...
Bends were created in the panels by removing the carbon laminate layer on the inner side of the bend. Some of the Nomex core was removed to allow the panel freedom to bend. The open cut was then filled with a lightweight epoxy, bent to the appropriate geometry and held to cure. Several layers of plainweave carbon fiber cloth were added to inner side of the bend to add strength.
Goal: Discuss sesign and analysis methodology used by UofT
The univeristy of Toronto decided to use composites for the following reasons:
- Using composites in addition to steel tubing allows for a lighter chassis in comparison to a full steel frame chassis.
- Increased torsional stiffness
During their design process, they believed it was important to note that their chassis was designed for optimal stiffness, not maximal. It's important to note, seeing as the closer you get to a "perfectly stiff chassis" results in more dead weight (i.e. the weight to stiffness ratio begins decrease exponentially).
Composite Design and Optimization
For their composite design they focused on using 2 laminate layers with one core layer. Examples of their panel design can be seen above.
During their optimization process they focused optimizing the following three parameters:
- Laminate thickness
- Core thickness
- Ply directions
By optimizing the above mentioned parameters, their goal was to improve the stiffness to weight ratio. From their optimization, they found that ply direction affected their design and results the most out of the three as it greatly impacted force loading results and projected performance of the car. UofT used Altair OptiStruct in order to determine the optimal direction of plies.
UofT's process for simulating a side impact:
- simulated with dynamic loading
- to show how the chassis was being loaded, they decided to look at the amount of energy absorbed at different points in the chassis, as well as how the car maintained and distributed the load.
- to estimate minimum laminate and core thickness allowable, they used generalized panel calculations, in order to find the safest structure
For examples of equivalency calculations used please, see appendix B in the above linked document.
Chassis Design and Optimization
To begin with, UofT defined what their optimal torsional stiffness range was. They decided on this to ensure their design was optimized and not maximized. First they determined what the difference between the expected front and rear stiffness values and then multiplied them by 6-8. For example, their front stiffness value was 663 ft-lb/deg and their rear stiffness value was 343 ft-lb/deg; as a result their optimal stiffness range was 1950 ft-lb/deg to 2800 ft-lb/deg. In their simulation software, they represented all of the tubes as beam elements, and applied a linear analysis. While performing their baseline analysis, they represented their composite panels as shell elements with isotropic properties (which is not the case in reality) they did this in order to simplify the analysis greatly, and focus mainly on optimizing the design of the steel frame.
The main advantages of this are:
- faster simulation times and simplified analysis set up
- faster design iterations
- focus on improving relative stiffness of steel geometries.
After they had roughly optimized the design of the steel geometries, they began improving the representations of the composite panels. They switch the composites to be represented as "advanced shells" specifying directional properties (representing the panels as orthotropic materials) as well as the ply direction and thickness.
Model constrains
Constraints in the model included limiting the movement of the left, rear, lower suspension point from translating in the x, y, and z directions. The right, rear, lower suspension point was restricted in the x (longitudinal), z (vertical), but allowed to translate in the y (lateral direction). Allowing this lateral movement serves to avoid over-constraining the chassis, which would produce stiffness values much higher than is realistic.
A torque of 2000 ft-lb was applied at a center point of the forward axle of the chassis. Because all analyses carried out in these studies were linear, the actual value of the torque applied is arbitrary, though the relative angular deflection values of each iteration of the chassis under this load are of interest in optimizing the design.
Composite Strucuture Design and Optimization
In this section, the article was very lacking, no new information was discussed and only mentions, they just specified number of initial layers, material properties of each layer and its orientation, and then boundaries of how much each layer can rotate. Afterwards, their software just gave them an answer for what the optimal design would be.
Lastly, it is important to note this paper focuses solely on optimizing stiffness, and does not focus on or necessarily discuss how to design and simulate and ensure their chassis is safe for impact and rollover cases.
For further information about their materials and manufacturing process, please refer to chapter 3.
For further information about how they validated their design and performed physical testing please refer to chapter 4.
Stanford's Solar Car Sundae
...