TEAM P18318: Pressure Coasters. Restrain Your Enthusiasm Robert Cybulski, Robb Foote, John Stelmack, Mike Troise, Alison Wright

Transcription

1 TEAM P18318: Pressure Coasters. Restrain Your Enthusiasm Robert Cybulski, Robb Foote, John Stelmack, Mike Troise, Alison Wright Abstract: A roller coaster parts manufacturer has designed a hydraulic locking cylinder used for passenger restraints that provides infinite adjustability along with keeping the passenger safe while riding. The hydraulic cylinder results in a system that requires great force to actuate, a large problem for younger riders. A new hydraulic locking cylinder has been designed that automatically extends and retracts via a ball screw system connected to the piston and rod of the hydraulic system and is driven by an electric motor. The resulting cylinder improves ride loading cycle time and eliminates the need to operators and riders to physically lower the ride restraint. The paper will outline the design of the system and includes results of preliminary testing. Background and Motivation: Passenger restraint systems are essential to the success of an amusement park ride. These systems must be easy to use and safe. Traditionally, passenger restraint systems have used ratchet and pawl systems. While these systems have met most of the amusement park desires, the prevalent use of moving parts in the systems have posed reliability challenges to park operators. These systems are almost always too tight or too loose for most riders. In recent years, hydraulic restraint systems have become more widely employed because of their simplistic design and infinite adjustability. The benefits of these systems are preferred by ride manufacturers but unfortunately, there are significant drawbacks to current system designs. Since the systems are not automatic, users must position the restraints themselves which requires much more force than ratchet and pawl systems. This has made positioning restraints especially difficult for younger riders and has extended the time needed to load a ride, extending ride throughput compared to ratchet and pawl systems. Advanced Concepts in Manufacturing (ACM) has developed a hydraulic system/cylinder specifically designed for the amusement industry. The system is plagued with the same usability concerns as other similar hydraulic systems. As the project s primary customer, ACM would like to build an automatic raise and lower hydraulic restraint system to eliminate the common complaints associated with hydraulic systems. ACM is the project s primary stakeholder since the demonstration of a working prototype is desirable for the development of future hydraulic restraint systems. Similarly, ride manufacturers are significant stakeholders in the project as they would like to install the most user-friendly and desirable ride possible. Additionally, park owners, ride operators, and riders are stakeholders since easier to use restraints will results in fewer complaints from riders benefitting all parties. Faster ride throughput will also benefit all as faster lines are more desirable for riders and when riders are happy, park owners and ride operators are satisfied as well. ACM had many requirements for a new system. The most important of these is that the new restraint needed to automatically extend and retract without physical input from the rider. Of similar importance was the safety while the system was in use. The new design could compromise rider safety and needed to meet all applicable ASTM requirements, specifically, ASTM F2291 Standard Practice for Design of Amusement Rides and Devices. Another important requirement was that the new system needed to be similar in weight current systems. Weight is very important on rollercoasters and every

2 pound added to a design counts. Similarly, the size of the new system needed to be comparable to current systems since space is very limited on ride vehicle designs. Finally, movement time requirements were specified to ensure the system would be efficient and maintain or improve ride capacity. These qualitative requirements are quantified by the engineering requirements summarized in Table 1. Table 1. Engineering requirements based on customer and ASTM specifications. Mechanical Design and Feasibility: To fulfill the requirements set forth by ACM and ASTM, two promising design options for cylinder movement were considered: a ball screw driven hydraulic cylinder and a hydraulic pump driven cylinder. Research and a feasibility studies were performed for each design choice before selecting a design. Due to the added weight of a hydraulic pump and motor, added current draw of a hydraulic pump, and the possibility that using a pump could introduce air bubbles into the system, the ball screw driven system proved to be the best design choice. There were concerns of threads breaking and contaminating the hydraulic fluid in the screw drive system; however, analysis showed that this problem could be addressed by using filters throughout the system. The ball screw driven system is essentially a combination of current passive hydraulic cylinders and current electromechanical linear actuators. The rod and piston in a passive hydraulic system are moved to the appropriate location in the cylinder when the patron or operator positions the restraint. As the piston is moved, fluid flows between the two sides of the cylinder. Once the restraint is in the proper position, the locking valve is locked by removing power to the system. Closing this valve prevents fluid from moving between the two sides of the piston and holds the restraint in place throughout the ride. The system was designed as a locking retract cylinder as requested by the customer. This means that when the cylinder is locked, the locking valve will not allow the rod to retract into the cylinder. A locking retract configuration allows for easy implementation on a standard lap bar system as shown in Figure 1.

3 Figure 1: Side view of lapbar in up (left) and down (right) position with movement of hydraulic cylinder demonstrated. The ball screw driven cylinder s hydraulic components and operation are similar to current passive hydraulic systems; however, the force needed to move the piston within the cylinder is provided by the movement of a ball screw system. To achieve this motion, the piston and rod of the hydraulic system are connected to the nut of the ball screw system. As the nut moves along the screw, the piston moves within the cylinder and the rod extends and retracts in and out of the cylinder. The piston s motion forces the hydraulic fluid to move to the proper locking locations within the locking mechanism. Once the restraint is properly positioned, the hydraulic system is locked and power is removed from all components. As with most linear actuators, the radial motion of the screw is provided by a motor. The motor is connected to the screw by a pulley and belt system. The motor was positioned to the side of the system due to the need for a ride chassis connection point directly behind the screw along the axis of the cylinder. Prior to beginning detailed design of the machined hydraulic components, important design decisions had to be made regarding the ball screw, motor, and belt drive system. Using ER 14, the assumed velocity curve shown in Figure 2, and the stroke of the cylinder, the maximum linear acceleration of the rod was found. Dimensions and masses for the restraint system were then assumed to calculate the required linear force provided by the cylinder. Ball screw calculations were then used to translate the linear force of the rod to a torque provided by the screw. This torque on the screw was used to specify and source a motor and belt drive system. Full detailed calculations of this process can be found on P18318 s EDGE website.

4 Figure 2: Initial velocity curve used for motor and ball screw calculations (more detailed profile developed during testing). Once design of the drive system was completed, focus shifted to designing the machined components needed to support the drive system and properly direct fluid flow. Many of the hydraulic components were based on passive hydraulic system designs already completed by ACM. Slight modifications were needed to incorporate the ball screw and many components needed to be scaled in size; however, these components were considered service proven. The hydraulic locking components from the current ACM passive cylinder including the manifold, locking valve, manual release valve, and accumulator were able to be reused. To accommodate this reuse, a new intermediate manifold was designed to transfer fluid from the newly designed hydraulic cylinder to the original ACM locking components. Multiple views of the assembly CAD model are displayed in Figure 3 and 4. Figure 3. Isometric view of full assembly CAD model.

5 Figure 4. Section view (left) and rear view (right) of assembly. A major design challenge was to restrict the radial motion of the nut. As explained, linear actuators feature grooves in which the nut rides, preventing radial motion and forcing the nut to move axially along the screw. Since sealing a non-circular surface is difficult, the typical grooves could not be used in this design. To mitigate this challenge, unique positioning rods were implemented. These rods run the length of the cylinder and pass through the piston. The rods are fixed at both ends in the insert and gland components. The system is pictured in Figure 5. Figure 5: Positioning rods highlighted as yellow components. Although many parts were considered service proven, analysis needed to be performed on new parts or parts that were significantly modified. Two components that required this additional analysis were the rod and the rear cover. To effectively analyze these components, loads had to be determined that would be applied during the analyses. Since the rod and rear cover are loaded along the same axis, the same forces could be applied to both. The loads were found by using the assumed restraint geometry and Section 7 Acceleration Limits of ASTM F2291 Standard Practice for Design of Amusement Rides and Devices. In section 7, the allowable duration of sustained forces on a rider are specified for the three principal

6 directions. Since, the highest forces are seen in the vertical direction and this is the main direction affecting the cylinder, only accelerations in the vertical z-direction were considered. Figure 6 in Section 7 pictured in Figure 5 specifies the allowable duration of forces in the positive and negative vertical direction. Using this figure as well as specifications regarding the transition time between positive and negative forces, a g-force plot shown in Figure 6 was generated for an event in which the rider experiences negative g-forces (traveling over a hill) before quickly switching to positive g-forces (in a valley). Using the geometry of the restraint and assuming the weight of a rider is 170 lb as specified in Section 8 Loads & Strengths, a plot of the force acting on the cylinder vs. time was found. The plot of the g-forces on the rider and forces on the cylinder is shown in figure 1 below. These force values were then used throughout the calculations for the rod and rear cover. Figures 6 and 7: ASTM plot of vertical g-forces vs. time (left) and P18318 g-force/load case used for analysis (right) Unlike the typical rod in a hydraulic cylinder, the rod used in this product featured a large hole drilled along its axis to allow for the ball screw to pass through it. Since the load on the rod should be transmitted axially along the rod, the failure case of greatest concern was buckling. A buckling calculation was performed on the hollow rod assuming that one end was fixed while the other was free. Assuming a worst-case scenario, the force applied to the rod was taken to be the maximum force derived from the

7 ASTM standards as explained above. The results of the analysis confirmed that the factor of safety was sufficiently high according to the customer s specifications. The rear cover was analyzed by performing finite element analysis. Three load conditions were tested using the finite element model: 1. Maximum compressive force due to a negative G moment (passenger weight of 170 lb) 2. Maximum tensile force due to a positive G moment (passenger weight of 170 lb) 3. Maximum compressive force due to a negative G moment (passenger weight of 300 lb). The forces of each analysis were applied as a bearing load on the bearing face of the rear cover in the appropriate direction (shown in Figure 8). The cover was supported as appropriate based on the direction of the applied force and the adjacent components. To establish the legitimacy of the results, mesh refinement was performed until the solution converged (shown in Figure 3). Different types of load applications were also applied to ensure that the type of loading did not significantly affect the results. The Von-Mises stress and total deformation of each load case was analyzed and showed that the factor of safety during the compressive loads is less than desired. This is acceptable for the current prototype as it will not be exposed to these types of loads; however, modifications and further analysis should be performed to ensure that the component will not fail during actual use on a ride. An example of one of the analyses is displayed in Figure 9. Figure 8. Converged mesh (left) and loading (right) of FEA of rear cover component.

8 Figure 9. Sample of FEA results of rear cover when loaded under the worst-case scenario with 170 lb. patron. Von-Mises stress is plotted in the left image while total deformation is displayed in the right image. Electrical and Software Design The electrical and software design goal of this project is to seamlessly apply the entire interface to any existing PLC unit. Along with any existing PLC unit, the available power for this project is limited to 24VDC. Given the assistance of any current system, the scope of the electrical is to isolate controls from the PLC unit to the main controller on board and efficiently power the selected controller, motor, current sensor, force sensor and IR sensor. Once each item is efficiently powered, the software will take over the controls from the PLC unit as Forward, Reverse, or Stop commands are received to drive the motor. The controller will send a PWM signal to maintain efficient motor control. Continuously monitoring control devices dictate when the motor control stops. Electrical To make a seamless environment with any PLC unit, the 24VDC relay signals coming into any controller must be isolated to prevent any possible spike in relay controls that could possibly harm the controller. Therefore a 4N35 optocoupler (Figure 10), is used in the control circuitry to isolate the incoming 24VDC and lower that signal to properly be received by any controller.

9 Figure 10. Optocoupler Circuit, PLC input signals come from the left and output low voltage signals on the right side of the circuit To power the controller for this system, a specified voltage range is required. Creating a supply voltage values from the main power supply, a voltage regulator or buck converter is required. Voltage regulators are effective by the means of producing very little output ripple. Unfortunately the design of voltage regulators draw more current than a buck converter. That excess current is then released as heat, which makes them inefficient for this system. This is the reason a buck converter was selected to power the external devices as well as the controller. Figure 11 below represents the basic efficiencies of buck converters vs. voltage regulators. The blue slope is the efficiency of a buck converter, which is why the buck converter was chosen. Figure 11. Efficiency Slopes of a Buck Converter (Switching Converter, Blue) and a Voltage Regulator (Linear Regulator, Red) Figure 12. Buck Converter Circuit

10 Figure 13. Voltage Regulator Circuit Figure 12 and 13 show the circuitry of a buck converter and voltage regulator respectively that are each designed to power the controller. The controller that will be powered by this type of converter is the Arduino Uno R3. This device requires a supply of 9 to 12 Volts. To power this device a D24V22F5 buck converter will be used to take an input of 24VDC and step it down to 9VDC. The Crouzet series 8989B motor was chosen, which can output a torque of 2Nm and a rated current of 4.9A. An H-bridge is necessary to drive this motor to not only go forwards but reverse as well. Figure 5 below shows the function of an H-Bridge. The H-Bridge that will control the motor is the DROK L298 (Figure 14). Figure 14. H-Bridge Typical Circuit Figure 15. H-Bridge that will control the Motor As per the original design and theory of this type of automated system, a force sensor will be used to simulate making contact with a patron. This will integrate with the controller and motor controls. If it detects a force above 4 lbs it will be triggered and brake the system. For testing purposes an IR sensor will be introduced to brake the motor on the test stand at different extensions. It will also be used to do perform distance sensing. This will send signals to the controller to either stop the motor or continue. For example, if the motor has driven the system at x distance away from its mounted position a signal will be sent to the controller to brake the motor. A circuit breaker switch will be used, to simulate the train coming into the main station when 24VDC is applied to the solenoid, to release the hydraulic cylinder for motor control. To simulate the train when ready to leave the station, the power will be switched off. Software Design The hydraulic system will be controlled by an Arduino Uno R3. This controller will be the brain of the entire operation. It will receive signals from PLC controls, and then output signals to drive the H-bridge which then drives the motor. While controlling the motor, it will receive signals constantly from the PLC controls as well as the force sensor and IR sensor.

11 The beginning stages of the software in the will identify the input and output signals. The main input signals are the PLC relay controls. These are isolated by the optocoupler previously discussed. These analog signals will be sent to digital pins on the controller and thus be interpreted as digital. Therefore if the forward, reverse or stop signal is sent to the system, a digital value of 0 will be received. When those controls are not on, then they will send a digital value of 1. After the controller receives these signals from the PLC unit it will drive the H-bridge with a PWM depending on which signal is received. It will either drive the motor in reverse, forward or even stop/brake if it is already in motion. A redundancy form of logic will be introduced into the system to prevent the controller from receiving a reverse input if the forward input is already on and vice versa. The only way to completely stop the current function and run the opposite, is to turn the stop function on and the current function off. This will prevent any errors in the motor operation. During motor operation in the forward direction the force sensor will be active. The controller will constantly be reading the analog input by the force sensor to determine when a value of 5lbs is reached. It will then brake the motor and break from the forward motor function. In summary while the motor is in either the forward or reverse direction, a PWM will be applied to the motor. In the reverse function of the motor, it will be driven with PWM over so much time until the return position has been reached. At that time, the motor will brake and the function will break out and continue to loop for PLC input control signals. If the forward function of the motor is in operation, the motor will be driven with PWM over so much time until the IR position sensor inputs a specific threshold value. After that, the motor will brake, and will break out of the forward function to continue to loop for PLC input control signals. Testing: Following the completion of detailed design, production drawings were sent to both the RIT machine shop as well the customer s machine shop. Simple components were machined at RIT while more complex hydraulic parts were sent to the customer s shop. Simple subsystem tests were performed throughout the assembly process. Most tests consisted of ensuring that the components and seals would properly assemble. This led to minor modifications to some parts. Once modifications were implemented, the basic operation of the ball screw drive system was tested by turning the screw by hand to mimic the torque supplied by the motor. As expected, the ball nut provided very little resistance to motion. However, a design flaw was found which involved the screw not being properly held by the end support. It was determined that this flaw would impede efficiency but would still allow the cylinder to operate properly. Although design changes will be needed in the future, assembly was completed, and the system was mounted on the test stand. The test stand was constructed from an 80/20 base and aluminum machined components. The cylinder was mounted fixed on the motor end and connected to a lever on the rod end to allow full extension and retraction as seen in Figure 16. After mounting the cylinder on the test stand, the system was tested with no hydraulic fluid to ensure that the motor would properly extend the rod. This test was successful as the system was able to fully extend and retract. Data was not recorded for these tests since they do not capture actual operation of the cylinder due to the lack of oil. Although data was not recorded, the ability to vary motor speed and torque was verified. Following validation of proper movement, the cylinder was filled with hydraulic oil. This was done by connecting each side of the cylinder to a reservoir of hydraulic oil and then actuating the cylinder in and out until air bubbles ceased to exit the cylinder.

12 Figure 16. Side view of test stand with cylinder mounted. Cylinder retracted (left) and extended (right). At the beginning of the project, many tests were planned, including vibration testing, testing to failure to determine life expectancy, full ASTM load testing, and oil cleanliness testing. Due to time, resources, and budget constraints, many tests could not be performed. At this time only simple movement tests with oil have been performed. These tests were mainly used to determine if the system can actuate in the required amount of time (results of these tests are discussed below). Due to time constraints, further testing will be performed before customer handoff and will be documented on P18318 s EDGE page as results become available. These tests will validate the use of the force sensor and IR sensors to provide input on the cylinder s motion. Tests may also be performed with low loads to ensure that the cylinder will be able to move an actual restraint. Once these tests are completed, the system will be sent to ACM for a more comprehensive test plan. ACM will test the amount of weight the system is able to move as well as apply loads to the cylinder to determine if the hydraulic locking system is working properly. As previously mentioned, at the time of this publication, only simple movement tests have been performed. Most of these tests were at low speeds to validate proper system movement and sealing. One test was performed at a higher speed to determine if the 5s movement time was achievable. During this test, the system extended in 5.52s and retracted in 4.85s when filled with oil and no load applied to the rod end. The system was slightly slower in the extend direction due to the need for the motor to overcome additional friction from the pulley contacting the rear cover. Tests at higher speeds were not performed due to the insecurity of the ball screw. Further testing will need to be performed with loads applied to determine the minimum movement time under load. An image of the test setup can be found in Figure 17. Figure 17. Cylinder assembled and mounted in test stand. Future Recommendations:

13 Throughout every phase of the project, methods for improving the system were documented. As expected with any prototype, many beneficial design changes could be made to improve the product. Many of these changes are simple model or drawing changes to allow for easier fabrication, sourcing, and assembly. A major design change that is needed is a redesigned ball screw to allow for proper operation of the end support which holds the screw. The current configuration causes one of the pulleys to rub against the rear cover which slows the system as it must overcome the friction between the pulley and the rear cover. Other improvements include the inclusion of filters to prevent valve clogging and a better method to restrict the radial motion of the nut. The most critical design change is reduction of the weight of the system. In the current state, the prototype cylinder weighs approximately 30 lb. without hydraulic fluid. This is well over the required weight specified by the customer. This large weight increase is largely due to the increase in size of the steel inner and outer barrels as well as the addition of the steel ball screw and the motor. The increase in size of the barrels was driven by the required piston size increase needed to accommodate the ball nut and positioning rods. By redesigning the positioning rod system, the barrel sizes could be decreased. This also allows all other components to be decreased in size, decreasing the overall weight of the system. Along with increasing the size, of the barrels, the wall thickness of the barrels was also increased due to available stock sizes. Further research and sourcing should be performed to find stock with a thinner wall and lower weight. Finally, after testing is performed and motor performance is analyzed, the motor choice will be reviewed, and a smaller, less powerful motor may be acceptable for this application. To better understand how the cylinder will hold up to the amusement ride environment, it is recommended to perform vibrations testing, accelerated testing, and maximum G-force testing. Additionally, the use of filters was considered to catch debris and confine them to the inner barrel. Before integrating this into a further design iteration, there should be some testing to determine the effects the filters have on the fluid flow in the cylinder. The purpose of vibration testing is to ensure the cylinder can withstand the stresses and strains experienced at different frequencies of vibrations and won t fail from fatigue. This testing can also ensure proper sealing between parts. The cylinder is to be fully assembled, filled, and sealed and tested at different frequencies for both long term and short term. Accelerated testing for a cylinder determines the number of operating cycles until failure. This can be accomplished by running the cylinder continuously at high loads, extreme temperatures, and great speeds. The most effective in this instance, and easiest to extrapolate results from, would be exposing the cylinder to large loads. This will create stress on critical components including the motor, ball screw, and oil. Debris from component wear can get into the hydraulic oil and clog up orifices, inhibiting fluid motion. A possible remedy to this is introducing filters on each side of the barrel so any debris is confined and can t reach the orifices. Potential issues that need to be explored include determining filter mesh size, if the filter can withstand the forces put on it from the fluid, the effects on the oil flow rate, and if it causes debris to collect and disrupt the flow. Acknowledgements:

14 Eric Wellington Advanced Concepts in Manufacturing Primary customer, hydraulic guidance, financial assistance. Don Pophal Project Guide Ed Hanzlik Project Guide ASTM International Project sponsorship Machine Shop Staff, Jan Maneti, Craig Arnold, Rick Wurzer Design guidance, fabrication, assembly assistance Gary Hodenius Drafting and manufacturing guidance Jim Styner Design guidance Chris Williams Calculation Guidance