Engineering Insights: Composites Lessen Load in Fuel-Cell Demonstrator

Fuel-cell systems might eventually revolutionize propulsion for ground transportation with energy-efficient power that produces little or no pollution. However, significant technical challenges must be overcome before fuel-cell vehicles (FCVs) are cost-competitive with internal-combustion, battery-electric, and hybrid-electric vehicles. Automakers are using small fleets of demonstration vehicles to work out these kinks in the hope of being first to market with commercial FCVs. Ford Motor Co. (Dearborn, Mich.) was among the first to put a fleet of FCVs on the road. In 2005, 30 units — heavily modified versions of Ford’s production compact, the Focus — were deployed to city governments and research organizations for real-world testing.

One challenge is offsetting the higher mass of the fuel-cell and hydrogen storage systems (approaching 400 kg/900 lb) to meet vehicle weight targets. To achieve the desired curb weight of 1,600 kg/3,527 lb for the Focus FCV, Ford responded with an aggressive lightweight-ing strategy that included aluminum hoods, front fenders, wheels and suspension com-ponents; lighter steel body panels; titanium suspension springs; lightweight front and rear glass; polycarbonate side windows; and a carbon/aramid composite underbody shield and fuel-cell cover. Ford also replaced the steel decklid with on of lighter and cost-effective carbon fiber-reinforced plastic (CFRP).

THE ROAD TO CARBON FIBER

Before the Focus FCV, automotive CFRP was used primarily on sports cars and supercars, usually driven by the need for aggressive styling, low mass (to boost performance), and tooling cost reductions on these low-to-moderate volume vehicles. Although styling and speed were unimportant for the Focus FCV, mass was critical to optimizing driving range and cost savings that, if they could be reaped, would help offset the investment for the fuel-cell system.

Ford asked Multimatic Inc. (Markham, Ontario, Canada), an experienced automotive body/chassis systems supplier, to design and produce several components for the FCV, including the decklid. Goals for the latter were that it be produced in low volume at reasonable cost, yet meet all OEM performance and appearance requirements at less than 4.75 kg/10.5 lb — 56 percent less mass than the production steel part. While the decklid could be altered, design changes would be limited by the fact that the exterior styling surfaces were to remain essentially unchanged and hardware and trim components from the steel decklid would carry over into the new component.

The benchmark was a multipiece steel decklid weighing 10.8 kg/23.8 lb and measuring 430 mm long by 1,260 mm wide by 450 mm high (17 inches by 50 inches by 18 inches). Components targeted for conversion included the Class A outer panel, the inner panel, and local reinforcements for latch, hinges and rubber bumpers that mounted on the inner panel. Carryover items included the wire harness and clips; emergency release and fasteners; latch hardware and fasteners; bumper stops; hinges and reinforcements; and the interior trunk liner, all of which attached to the inner panel and therefore would impact its design. Additionally, the outer panel provided attachments for the carryover stop lamp, lid handle and exterior trim.

Initially, Ford considered the decklid in aluminum because, at that time, other closure panels were being converted to aluminum, which could be stamped using the steel part’s legacy production tooling. However, the steel decklid’s tools did not reproduce its complex geometry properly in aluminum. The cost of new tools could not be amortized over 30 vehicles. Given the design criteria, composites were the logical next choice.

DESIGNING WITHIN CONSTRAINTS

The panel geometry was carried over from the steel design to maintain functionality. This permitted bonding of the latch reinforcement to both inner and outer panels. The design also accepts the stock trunk liner, although it was subsequently eliminated. While the preliminary composite design was similar to the steel decklid, several factors prompted alterations. A hydrogen vent duct was added to the outer panel to bleed off excess pressure from the hydrogen-storage tank. This change altered the local geometry of the inner panel’s seal plane as well. Additionally, the inner panel geometry was greatly simplified to aid manufacturing feasibility of the composite design.

The Class A outer panel was designed to be structural. It combines solid laminate with honeycomb-core sandwich construction in selected regions to improve stiffness and strength while reducing mass. Because the inner and outer panel would be bonded along the outer perimeter and in the interior region to meet structural stiffness and strength targets, the interior bond line was positioned in the cored regions of the outer panel to avoid read-through.

An early design goal was to find autoclave-curable materials that could provide a Class A surface with minimal hand finishing on the decklid’s complex geometry. Under consideration were carbon fiber prepreg fabrics in differ-ent resin systems; carbon fiber semipregs; a range of core types and densities; various surface veils; several in-mold and conventional primers; and a variety of surface-preparation methods. Based on manufacturability and OEM paint-plaque sample evaluations, two systems were selected: For the solid laminate and the sandwich skins in cored sections, Multimatic selected Advanced Composites Group Ltd.’s (ACG, Heanor, Derbyshire, U.K.) ACG MTM49-3/CF0301, a 200 g/m² (5.8 oz/yd²) carbon fiber/epoxy fabric prepreg with 0.21-mm/0.0083-inch ply thickness. For cored areas, Multimatic chose Euro-Composites Corp.’s (Echternach, Luxembourg) ECA 3.2-48 aramid honeycomb core, which has a 3.2-mm/0.13-inch cell size and a density of 48 kg/m³ (3.0 lb/ft³) at 6.35-mm/0.25-inch thickness. To prevent galvanic corrosion anywhere the CFRP otherwise would be in contact with steel (i.e., the hinge reinforcements and latch hardware), ACG’s 300-g/m² (8.9-oz/yd²) fiber-glass/epoxy fabric prepreg (ACG MTM49-3/GF0103 at 0.23-mm/0.0091-inch ply thickness) was used as a barrier layer. To bond CFRP components, Multimatic used 3M Co.’s (St. Paul, Minn.) DP-460 epoxy paste while ACG’s film adhesive — ACG MTA 263, at 189 g/m² (5.5 oz/yd²) — was used to join core to skins.

To reduce weight and offer additional galvanic-corrosion protection, Multimatic also opted for unique, lighter weight, e-coated mild-steel hinge reinforcements.

When materials selection was complete, detailed design began. This involved determining the shape of the inner panel and wiring package; engineering the attachment and interface for all mating components; and incorporating the new vent and carryover hardware. Because composite panels are usually thicker than steel counterparts, thickness differences between the baseline steel and composite materials and the design of the inner panel required close attention to ensure correct fit of carryover components. Maximum allowable thickness for each mating hardware and trim component was defined and a workable solution determined: In some cases, local panel thickness was decreased to accommodate hardware while, in other cases, fasteners or clips compensated for increases in material thickness.

Multimatic engineers completed a CAD model in I-DEAS software, now marketed by Siemens Automation & Drives (Plano, Texas), and then a detailed finite-element model was built and CAE analyses were performed in ABAQUS/Standard software (Dassault Systèmes, Providence, R.I.) to assess structural performance under six previously identified design validation load cases. All components were modeled in fully integrated shell elements, using the nonlinear analysis option. The code’s composite shell section definitions were used to define composite layups, and the honeycomb core was assumed to be a ply within the composite-shell section to evaluate various core thicknesses. Local coordinate systems were applied to define orientation for the orthotropic composite materials. All decklid components were assumed to be bonded together with a structural epoxy adhesive; bonding was simulated using the model’s rigid-beam multipoint constraints. Material properties were assumed to be linear elastic, so material damage was not simulated in the analysis. Permanent set was evaluated with physical testing later in the program.

Multiple laminate configurations were assessed to find a design that minimized mass yet met Ford’s require-ments. Analysis indicated two dominant load cases would drive the design: front-corner deflection and waterfall deflection. While each load case had different deflection contours, both cases caused local bending in the waterfall region of the decklid (the area where the styling line cascades downward from a horizontal shelf). Increasing section size in the target regions was not an option because of the carryover interior sealing surface. Instead, material thickness was increased and a combination of 0°/90° and ±45° ply angles was used to satisfy the deflection requirement.

Decklid thickness varied from 1.1 mm to 2.7 mm (0.04 to 0.11 inches) in the solid laminate regions, while the sandwich laminate was constant at 7.2 mm/0.28 inch. The predicted mass (including hardware and adhesive but not primer/paint) was 3.9 kg/8.6 lb.

To minimize program investment, prototype and production parts were produced on the same tooling with only minor modifications for hole and slot locations. All decklid components were manufactured on single-sided tooling using a vacuum bag/autoclave process. Inner and outer panel tools were carbon fiber composites (produced by Multimatic with ACG’s LTM12 carbon tooling prepreg from epoxy patterns made with ACG TB650 epoxy tooling block), while the latch reinforcement and vent tools were aluminum to facilitate late design changes. Prepreg and core materials were NC-cut on a Gerber Technology (Tolland, Calif.) automated ply cutter and organized into individual kits to facilitate workflow. All decklid components were laminated according to the layup scheme defined by analysis. To minimize postcure finishing, outer panel tooling was coated before lamination with Duratec in-mold vinyl ester primer from Hawkeye Industries Inc. (Marietta, Ga.). After autoclave cure, components were trimmed using a combination of router fixtures, a 5-axis waterjet supplied by Flow Automation (Burlington, Ontario, Canada, a wholly owned subsidiary of Flow International, Kent, Wash.), and a BM1600 vertical machining center from AWEA Mechantronic Co. Ltd. (Hsipu, Taiwan).

Finished CFRP components were abraded, wiped with acetone in the bond areas and bonded together in a fixture with ambient-cure 3M paste adhesive. Multimatic maintained a minimum bond gap by mixing 0.25-mm/0.010-inch solid-glass microspheres into the adhesive. Assembled decklids underwent 100 percent surface and detail inspections with a Mitutoyo Corp. (Kawasaki, Japan) BN715 coordinate measuring machine to verify dimensional accuracy.

Prototype components were used to develop the manufacturing process, verify DV performance, develop all wire routing, evaluate fit and finish, and determine vehicle-level performance. When all production part approval process (PPAP) requirements were met, primed decklid assemblies were delivered to Ford for topcoat paint and final trimming (locks, lights, decals, etc.) and installation in the vehicle.

The composite decklid’s as-molded mass of 4.3 kg/9.5 lb and primed/painted mass of 4.5 kg/9.9 lb were under the 4.75-kg program target and represented a 60 percent mass savings vs. the steel decklid.

Duane Grider, supervisor, ZEV – Fleet & Vehicle Program Engineering at Ford, says, “Ford’s Focus FCV has been a very successful demonstration project. We see weight reduction as a continual challenge as we design more fuel-efficient vehicles. This is an opportunity for composite suppliers to provide cost-effective, weight-reduction solutions that meet all of Ford’s design criteria.”