An assessment of the mechanics of protected and unprotected head impacts in cricket using representative impact conditions and a sport-specific anthropomorphic test device

The fundamental aim of the research presented in this thesis was to investigate the mechanics of protected and unprotected head impacts in Cricket through the development of enhanced, laboratory based projectile tests. In order to develop an improved test methodology, research was conducted to; (1) establish the properties of cricket balls at realistic loading rates, (2) determine appropriate impact conditions to represent those seen in real-life impacts, and (3) develop a novel, sport-specific headform (LU headform), that could be used in protected and unprotected impacts. An additional aim was to assess the validity of the impact attenuation test currently specified in the British Standard for head protectors for cricketers (BS 7928:2013), by comparing the mechanics of the impacts observed in these drop tests to those observed in more realistic projectile impacts.
The properties of two elite level Cricket balls (Dukes Special County and Kookaburra Turf) and one Cricket training ball (BOLA) were assessed at three nominal initial loading rates c. 4.5, 18 and 31 m/s). This study found that Cricket balls exhibit different characteristics based on the construction and materials used, the orientation of the ball at impact, and the level of wear. The Dukes ball were at least 15% stiffer than the Kookaburra balls, and the Kookaburra balls were at least 9% stiffer when impacted perpendicular to the seam than when impacted parallel to the seam. The BOLA ball was found to be at least three times more compliant than the two Cricket balls tested. After 20 repeated high-speed impacts, the Dukes and Kookaburra balls were found to be at least 36.3 and 20.5% more compliant respectively. As a result of these findings, the Kookaburra ball was selected for future impact tests due to improved ball to ball variation and reduced ball degradation. Non-destructive 4.5 m/s impact tests were also integrated into projectile tests to ensure that the ball properties remained within 5% of the pre impact stiffness.
A representative impact speed was determined by using ball tracking data collected during match play. A database containing measured release speeds of 447 elite level bowlers was used to identify the maximum observed release speed (43 m/s). Ball tracking data was also used to determine the release speed and the speed at which the ball passed the batsman (inception speed) from a sample of short length deliveries. A ~worst case impact speed was determined by dividing the highest release speed in the database by the lowest percentage change in speed from release to inception. This resulted in a nominal impact speed of 34.7 m/s. Three anatomically anchored impact locations were defined to generate linear and angular motion in six degrees of freedom. Each headform was suspended using bungee cords to simulate the passive stiffness of the human neck and therefore a ~worst-case scenario with respect to the observed dynamic response.
Due to the construction and materials used in commercially available headforms, the development of a headform that was capable of producing realistic first order dynamic responses during projectile impacts was deemed necessary for the assessment of protected and unprotected scenarios (LU headform). External soft tissue, bone and brain components were modelled in Siemens NX 10.0 based on CT and MRI scans. The geometry of the model was manipulated to match that reported for a 50th percentile UK male, and tissue thicknesses were measured to be within previously reported ranges. The inertial properties (including principal moment of inertia, Moment of inertia about axes at the centre of gravity and location of the Centre of Gravity) were theoretically calculated and empirically measured to be closer to the average values reported for human heads than commercially available headforms. The inherent validity of the LU headform was established through material tests which showed each component to be comparable to values reported for human tissue. In addition, the resonance frequency of the skull component of the LU headform was found to be comparable to that reported for dry human skulls. The response of the LU headform during drop test and projectile impact tests was found to be within the range reported in human cadaver responses with similar impact conditions.
This research facilitated the development of a test methodology to investigate the mechanics of head impacts observed in Cricket, that was superior to that achieved in previous research. Four headforms (EN 960, Hybrid-III, NOCSAE and LU) were instrumented with a ±2000 g triaxial linear accelerometer and ±6000 deg/sec triaxial angular rate sensor and subjected to projectile impacts at three locations, when protected by two helmet types used at the professional level in the 2018 season. The Hybrid-III and LU headforms were also subjected to unprotected impacts at the same three impact locations.
Headform type and impact location was found to significantly influence the observed dynamic response during the projectile impacts. The EN 960 headform was found to be unsuitable in this impact scenario as the sensor mounting block introduced non-biofidelic frequency artefacts that could not be removed without significant signal distortion. When using the LU headform, frequency artefacts were also evident in the measured signal that were in line with the measured resonance frequency of the skull component. This concurs with the findings of previous research using human cadavers with similar impact conditions (Raymond et al., 2008) and was therefore considered a legitimate response phenomenon. The NOCSAE headform produced a response comparable to the LU headform. Alternatively, the Hybrid-III responded predominantly as a rigid body and therefore overlooked the potentially important resonance frequency excitation. As a result of the resonance frequency excitation, the LU and NOCSAE headforms produced higher peak resultant linear acceleration, impulse, head injury criteria (HIC), maximum angular velocity and brain injury criteria (BrIC) values, in addition to longer contact times.
This study is the first of its kind to assess unprotected head impacts in Cricket and therefore the first opportunity to compare the true performance of helmets relative to unprotected scenarios. Both helmet types tested were found to reduce the linear and angular response observed during impact, and prevent skull fracture that was observed at one location when using the LU headform. When using the Hybrid-III headform, the peak resultant linear acceleration, maximum angular velocity, HIC and BrIC values were reduced by at least 39.9, 37.9, 69.6 and 32.3% respectively when helmeted. When using the LU head form, peak resultant linear acceleration, maximum angular velocity, HIC, BrIC calculated using maximum angular velocity and BrIC calculated using steady state angular velocity were reduced by at least 32.9, 25.7, 15.0, 32.05 and 13.86% respectively when helmeted. When using the LU head form the HIC values were found to be alarmingly high, even in the helmeted scenarios. However, as the HIC was not developed for short duration impacts where resonance frequency excitation is likely to occur, this metric should be used with caution. These results highlight the need for further investigations into injury mechanisms in projectile head impacts and the development of injury metrics specifically for this impact scenario. Repeated impacts at one impact location on the helmets were found to significantly reduce their ability to attenuate the observed linear and angular response, despite minimal external damage.
Drop test were also conducted as specified in the current British Standard (BS 7928:2013) with the same helmet types and impact locations used in the projectile tests. The drop tests were found to produce contact time and time to peak resultant linear acceleration at least 61.4 and 59.9% longer than the projectile tests respectively. With the exception of one helmet type at one impact location, peak resultant linear acceleration was at least 161.8% lower in the standard drop tests. From these results it can be concluded that although the standard tests achieved what it intended to (i.e. preventing skull fracture), the mechanics of these impacts are unrepresentative of those seen in real life impacts.
Overall, the research presented in this study facilitated the assessment of laboratory-based head impacts representative of those seen in Cricket and highlighted the importance of suitable human surrogates during testing. The results presented in this thesis can be used in future work to; improve helmet design (through access to more realistic response data), develop and validate finite element models to investigate injury mechanisms, and in the development of impact specific injury metrics