This article explains rapid advances in science and computer technologies, especially simulation software that supports engineering analysis. The author's firm created an engineering simulation using physics-based Mechanical Event Simulation software from Algor Inc. of Pittsburgh to verify the integrity of a cinematic motion simulator ride. With the help of Algor’s technical support, the engineers determined that replicating the actuating movement of the hydraulic cylinders using Algor’s new actuator element technology and Mechanical Event Simulation (MES) software was the best method of evaluating real dynamic loads over time. The MES results showed that the stresses experienced by the bearings under loading from the six cylinders were well within the acceptable range. The actuating motion produced by the MES appeared to be very realistic when the analysis replays were viewed in real time.
The job of an engineering consultant specializing in amusement rides, theme park design, and water parks can often seem more like play than work. He works on countless theme park simulations-flight, free-fall, explosions, train rides and space travel, just to name a few.
While it is fun and exciting, engineering in the entertainment industry is also a very serious business. Millions of people take tens of millions of thrill rides each year in amusement parks all over the world. Assuring their safety as they roar through the loops of a roller coaster at 60 mph, or as they dive and dodge exploding asteroids in a flight simulator, is a demanding engineering responsibility.
Fortunately, rapid advances in science and computer technologies, especially simulation software, now support engineering analysis. However, for slide rule-trained engineers, realizing the benefits of the latest simulation tools can require some extra time and training. Those efforts paid off recently when the author's firm created an engineering simulation using physic s-base d Mechanical Event Simulation software from Algor Inc. of Pittsburgh to verify the integrity of a cinematic motion simulator ride.
A family entertainment complex in Edinburgh, Scotland, purchased a new TurboRide from Iwerks Entertainment Inc., headquartered in Burbank, Calif. Iwerks contracted Edward M. Pribonic P.E. Engineering and Consulting to verify that the American-built simulator would meet the requirements set by the British Fairground Standard for amusement ride safety.
The Iwerks TurboRide consists of two, four or eight seats mounted on motion bases that are arranged before giant flat or domed 180-degree theater screens. The motion bases move in synchronization with action-packed point of view movies. The large screens, coordinated movement, and booming digital audio transport riders through the virtual twists and turns of a space voyage, whisk them through the human body, or race them across the finish line of an Indy car-type race.
The Edinburgh installation consists of 12 four-seat motion bases. The fiberglass seats are supported by a welded steel frame of rectangular tubing, known as the flying platform. Six double-acting hydraulic cylinders connect the flying platform and floor-mounted base unit. Each cylinder is fastened at one end to the base unit via a cast iron bearing. The base unit has three mounting plates, each with two bearings to accept the lower ends of the cylinders. The computer-controlled hydraulic cylinders extend or retract independently, providing the seat and occupants with roll, pitch , yaw, heave, surge, and sway motion with six degrees of freedom in coordination with the on-screen adventure action.
After reviewing the simulator design, Iwerks, its client, and the author determined that all components of the ride should be structurally analyzed with the bearings as the nu in focus of the engineering analysis. Engineers needed to verify that the simulators would safely withstand the dynamic loading caused by the actuating cylinders, the weight of the entire assembly, and the presence of four adult passengers. Of special concern was a portion of the iron bearing housing beneath the bearing inserts. It was necessary to prove that there was no possibility of a catastrophic failure of this part.
While the goal of the project was straightforward, the method of achieving it was not. The first approach to the problem was to perform a linear static stress analysis. However, it was discovered that the inherent limitations of static analysis made it unsuitable for studying the dynamic nature of the simulator. The problem arose in determining loads created by the accelerations of the multiple double-acting cylinders to use as input into a linear static stress analysis.
Control valves connect to the top and bottom. of each cylinder. As the ride begins, computerized controls lift all six cylinders to their neutral positions at half the cylinder extension capability (the maximum extension is 25. 25 inches) by raising fluid levels in the lower portion and releasing fluid in the upper portion of the cylinder.
Computer commands conduct unique multidirectional extension and re traction sequences for each of the cylinders. An accumulator provides for fluid surges, while a central hydraulic power unit provides a constant fluid pressure of 2,000 pounds per square inch.
Due to the complexity of movement, it was not feasible to calculate the loading on the bearings from the six independently moving cylinders. A detailed solid model of the bearing was created in Solid Works and captured directly in Algor using In CAD Plus for Solid Works, without translation to a neutral file format. The model was fully constrained at the stainless steel inset in the center of the bearing. The engineering team decided to analyze a worst case scenario, applying an artificially high static load to the bearing.
In addition to the bearing analysis, Iwerks requested another analysis to verify that the welds of the motion base also would withstand the dynamic loads created by actuator movement and the presence of simulated passengers. A detailed solid model of the motion base was created in Solid Works and transferred to Algor in the same manner as the previous model. A load that would have been distributed to three mounting plates was applied to just one mounting plate in the vertical and horizontal directions.
The points where the seat is attached to the frame were fully fixed for the static analysis. The results showed only moderate stress levels in the structure despite an artificially high load case. The welds in the seat frame performed well under the applied loads.
The static analyses conducted separately on the bearings and seat frame yielded low to moderate stresses as well. While the results were favorable, both the consulting firm and the client were not convinced that the results portrayed an accurate picture of the mechanical system behavior because of the number of assumptions and simplifications made in the static analysis.
The author recommended building a complete, fully detailed, full-motion computer model of the simulator that could be used to run the motion profile at hand as well as new motion profiles as they are developed for future films. Stresses on the equipment can change with every new motion profile, so Iwerks needed an engineering model that could run each motion profile and produce results.
With the help of Algor's technical support, the engineers determined that replicating the actuating movement of the hydraulic cylinders using Algor's new actuator element technology and Mechanical Event Simulation software was the best method of evaluating real dynamic loads over time. In addition, a reliable computer model based on a de tailed CAD solid assembly would help all involved to better understand the dynamics of the design and apply this knowledge to future programming and simulator design decisions.
At the onset of the MES, the dynamic load calculation problems encountered in the linear static stress analyses were eliminated. MES simulates motion and flexing simultaneously to calculate stresses over time, so that forces are determined intrinsically by the software. MES is physics-based, not assumptions-based. Therefore, engineers could rely on the "known" physics of the event-the. weight of the simulator and passengers, gravity, pressure, and displacement of each cylinder over time-to unfold as the event was processed. Using this data, Pribonic Engineering and Consulting was challenged to simulate the six degrees of freedom motion capability of the simulator ride in order to determine dynamic stresses in the bearings.
For a simple MES, such as a manually shifted lever, the engineer simply specifies a prescribed displacement or, if it is known, the force needed to set up the MES. Then, the software will compute the acceleration and resulting stresses. In the case of the simulator, applying prescribed displacements for each cylinder to get dynamic motion and stress results over time was not feasible. At best, stresses could be determined at just one instant in time if this approach was used with the available software capabilities. Such an analysis would yield a similar result as a motion load transfer analysis, in which loading determined in a kinematic analysis is applied to a static stress analysis. The client had already agreed that this would not meet the safety analysis requirements for the project.
The complex actuating movement could be simulated using a new actuator element technology. Actuator elements are engineering elements (like contact or dashpot elements) that replicate linear extension and contraction movement in three-dimensional space, typical motion for hydraulic and pneumatic cylinders and for electric solenoids. An actuator element, which appears as a line, was used to represent each cylinder in the finite element model.
To proceed with the MES, the engineers created a detailed CAD assembly of the simulator that would be captured for finite element modeling. Engineers at Pribonic Engineering and Consulting modeled more than 100 individual components based on Iwerks' drawings and merged the components into five subassemblies. The surfaces of the subassemblies were meshed separately.
Once the subassemblies were meshed, they were merged into one model in Superdraw Ill. While some detailed surface matching was needed to align welded components in the seat frame, very little surface mesh enhancement was performed. The initial finite element model contained approximately 500,000 elements-far too many for a reasonable analysis processing time.
With the help of Algor's technical support, the consulting firm was able to reduce the overall number of finite elements in the model to about 141,000 elements. This was due in part to solid mesh engine, which automatically creates better aspect ratios for each solid element based on the quality level chosen by the engineer. Flexible brick elements were 'also replaced with kinematic elements where possible to reduce the analysis computation time. Kinematic elements behave just like flexible finite elements, but do not produce stresses. Engineers can insert kinematic elements in areas of the assembly where dynamic effects are essential but for which stresses are of secondary importance. This saves time, and the engineer can focus the analysis on the part of the mechanism being optimized-a set of bearings in this case.
After the model size was reduced, the consultant defined the analysis type, unit system, element, and material properties by group. A translator program was used to read in the displacement vs. time load curves for each actuator element. The program extracted the load curves for an 11.6-second segment of the motion profile provided by Iwerks. This particular segment was chosen because it contained the most extreme range of accelerations across all the load curves. Material properties for steel and cast iron were defined using Material Library Manager. Global analysis parameters included the duration of the event and capture rate. A rate of 30 captures per second was chosen to match the Iwerks motion file data points. Gravity also was applied to the model.
During processing, Algor's built-in visualization capabilities and Monitor utility were activated so engineers could watch the event unfold as it was processed. The processing took 2.5 days of actual computing time. The analysis was run on a Pentium II 400-Mhz machine, with some 500 MB of RAM and a 12-gigabyte SCS II hard drive . Algor enables "what you see is what you get" visualization by showing the movement of the mechanism and stresses as they occur over time. If an error had been found at the beginning or during the run , the analysis could have been stopped and the problem fixed without waiting for the entire run to complete.
The Monitor utility works like a virtual oscilloscope, displaying velocity, displacement, acceleration, reaction forces, or maximum stresses vs. time for a specified node. Using Monitor, acceleration vs. time data was viewed for the six cylinders.
As soon as the processing was completed, analysis replays of the MES were viewed in .avi format. These served as visual aids, helping the company explain to Iwerks the dynamics of the event in terms that nonengineers could readily understand.
The tensor stress output was reviewed normal to the base of the bearings and axially through the pivot point for the MES. Tensor stress output was chosen because of the brittle properties of cast iron. However, von Mises stress output was used in a general comparison with the linear static stress analysis to pinpoint timesteps with high relative stresses in the MES.
The MES results showed that the stresses experienced by the bearings under loading from the six cylinders were well within the acceptable range. A comparison of the maximum stresses found in the MES with those of the linear static stress analysis showed that the static results were extremely conservative. The accuracy of the MES stress results for the simulator was important to ensuring that the ride met the assigned British Fairground Standard.
The actuating motion produced by the MES appeared to be very realistic when the analysis replays were viewed in real time. However, to verity the accuracy of the actuator elements used to produce the motion, the MES output was compared to actual test data.
Using one of the motion bases produced for the Scottish TurboRide, engineers placed accelerometer test equipment at key areas of the simulator. One such point was placed in the left rear seat of the motion base. Data acquisition software compiled the acceleration data as the motion program was run on the simulator.
Acceleration vs. time data for a node in a similar location as the accelerometer testing was acquired from the Monitor program. The engineers found that the MES and physical acceleration testing results for the same time frame correlated very well, giving the consultant and Iwerks a great deal of confidence in the accuracy of the results.
Through this engineering challenge, Pribonic Engineering and Consulting recognized the value of developing virtual prototypes using MES. The firm can represent product designs using detailed CAD assemblies and simulate complex, dynamic behavior.
In this case, the virtual prototype verified that the ride would withstand stresses caused by the high-speed actions. Just as important, its true-to-life form will enable it to serve as a virtual prototype for future product developments by Iwerks.