- Dr. Dermot Brabazon, School of Mechanical and Manufacturing Engineering, Dublin City University
Evaluating the impact characteristics of Irish hurling balls to measure performance and define quality standards for the sport.
Rapidly designing an easy-to-use automated test system rapidly with NI LabVIEW and NI vision hardware and software to capture and analyze ball impact data at high velocities.
Fiachra Collins - Dublin City University
Dr. Dermot Brabazon - School of Mechanical and Manufacturing Engineering, Dublin City University
Dr. Kieran Moran - School of Health and Human Performance, Dublin City University
In recent years, variability in behaviour of the sliotar, the ball used in the Irish sport hurling, has become evident in championship matches. The resulting controversy among players and the media has led the Gaelic Athletic Association, the governing body for the sport, to adopt a single standardised core for the sliotar. To produce a core with consistently desirable playing performance, it was necessary to characterise the dynamic impact properties of approved sliotar cores. Data had to be analyzed on the coefficient of restitution (ratio of velocities before and after impact), impact force, deformation characteristics, dynamic stiffness, and contact time.
To test within the conditions occurring in the sport, the characterisation requires impact velocities of up to 86 mph and impact durations in the region of 2 ms. This requires high data acquisition rates for data capture and analysis. We designed an automated test system using LabVIEW and NI vision devices to rapidly acquire the data and analyze it in a user-friendly graphical user interface.
In this application, a pneumatic system propels sliotars using a modified actuator with large ports and a light polymer piston. An air reservoir acting as a buffer is located immediately before the directional control valves to avoid restrictive choked flow developing in the narrow apertures of the pneumatic system. Controlled by LabVIEW via a multifunction data acquisition (DAQ) board and an NI SCB-68 breakout box, the pneumatic system projects the ball vertically upward with precise aim and zero spin to strike an impact plate. The steel impact plate is firmly supported to ensure a negligible contribution to the energy loss of the impact and it is inclined at 4 degrees to ensure the ball does not rebound back down the barrel into the firing area.
The impact characteristics are measured from data acquired through a compression load cell via an NI PCIe-6351 and a high-speed camera via an NI frame grabber. Due to the short exposure times associated with high-speed footage, the impact area is illuminated by a 400 W halogen light with a cooling fan installed beside the impact plate. After impact, the ball rebounds within an enclosed frame and rolls into the water conditioning unit. This allows us to test the balls under wet conditions to investigate the effect of rain on performance. The ball finally rolls into a feeder channel that can accommodate up to 14 balls tested continuously in a cycle.
Nine synchronized LabVIEW VIs control the entire system with a user-friendly interface. The top-level VI allows the input of the ball identification names and the desired projection velocity, executes the subVIs in the correct sequence, and displays the impact characteristics following analysis of the acquired data.
Using LabVIEW to control the pressure in the reservoir and the triggering signals to the directional control solenoid valves, the actuator can accelerate up to 38 m/s in 12 ms. The timing of the directional control valves is crucial in permitting the maximum use of the extension stroke to accelerate the ball and decelerate the piston before it strikes and fractures against the end of the stroke.
We used a Mikrotron MC1302 high-speed camera to acquire high-speed footage of the impact via a base Camera Link connection to the NI frame grabber. This camera has a full frame resolution of 1,280 by 1,024 pixels at 100 frames per second (fps); however, we found that 5,000 fps at 220 by 110 pixels proved to be the optimal setting for recording the impact duration. Serial commands in LabVIEW specify the frame rate, resolution, and exposure settings of the camera.
An LL ring buffer algorithm delivers high-speed image acquisition, which creates the memory allocation for the frames before acquisition, and rapidly saves each frame in an uncompressed AVI codec for subsequent image processing. To prevent excessive use of the CPU and disk space and reduce the time taken for image processing, the duration of acquisition is set to 90 ms. This requires precision timing and triggering of the camera when the ball is projected.
The signal voltage from the compression load cell is amplified to the 0 to 10 V range by an RPD DR7DC transducer amplifier. This voltage is acquired at 10 kHz by a high-rate analogue input and converted to the force reading by the calibration built into the program.
In the interval of time between impact and the ball re-entering the feeder channel, the test system automatically analyses the data, displays it to the user, and saves it in spreadsheet format. The top-level LabVIEW VI passes the path of the AVI to the image processing algorithm, which uses NI Vision Development Module to perform a variety of functions to threshold, filter, and sharpen the image. This ultimately isolates the ball from the background and tracks the centre coordinates and pixel area from frame to frame via the shift register. The resulting measurements include the following:
Using LabVIEW, the program analyses the force reading to evaluate the peak force, contact time, and impulse. The impulse is the change in momentum of the ball and is calculated from the integral of force with respect to time. The force is plotted against the deformation to produce a hysteresis curve with the slope corresponding to the dynamic stiffness and the area enclosed within the hysteresis loop corresponding to the dynamic energy loss.
By implementing the intuitive programming of LabVIEW and ready compatibility with both NI and third-party hardware, we were able to quickly set up the system. In addition, for system calibration, the application had excellent repeatability with accuracies of ± 0.14 m/s, ± 0.58 mm, and ± 0.7 N for velocity, deformation, and force measurements respectively.
Through data analysis we determined the variation in performance between different brands and even within the same brands of sliotar cores. This variation becomes more pronounced at higher velocities because the results are more scattered with increasing impact velocity. This is due to different material properties, nonuniform ball construction, nonuniform mass distribution, and imperfect sphericity, which are results of insufficient quality control in the manufacturing process.
The velocity-dependent deviation in performance would not be apparent with current sliotar regulation testing. Using LabVIEW and NI vision hardware and software to build a high-performance test system was critical to comprehensively characterise a ball’s impact at high speeds. In addition to investigating the relationship between the material properties and ball performance, the system can also aid in future enforcement of the new quality standard through ball testing and quality control.
Dublin City University
School of Mechanical and Manufacturing Engineering,Dublin City University,
Glasnevin, Dublin 9,