and life cycle assessment applied in this approach.

Weight minimization
Weight minimization is applied to all material candidates in order to derive the lowest optimal mass for the design target. For the weight minimization, constraints such as available space for the part or maximum allowed deformation etc. are defined based on the requirements of the design target. The general constraint optimization problem is formulated as: subject to , 1 … 5 f 0 is the weight function which is a function of the design variables x i . f k is the constraint function while b k represent the constraint values. and define the lower and upper bounds for the design variables x i . These bounds could be for instance the allowed wall thickness of a structure.

Life cycle models and assessment
Life cycle models of each candidate material are created and assessed according to [25]. The functional unit (reference measure of the assessment) [26] defines the function and operational life of the design target and remains constant for all design alternatives in order to provide comparable results. Environmental inventory data (inflows of energy and materials and outflows of emissions and waste) are collected for all stages and processes involved, either from generic material databases e.g. [27,28], suppliers, or manufacturing sites. The potential environmental impact of the design alternatives can be estimated with life cycle impact assessment (LCIA) methods [29] that quantify the impact of resource use and emissions [26]. The most appropriate LCIA method is determined by the availability of data, quality of the model, as well as the aspects that the designer wants to consider.
Different software tools are available to facilitate data collection and calculation procedures.

Results analysis and material selection
In order to fulfil the functional requirement of low life cycle environmental impact, a comprehensive analysis of the results is needed, which may consider the following aspects:

Case study: material selection for a truck roof panel
A material selection case study for a truck roof panel is used to test and illustrate the suggested approach. The case study was also used to evaluate the performance of SrP from a lightweight design and environmental perspective. Since this is a theoretical case, specific manufacturers and materials suppliers were not available when conducting the data inventory. The majority of data considered are based on generic databases and literature.

Definition of design target; setting requirements and constraints
The truck roof panel represents the design target of the material selection process. The roof is used by a 40t diesel truck and designed for a life cycle distance of 1 000 000 km (indicated as average life time mileage in [30] for heavy freight vehicles). The roof size is assumed to be 2.2m x 2m with a maximum allowed roof thickness of 100mm and it is designed to fulfil specified functional and safety requirements such as the truck cabin safety test standard ECR29 [31].

Selection of material families and candidate materials
Representative materials were selected from the material families which could fulfil the design target; metals, hybrids and polymers. Mechanical properties of the selected materials (assumed to be linear elastic up to yield stress) are presented in Table 1. Steel and aluminium are the most commonly used metals in the vehicle industry. To have the roof as light as possible, steel and aluminium alloys with good mechanical properties are needed.
Hence, a dual phase steel DP800 and an Al6061-T6 were selected. Hybrids, such as fibre reinforced polymer composites are also widely used in automotive applications. Commonly used hybrids are carbon or glass fibre reinforced vinyl-ester (C/VE and G/VE respectively). To obtain relatively high stiffness and strength, the volume fractions of fibres in the composite were assumed to be 0.58 vol.
% and 0.54 vol. % for the C/VE and and 50% high tenacity PET (fibre material) [10]. Due to the reinforcing fibres, SrPET composite material has higher mechanical properties than the base material PET, but is expected to have higher environmental impact than of base material PET. To investigate if this higher environmental impact is compensated by the lower weight of SrPET, an unreinforced PET was also included in the study.

Weight minimization
To use the material efficiently the truck roof was designed as a corrugated sandwich panel. To facilitate recycling the panel should consist of the same face sheet and core material. The geometry of the sandwich panel (Fel! Hittar inte referenskälla.) was defined by weight minimization for all materials.

Figure 2: a) Schematic of sandwich beam b) corrugated unit cell
The weight minimization problem was simplified to a corrugated sandwich beam with a length L=2.2m, simply supported edges and with a uniform distributed load q=22.3kN/m. Five design variables were used ( Table 2). Bounds were chosen to ensure a manufacturable sandwich structure while limiting the design space without influencing final result. The objective function f 0 is the beam . x 3 ω corrugation angle 10 -80 x 4 t w core member thickness 0.5 -15mm x 5 t c core height 10 -90mm Weight minimization was performed with a constrained nonlinear multivariable optimization function fmincon in Matlab R2012b. The Matlab global search algorithm MultiStart, which generated 10 000 uniformly distributed starting points within the bounds, was employed to search for a global minimum.

Constraints
The constraints considered were geometrical and stiffness constraints. The geometrical constraint was determined by the available physical space, which defines a maximum allowed sandwich structure thickness of 100mm (sum of face sheets and core thickness). The stiffness constraint defines a maximum roof deflection of 100mm. To ensure that the roof deflection is not exceeding the maximum, the sandwich structure is not allowed to soften. It was assumed that this structural softening appears when the material yields or buckles. This results in five more constraints.
Deflection of a sandwich beam is the sum of deflection due to shear w s (assumed to appear only in the core) and due to bending w b resulting in w=w s +w b =0.125qL 2 S -1 +5/385qL 4 D -1 100mm [6] where q is the distributed load, S and D are the shear stiffness and flexural rigidity of the sandwich structure respectively. S= G xz d 2 t c -1 where d=t fc /2+t c +t ft /2 is the distance between the centroids of the faces and core shear modulus G xz =0.5Et w L cm sin(2) for corrugated core defined by [35] where L cm =t c /sin(ω) is the core member length. The flexural rigidity D of the sandwich structure is defined as: where the core modulus E xx =Et w cos 3 ()t c -1 [35] and the position of the neutral axis is * (5).
The material failure constraint for the top face sheet (compression stress) for materials with linear behaviour is: where M x =0.125qL 2 is the maximum bending moment in the beam. Similarly, material failure for the bottom face sheet (tensile stress) is: In the sandwich core, the core members are loaded in tension and compression and therefore the core member material constraint is: where T x =0.5qL is the transverse force in the sandwich beam. The core members and the top face sheet are loaded in compression and can fail by buckling. The buckling constraint for the core member is (assuming pinned ends): where L cm =t c sin -1 (ω) is the core member length. For the top face sheet, the buckling constraint is (assuming pinned ends): where L critf =2t c tan -1 (ω) is the length of the top face sheet.

Weight minimization results
Results obtained from weight minimization are listed in Table 3. The table also shows weight savings compared to steel which can be considered as the reference material for the truck roof. As expected, the highest weight saving was achieved by C/VE. Unreinforced PET resulted in a 23% heavier roof.

Life cycle models and assessment
Life cycle models were built of the truck roof, for each material candidate ( Figure 3). The functional unit is "one truck roof" with design specifications according to the design target and a life cycle distance of the truck of 1 000 000 km. Results from the weight minimization provided the total mass of the roof and the amount of each material needed. The geographical boundary is Europe, meaning that inventory data was representative for Europe. Transportation of materials and parts were excluded due to lack of data and expected minor impact on results. LCA models were built using the SimaPro 7.3 software [36] and environmental inventory data from the ELCD v2.0 and Ecoinvent v2.0 and databases [27,28]. Other sources of information, mainly academic publications, were also used.
A summary of assumptions is presented below. Details on the process data can be found in the supplementary material of this paper.

Production of raw materials and truck roof manufacturing
Data on raw material acquisition and production were obtained from databases [27,28]. Production of carbon fibres (CF) was modelled according to [37]. Manufacturing processes for the truck roof were modelled based on the properties of the selected materials. Metal structures go through a sheet rolling and cutting process. To join the structure the solid -state joining process friction stir welding was assumed. The CF and GF composite corrugated structures were manufactured by a 3D fibre weaving process. The woven CF and GF structures were then infused with the vinyl ester resin.
For the SrPET sandwich structure, a commingled yarn of PET fibres and PET matrix was used which was consolidated by a hot compaction process. Finally, the unreinforced PET structure was assumed to be manufactured by a PET injection moulding process. Energy requirements of the manufacturing processes are presented in Table 4. Electricity was modelled as average European mix [27,28].
Waste from cutting of polymer and composite structures was assumed to be 10% of the structures' weight. Thermoplastic polymer process waste was recycled and C/VE and G/VE process waste was assumed to be incinerated in municipal waste incineration facilities, modelled based on [43].

Use phase
The use phase was modelled as operation of a diesel truck of EURO 5 emission standard [44]. Energy use and emissions during the reference life time distance of 1 000 000 km were allocated to the truck roof assuming a linear relationship between weight and fuel consumption, resulting in a fuel consumption of 7.5 10 -6 l km -1 kg -1 material (which is close to the values estimated by [2]).

End of life (EOL)
EOL treatment is determined by the properties of the materials and available waste management technologies. Recycling of metals from end of life vehicles (ELVs) is well-established with high material recovery rates [5,45]. Polymers and composites usually end up in the automotive shredder residues (ASR) that is either incinerated or landfilled [46,47]. However, there are large geographical differences and uncertainties about future practice at EOL. The uncertainties are explored by analysing different treatment alternatives ( Table 5). As a reference case, recycling (R) of metals and thermoplastic materials and incineration (I) of thermoset composites were assumed, representing the "best case" of technologies available today. Incineration of all structures (apart from metals) was modelled as a moderate case, and landfill (L) of all structures as a worst case.  L L L L L L 95% recycling efficiency was assumed, while remaining material was landfilled [45]. Benefits of recovered materials were included as avoided burdens, assuming replacement of production of equivalent primary materials. Waste incineration was assumed to produce electricity only (no heat).
Although heat recovery from waste incineration is significant in some regions, it is low in the EU as a whole [48]. Electricity production was calculated based on the thermal content of the materials (Table 6) and an average 20% energy efficiency [49]. Benefits of recovered electricity were also included as avoided burdens, using average European electricity mix [28] as the avoided electricity.

Environmental impact assessment
Impact assessment was done using the methods Cumulative Energy Demand (CED) v. 1.08 [29,51] (unit: MJ), and Global Warming Potential (GWP) [29,52] (unit: carbon dioxide equivalents, CO 2 eq.), as implemented in SimaPro [36]. The assessment was limited to these two methods due limitations in the inventory of life cycle data.  results in a negative value for CED and GWP for the metal and polymer structures due to recycling and recovery of materials. Similarly, electricity recovered from incineration of composite structures leads to avoided impact of CED, while no net savings of GWP are obtained when EU average electricity is the avoided electricity. During manufacturing the most significant processes for both CED and GWP is the acquisition and production of raw materials. Steel is the least energy demanding raw material ( Table 7). The lighter materials (aluminium, C/VE and SrPET) use more energy during acquisition and production, thus resulting in greater emissions. SrPET results in slightly higher CED and GWP, despite its lower weight compared to unreinforced PET, since additional manufacturing steps are required. Manufacturing the roof per se represents a lower share of the impact during this stage. Processing of polymer and composite structures is more energy demanding thus resulting in higher GWP compared to metals.

Analysing variations during the life cycle of the design target
The environmental assessment entails uncertainties since not all stages of the life cycle are known at this stage in the design process. Trucks or parts of a truck are often used longer than they are designed for e.g. as spare parts for other vehicles [53]. This would increase the operational life of the component, but may also influence how the component will be treated at EOL. In addition, lightweight design is rarely the only strategy for environmental improvements is vehicle design [5].
More efficient engines are developed and alternative fuels may be used. Such uncertainties may influence results and were investigated with sensitivity analyses.

Variations in energy efficiency or fuel type
More energy efficient engines would reduce the relative importance of the use phase impacts for all studied structures. In Figure 5 this would reduce the gradient of the plotted lines, meaning that all breakeven points between materials would move to larger distance, i.e. the benefits of light-weight materials would be less profound. Use of renewable fuels has a similar influence on GWP. According to the European Renewable Energy Directive (RED) [54], renewable fuels are those that lead to at least 60% reductions of greenhouse gas emissions, and the life cycle GWP impact is significantly reduced. However, more energy demanding fuel production and lower fuel efficiency of ethanol [55] increases the influence of weight on CED, shifting the breakeven points in between materials Figure 5 to an shorter travelled distance. In contrast, the much lower GWP of the ethanol truck significantly reduces the benefit of light-weight materials; high GWP in manufacturing would not be compensated

Variations during EOL
Two EOL alternatives were considered in addition to the reference case (Table 5). EOL remains the stage with the lowest contribution in terms of environmental impact, although the life cycle impact of the design alternatives vary between the best and worst EOL alternatives. Recycling of the aluminium, PET and SrPET structures has the potential to save significant amounts of energy and emissions. Such savings were less significant for PET and SrPET when the moderate alternative with incineration was assumed. If all structures are landfilled, total environmental impact increases for all design alternatives. For the metals and thermoplastic structure this increase is more significant since they lose the potential for material recovery. Weight reduction in that case has a significant influence on the results, so that the light C/VE structure would be the best alternative followed by the G/VE and aluminium. These structures are among the ones with the highest impact from manufacturing, but this is compensated earlier than 400 000 km when compared to the heavier but less energy and emission intense steel structure.

Discussion and conclusions
The need for a systematic comparison during material selection and substitution in order reduce life cycle environmental impact of products has been stressed by many [3,5,18]. This paper presents an integrated material selection approach for vehicle design that combines weight minimization and environmental life cycle assessment and its application to a hypothetical material selection case study for a truck roof. The case study also served to assess the environmental performance of SrPET, The study showed that weight minimization and LCA can be performed in a systematic and integrated manner during the design and material selection process. The suggested approach assists in moving from (partly) ad-hoc processes often seen today [13,15,23], to more informed and systematic decision making processes. This reduces the risk for sub-optimizations and shift of the environmental burdens along the different life cycle stages of the product. Material selection however, is rarely a straightforward process. In practice, the selection of the best alternative depends on more parameters in addition to environmental impact and weight; leading to a multicriteria assessment [16,18]. Designers must define decision criteria and handle trade-offs between the different parameters. For the suggested approach to be widely implemented in practice, the necessary information, tools, and expertise need to be in place. This requires multidisciplinary groups including among others designers, material specialists and environmental experts, but also procurement or sales representatives. Such groups exist in vehicle manufacturing companies but rarely work in an integrated manner during material selection [13]. A systematic procedure may facilitate communication and collaboration among different groups during product development, the lack of which was often seen as barrier for implementation of eco-design practices [57,58].