Machinedesign 2631 090122 P26 Waterslide 0
Machinedesign 2631 090122 P26 Waterslide 0
Machinedesign 2631 090122 P26 Waterslide 0
Machinedesign 2631 090122 P26 Waterslide 0
Machinedesign 2631 090122 P26 Waterslide 0

Waterslide Simulator

Jan. 20, 2009
For aquatic daredevils, heaven would be an “endless” waterslide that fits in your back yard

Delta Computer Systems
deltacompsys.com

For aquatic daredevils, heaven would be an “endless” waterslide that fits in your back yard. That exact idea came to Zoz Brooks, host of Discovery Channel’s “Prototype This,” a series which shows how engineers and scientists design and build prototypes that solve the world’s problems — or just do cool things.

In a recent episode, the design team built a 30-ft-diameter doughnut out of waterslide sections. Hydraulic actuators lift, tilt, and spin the circular apparatus while a computer projects an image of a waterslide course inside. Riders surf inside on a cascade of water, with the actuators simulating the twists, turns, and bumps of a slide that, for the show, seems 450-ft high and 3,600 ft long and takes about three minutes from start to finish.

Because of the sheer size of the “slide,” designers chose hydraulics to power it. The doughnut weighs approximately 6,000 lb. And it’s easier to move heavy loads with smooth changes in direction using hydraulics rather than with electric drives. A hydraulic motor also turns the friction wheel that spins the doughnut.

Two pairs of linear hydraulic actuators lift and tilt the doughnut and keep the structure stable. Each actuator must synchronize with its mate so the superstructure doesn’t twist out of alignment. And like other motion-based simulators, the movements of all the axes must be coordinated with video images the rider sees.

The design team opted for an RMC150 motion controller from Delta Computer Systems, Vancouver, Wash., to manage and synchronize the actuators. The controller ensures smooth movements by executing special instructions called spline functions. Splines create smooth curves between target points so there are no discontinuities in the motion, which riders would feel as unacceptable bumps out of synch with the video. Brooks generated about 50,000 position data points to match doughnut motion with the ride video, which the design team subsequently boiled down to approximately 1,000 target points — one every 180 msec. Each point includes values for lift, tilt, and rotational velocity.

If that’s all the controller had to do, the design would have been straightforward and much like other motion platforms. However, the hydraulics presented additional challenges. One key issue is often referred to as a spring effect. The simulator had 50 ft of hydraulic hose between the valves and hydraulic actuators, and hydraulic pressure fluctuations made the hose expand and contract which, in turn, causes the cylinders to lag behind the target motion.

This made the slide impossible to control using simple PID algorithms. Designers could have compensated with a low P gain to make the ride more sluggish, but this would have taken out some of the thrills and chills. In retrospect, hydraulic valves could have been mounted directly on the cylinders, but the prototype was built quickly and this detail was sacrificed for the sake of expediency.

Engineers from Delta Computer Systems were called in to help program the controls and tune the axis, as well as address the spring-effect issues. Company spokesperson Dennis Ritola says adding a second-derivative gain in the control algorithm fixed the problem. He explains that the Delta RMC also supports acceleration control and that the new control algorithm is a PID + D2 rather than just PID. This reduces errors between target and actual acceleration of an axis at any point in the motion profile. It also generates early speedup or slowdown when a motion axis starts to lag.

To ensure the ride responds quickly to changing motion targets, Delta engineers also added feedforward gain to the control algorithm. Feedforward is a predictive element in the control-loop equation that anticipates what the control output needs to be. (The other gain parameters in the equation work to reduce errors, not predict the motion.) Parameters were graphically tuned with the company’s RMCTools software. When complete, the swirling doughnut is both responsive and “as smooth as silk,” says Ritola.

So far, the virtual-waterslide prototype is one of a kind, but it shows hydraulic-based motion simulators can smoothly and quickly move large loads and give realistic rides.

Sponsored Recommendations

Altech's Liquid Tight Strain Relifs Catalog

March 13, 2024
With experienced Product Engineers and Customer Service personnel, Altech provides solutions to your most pressing application challenges. All with one thought in mind - to ensure...

Industrial Straight-Through Cable Gland

March 13, 2024
Learn more about Altech's cable glands and all they have to offer for your needs!

Smooth Sorting with SEW-EURODRIVE!

Feb. 22, 2024
Sorting systems are essential when it comes to warehouse automation, material handling, and distribution. SEW-EURODRIVE’s automated sorting solutions increase capacity, reliability...

Setting the Agricultural Industry in Motion

Feb. 22, 2024
Agricultural products go through specialized processes of harvesting, storing, conveying, cleaning, drying, sorting, and handling to ensure that seasonal products reach the processing...

Voice your opinion!

To join the conversation, and become an exclusive member of Machine Design, create an account today!