Carbon Fibre Internal Structures

The systems problem

The design of the internal structures is fundamentally a systems integration problem, as they react and interface the loads/geometry of almost all the subsystems on the car.

The bulkheads are the direct mounting point for six high-load subsystems:

Front suspension mounts its control arms and shocks, and produces the most significant reaction forces within the bulkheads

Steering mounts its racks and columns directly to a few very precisely defined points to enable the asymmetric steering functionality, while a belt tensioning system also relies on bulkhead reaction forces to tension a timing belt

Latches holding the open-able topshell (solar deck) down are mounted directly to the bulkheads

The opening mechanism of the topshell mounts its hinges and struts off the bulkheads, on which significant loads are reacted from aerodynamic gust cases

The roll hoop also mounts one of its tube terminations onto a bulkhead

The battery is mounted onto the bulkheads and is by regulation required to be restrained in a 20g acceleration in any direction

The bulkheads also provide the geometric interface for many more mechanical and electrical systems:

Low and high voltage electrical cables pass through and around many bulkheads

Brake lines pass through bulkheads low-down (to avoid air entrapment); they also produce a force reacted to by the bulkheads via the suspensions

Rear suspension loads are reacted through the bulkheads and need to be considered as a support condition

Solar MPPTs mount at the same vertical height as the bulkheads, while providing a serviceability requirement that requires specific bulkhead beams to be able to take the load of a human sitting/lying on top

As a systems engineering problem, I needed to liaise across subteams, across many people and subsystems that had arrived at their own conclusions on ideal geometry, constraints and loads. I needed to reconcile systems and resolve conflicts all while attempting to achieve my own objectives in reducing weight and cutting down the number of bulkheads (everyone has an ideal bulkhead position, but we don’t want there to be 50 of them in the car! There are trade-offs that needed to be justified and made). After geometric and functional requirements, I re-calculated and re-ran all load cases to cross check and produce a consistent methodology for analysis, before moving into detailed composites design.

Geometric design

The capped “I beam” design is unusual in the world of solar cars, but was a deliberate choice made as a trade-off between serviceability and optimisation. Solar Car 4 was our team’s first asymmetric catamaran, and also the first time we were designing every structural and mechanical system in-house. This drove a clear priority: modular design that enables fast geometric change, as well as easy serviceability in its usage condition. As such, a beam structure was the natural architecture of the system, compared to semi-monocoque “torsion box” systems that are often seen in other teams (better weight optimisation, worse modularity, worse serviceability).

In future, a “torsion box” may be the better option to optimise load paths and trim weight, but given the inexperienced team I was working with, with constant revisions to subsystem requirements, and a team that was learning composites for the first time, I believe I made the right choice to use beams. It let us keep iterating up until very late into the timeline, without large re-tooling constraints that would otherwise hinder our development.

Analysis under uncertainty

We had no measured suspension load data, and outback highway conditions are unpredictable and often catch out even the best of teams anyway – delaminating mounting structures or snapping suspension arms. Furthermore, as an inexperienced team with constantly changing parameters, I decided not to run an inertial-relief simulation typical of this use-case, as I cannot validate or even be confident of the unknowns of final mass and placement. With the many uncertainties in design, and the practical inability to validate them until the car is driving (at which point it is too late to change anything), I chose to design against conservative representative worst case static cases. Fixed displacement constraints on the occupant cell capture loads going through the transverse bulkheads, while triangulated displacement constraints across the contact patches capture loads through the diagonal bulkheads. The whole structure was designed to the materials factors of the EuroComp composites design code, which by recommendation in consultation with professional engineers, sets a standard for a safe and robust first in-house composites structure.

Composites design followed a quick to detailed workflow:

Detailed FEA was conducted using Ansys Mechanical with ACP composites layup modelling. Suspension loads were modelled through a representative beam element model. A shell model was used to model the bulkheads, including the unidirectional capping to form the I beam, shear patches, and foam and solid core replacements. A range of load cases were analysed, including various suspension and internal mounting loads.

Hand calculations were used to work around the limitations of the shell element mesh, validating the strength of bonding and overlaminate joints between bulkheads, validating shear patch and core insert capabilities, and justifying points of artificially high stress.

Results

The completed bulkhead assembly weighs around 18kg before bonding into the aeroshell, which is within comparable range to equivalent structures on the best international teams. Despite the 50% car size increase, it still helped to trim weight overall, and it ran reliably without needing repairs across ~2000km of outback roads (1301km in race conditions), over unpredictable cattle grids, the occasional off-road excursion, and many times around the Hidden Valley Racetrack in Darwin.