This article focuses on the finite element analysis (FEA) that is a key ingredient in keeping a popular theme park ride up and running. Long before the ride entered service, Universal called in GLENCO Engineering Inc., to perform finite element modeling and analysis in order to evaluate the primary structural components of the Spider-Man ride vehicle cabin. The FEA initiated to increase the stiffness of the cabin floor also identified a location where the peak strains were higher than the values determined in the hand calculations and associated strain gauge testing. The FEA of the Spider-Man cabin structure significantly improved the design of the composite floor. Strains measured at locations selected without the benefit of FEA produced a misleading assessment of the design in its first prototype. FEA, however, identified peak strains at a location and direction that were not intuitively obvious to the engineers. The results were confirmed with dynamic strain measurements—verifying not only magnitude but also directions of the principal strains. The finite element model was a valuable tool that enabled the floor design to be precisely refined in one of the iterations.
The amazing adventures of Spider-Man ride at Universal Studios’ Islands of Adventure theme park in Orlando, Fla., averages several thousand riders a day. And it’s a safe bet that none of them is aware of the engineering technology beneath their feet.
Each ride vehicle consists of an open passenger cabin mounted on a translating motion base. Achieving the desired ride performance from the motion base meant minimizing the cabin weight. This led Universal to select carbon-fiber composite technology for one of the most critical structural components of the vehicle assembly, the cabin floor.
A ride’s safety is paramount, and that is ensured by built-in durability. In the case of the cabin floors, reliability is especially critical as composites generally cannot be repaired. And replacing the floors would be a huge task, since the entire cabin is built from the floor up.
Spider-Man represents a big investment, and is heavily promoted. The ride is expected to operate 16 hours a day, every day of the year, and Universal’s management is loath to disappoint vacationing customers and their kids by not having the ride functioning when the park is open.
Universal points out that the entire ride system, including every vehicle, is thoroughly inspected nightly. In addition, routine maintenance and rebuilds are scheduled at preset intervals for each vehicle. An unexpected maintenance task, such as replacing the floors, could cut into the availability of the ride.
Long before the ride entered service, Universal called in GLENCO Engineering Inc., to perform finite element modeling and analysis in order to evaluate the primary structural components of the Spider-Man ride vehicle cabin.
GLENCO specializes in structural analysis and strain gauge testing for the theme park industry by using methods that were developed for aerospace engineering. The company’s aerospace experience includes work at Allied- Signal, TRW Inc., and Hughes Aircraft Co.
The software used to analyze the Spider-Man ride was ANSYS Mechanical from ANSYS Inc. of Canonsburg, Pa.
Universal had a prototype cabin built before computer analysis was performed in order to get the creative and other design activities running. The prototype floor was designed using hand calculations based upon the premise that it would function like a cantilever beam.
Subsequent testing of the prototype with strain gauges located according to the cantilever assumption indicated that the stresses in the floor were low. However, Universal’s engineers felt that guests might be able to perceive the flexing and vibration of the floor. Thus, the initial testing indicated that the bending stiffness and associated natural frequency of the floor should be increased.
The finite element analysis initiated to increase the stiffness of the cabin floor also identified a location where the peak strains were higher than the values determined in the hand calculations and associated strain gauge testing. The cabin floor is attached to the motion base through a steel ring embedded within the composite layup, at about the center of the floor. FEA found that the peak strains in the composite did not occur adjacent to the forwardmost portion of the ring (the 12 o’clock position) as expected, but at the 3 o’clock and 9 o’clock positions around the ring.
The computer results were confirmed by dynamic strain measurements obtained during simulated operation of the prototype cabin assembly. The use of finite element analysis permitted the stiffness and strength of the cabin floor to be substantially improved in just one design iteration. The project stayed on schedule, opened to rave reviews, and has been a top draw ever since.
The Spider-Man ride tells the story of the triumph over Doctor Octopus, who threatens to destroy New York City’s Manhattan Island and all its inhabitants. It’s Spider-Man and each vehicle’s guests allied against a cast of ominous supervillains.
Spider-Man showcases the talents of Hollywood’s king of special effects, Steven Spielberg. Thanks to high-speed moving simulators, oversize 3-D graphics, custom-choreographed mechanical animation, and pyrotechnic special effects, the battle rages over, under, and around the guests. They dodge bombardments from a host of evildoers, including spewing water pipes and hobgoblins’ exploding pumpkin bombs.
Loads of Illusion
Spider-Man combines a translating motion base with 3-D video. With 25 large- format movie projectors and numerous smaller projectors, the action seems always to be in the guests’ faces. Other, more surreal experiences include being lifted 400 feet in the air, then plummeting back to street level. Although the motion base drops only about a foot during the fall, the sensation induced is amazingly realistic.
High-performance rides such as Spider- Man generate substantial accelerations and associated loads on the vehicle—at times comparable to four-wheeling over rough terrain. In addition to pitch, roll, and heave movements, each vehicle’s motion base is capable of rotating the cabin 360 degrees about the yaw axis in order to direct the guests’ attention continuously toward the action. The floor supports the loads generated by the entire cabin structure and up to a dozen adults, so that its integrity is critical to the ride’s operation.
The finite element model of the floor included directional, orthotropic properties matching the various regions of the composite layup. The elastic properties of the composite skins and internal ribs are specified based on the orientation of the carbon filament and glass fiber- reinforced materials in each section. This is notably different from the analysis of an isotropic metallic structure, where the properties are uniform in all directions. The floor skins and ribs have 15 discrete regions with differing composite properties. Assigning properties correctly was an essential step in obtaining analytical results that correlated with the strain gauge measurements.
The model of the entire cabin assembly comprises about 20,000 elements. Most of these are shells, while the rest are beams, spars, springs, etc. The solution included approximately 110,000 degrees of freedom. The model’s database file is approximately 20 MB and the results file for the full model solution (three load steps) is approximately 110 MB.
The model was created directly in the ANSYS Prep7 preprocessor from data provided by Universal—primarily 2-D engineering drawings and tabular summaries of the composite layup schedules. The analyses of the floor structure were mostly linear. Some nonlinear, large-deflection solutions were also performed for the floor and cabin door. All postprocessing was done in ANSYS Post 1.
Most of the work was done on a PC powered by a 300- MHz Intel Corp. Pentium II CPU. The machine has 192 MB of RAM and a 6-GB hard disk. Solving the model of the entire cabin with three load steps took 2Уг hours. A floor-only subset of the cabin model was solved in half an hour, again with three load steps.
For comparison, the FEA was rerun after GLENCO upgraded to a machine with a 550-MHz Pentium III, with 768 MB of RAM and a 54-GB hard disk. Repeated tests on the two machines produced identical solution results, despite the huge differences in CPU clock speed and available RAM. The solution time dropped to 18 minutes for three load steps with the ANSYS Iterative Solver.
The floor’s response to sudden pitch movements, the most severe loading condition, was the primary concern of Universal’s engineers. The resulting bending deformation causes the cabin’s nose to deflect up and down in a manner similar to the response of a cantilevered diving board.
The natural frequencies of the baseline finite element model corresponded closely to the values obtained empirically from earlier tests of the prototype. The validation was augmented by comparing results for both the bare floor and the completely assembled cabin. The added inertia of the body components reduced the frequency of the pitch mode threefold.
Further analysis with the model revealed that the flexural stiffness about the pitch axis could be nearly doubled with the addition of a central rib and by increasing the thickness of the floor’s internal longitudinal stiffener ribs. This change sufficiently increased the frequency of the floor’s pitch mode with only a negligible impact on the cabin weight.
The FEA also predicted that the circular constraint created by the steel ring produced peak strains that exceeded the levels desired by Universal and at unexpected locations. These strains, caused by flexing in response to sudden pitch events, occurred tangent to the steel mounting ring at locations closer to the cabin’s sides than at the forwardmost portion of the ring as expected based upon the cantilevered beam analogy.
Further examination of the results showed that the peak strains in the composite skin were predominantly shear. Consequently, the principal strains acted approximately at 45-degree angles to the centerline of the cabin floor. This finding revealed that more plies were needed in those directions in the production floor layup schedule.
Halving Peak Strains
Following confirmation, the model was used to evaluate modifications to reduce the peak strains in the floor skins. The orientation of fibers in the prototype floor was predominantly front to rear and side to side, an arrangement that was tailored to the cantilever beam assumption, but which did not address the shear strains produced by the circular constraint of the steel ring.
Two plies of carbon fiber material were added to the cabin floor around the embedded steel mounting ring. Oriented at 45-degree angles to the cabin centerline so as to directly counteract the maximum principal strains, the additional plies halved the peak strains. Establishing that only two additional plies were needed, and in just a small region, was of significant value to Universal because weight was critical in cabin design.
These analytical predictions were also confirmed with more strain gauge testing. This second set of tests used a complete ride vehicle, including the redesigned composite floor, operating on the actual ride track.
The finite element analysis of the Spider-Man cabin structure significantly improved the design of the composite floor. Strains measured at locations selected without the benefit of FEA produced a misleading assessment of the design in its first prototype. Finite element analysis, however, identified peak strains at a location and direction that were not intuitively obvious to the engineers.
The results were confirmed with dynamic strain measurements—verifying not only magnitude but also directions of the principal strains. The finite element model was a valuable tool that enabled the floor design to be precisely refined in just one iteration.