Several new wing designs were evaluated on the 3-D model, simulating per- formance at the same speed as used on track tests. Visualizations of the first few design iterations showed regions of flow separation caused by the original wing mounts. The wing mounting system was redesigned to address the issue, and further CFD tests were run.
A comparison of oil flow on the rear wings clearly indicated flow separation on the mounts of the original wing and significant improvement in this area with the new wing design. The new design had a small separation at the trailing edge, but that was removed when a 6-mm gurney was attached to the wing. The final CAD model was loaded into CNC machines to produce tooling blocks needed to manufacture the new wing.
A new rear wing was made in time for track tests 6 weeks after the project had begun. The track tests and further running at the Le Mans prequalifying event confirmed that the new wing reduced drag by 2.5% for the same level of downforce.
Lap times on Le Mans test day confirmed what the race itself would later prove: The Veloqx Prodrive Ferraris were easily the quickest cars in the GTS class–with both the new and old wings. Because of time constraints, the new wing could not be tested for durability, a key factor in the 24-hour race. Not wanting to risk going into the race without durability testing and knowing that its cars were still the fastest, Veloqx Prodrive decided to stick with the old wing design. The wisdom of the decision was borne out by the victory of the young racing team.
Buoyed by the success of the Le Mans project and continually looking for a competitive edge, Prodrive developed an additional rear wing for the Ferrari 550 Maranello. The goal this time was to provide greater downforce for the American Le Mans series. A similar design approach was used for the new wing, and once again, it went straight from computational flow analysis to the track for testing. The wing exceeded predictions from CFD analysis when tested on the car and was used for the rest of the season, as Prodrive teams won the GTS class in all three of the American Le Mans series races in which they competed.
150 7 Reverse Engineering in the Automotive Industry
Faced with these rigorous quality demands, the company wanted to move be- yond traditional inspection processes using coordinate measurement machines (CMMs). CMMs collect a sample of discrete points on a part, one at a time. The process is slow and does not adequately address surface-to-surface inspection required to verify the accuracy of sheet metal or free-form surfaces. Results are recorded in a 2-D geometric dimensioning and tolerance (GD&T) report that does not directly correlate with the 3-D CAD model of the part.
The CAI process uses noncontact scanners to collect millions of points in sec- onds. Software based on reverse engineering principles then processes the in- formation automatically to compare an as-built part to its CAD representation.
The process creates an interactive loop among the design, manufacturing, and quality control divisions.
The first step in the CAI process is to capture accurate geometry and dimen- sions by placing target points, which are used to align multiple scans, on the surface of the existing engine component. The engine part is then scanned with a noncontact white-light scanner that generates a polygonal model.
Jagged edges, holes, and surfaces on the polygonal model are smoothed out, and the model is cleaned to remove extraneous points or noise that might un- dermine the data. The completed files are then merged, aligned, and saved in STL format. The STL model is imported into CAI software, which automatically allows engineers to align and compare the STL model with the original CAD data
Figure 7.10. A point cloud of an automotive intake turbine scanned from the physical part (left) and the original CAD reference model (right)
Figure 7.11. A polygonal model of an automotive exhaust manifold reverse engineered from the actual part
to determine exactly where variations in the geometry occur and to analyze how deviations might impact the part’s functionality.
Figure 7.12. Color maps visualize the deviation between the scanned sheet metal part and its corre- sponding reference CAD model. A color reproduction of this figure can be seen in the Color Section (pages 219–230).
Figure 7.13. An HTML report of a sheet metal part and a PDF report of an intake turbine.
A color reproduction of this figure can be seen in the Color Section (pages 219–230).
152 7 Reverse Engineering in the Automotive Industry
Even miniscule differences between the physical part and the CAD data can re- sult in performance flaws and inaccurate engineering analysis, so the auto maker imposes a tolerance level of 0.02 to 0.03 millimeters–less than the width of a hu- man hair. If parts differ from the original design by more than that, they are sent back to be reworked. Part discrepancies are communicated through reports with visual images and numerical results that are generated automatically by the CAI software. Reports can be output in many standard formats, including HTML, PDF, Microsoft Word, and Excel, or as customizable graphics. The reports are also leveraged by suppliers, quality engineers, and others in the company’s sup- ply chain.
The European automaker is joining its Japanese counterparts in finding that the CAI process based on reverse engineering saves time and increases quality by ensuring dimensional accuracy and engineering performance. According to the manufacturer, design engineers have better information about the quality of parts early in the product life cycle, which saves the company money and pro- duces a better end product for consumers.