A Method for Assessing the Overall Impact Performance of Riot Helmets

246 JOURNAL OF APPLIED BIOMECHANICS, 2003, 19, 246-254 2003 by Human Kinetics Publishers, Inc. A Method for Assessing the Overall Impact Performance of Riot Helmets Jean-Philippe Dionne, Ismail El Maach, Ahmed Shalabi, and Aris Makris Med-Eng Systems, Ottawa The objective of the present paper is to investigate the overall impact performance of various riot helmets in a comparative study. The National Institute of Justice (NIJ-0104.02) and the Canadian Standards Association (CSA-Z611-02) standards regulate the use of riot helmets in North America. Both sets of standards have a number of requirements for impact performance. Impact tests carried out with the use of a drop tower apparatus compliant with NIJ test protocols demonstrated large differences in impact attenuation level among the helmets from six manufacturers in terms of frontal and lateral impacts to the shell, and face-shield deflection. For instance, the impact energy yielding a headform acceleration of 300 g s was measured for each helmet for frontal impacts on the helmet shell. Values ranging from 69 J up to 171 J were obtained. The energy levels of typical crowd-control threats, e.g., baton blows and projectiles, were quantified and compared with the impact energy values used in the standards. It is observed that the NIJ face-shield deflection requirement is low as compared to actual riot threats, whereas the CSA requirements are more in line with these threats. A novel method was devised to objectively assign a global impact performance score to each helmet. This method takes into account the frontal and lateral impacts to the shell as well as the face-shield deflection tests. It is based on the directional origin of the threat and the geometry of the helmets (frontal percentage area of the visor). From these global performance scores, it is possible to obtain a ranking of the various riot helmets used in the present comparative study. Based on the analysis of the global scores, it was found that appropriate protection of the face (through an impact resistant visor) is the key feature for a helmet that will be used in riot environments. Key Words: head protection, impact energy, crowd control, standards, drop tower The authors are with Med-Eng Systems, 2400 St. Laurent Blvd., Ottawa, Ontario, Canada K1G 6C4. 246

Assessing Riot Helmets 247 Introduction Blunt impacts to the head cause differential movement of the brain relative to the skull, disrupting the neural tissue and producing concussive and other traumatic brain injuries (Ryan, 1992). Although differential movement is the most relevant parameter related to injury prediction, it is difficult to measure experimentally. As a result, most commonly used criteria for head injury are based on the linear acceleration of the head. A linear acceleration value of 300 g s (1 g = 9.8 m/s 2 ) is used in a number of riot helmet standards, including that from the Canadian Standards Association (CSA, 1986) and that from the U.S. National Institute of Justice (NIJ, 1984). A value of 300 g s is also recommended for other head protection systems, such as cricket batting helmets (Stretch, 2001). A more conservative value of 200 g s has been suggested for sport impacts (Naunheim, Standeven, Richter, & Lewis, 2000). Figure 1 Posttest pictures of the riot helmets used in the present study. Riot helmet standards also require that manufacturers comply with specific impact energy acceptance criteria for the visor. There are no tolerance levels specified for the criteria, making it difficult to compare the actual level of impact attenuation offered by the various helmets, and eventually rank their performance. Based on this premise, a series of comparative tests involving riot helmets from 6 manufacturers (illustrated in Figure 1) was conducted by an authorized laboratory (Riot Helmet Testing Report, 2001). The relative impact attenuation of each helmet for three different impact sites was examined, and a global scoring mechanism was proposed and discussed. The impact energies generated in this study were also compared with those from typical riot threats.

248 Methods Three types of tests were carried out: the resistance of the helmet shell to frontal impact, the resistance of the helmet shell to side impacts, and the face-shield deflection test. The experimental apparatus, shown in Figure 2, conformed to the NIJ standard. For front and side impacts, a partial headform (size 7 1 / 4 ) was fitted inside the helmet. In the experimental procedure, the only relevant difference between the side and front impact tests was the orientation of the helmet assembly with respect to the fixed anvil, as shown in Figure 2. The helmet assembly was dropped from various heights (corresponding to different impact energies) onto a hemispherical anvil. The impact energy was computed from E = 1 / 2 mv 2, where E is the impact energy, m is the mass of the headform assembly (excluding the helmet), and v is the velocity at impact. An accelerometer was embedded inside the headform to determine the acceleration (or deceleration) of the headform when the assembly hits the anvil. The acceleration signals were processed through a computerized data acquisition system. In the present paper, the impact energy at which the helmet and headform assembly undergoes a 300-g acceleration is referred to as E 300. This acceleration value is used as the threshold in both the CSA and NIJ standards. Given the destructive nature of these tests and the limited number of samples (5 helmets) from the 6 manufacturers, the E 300 impact energy value was found by generating impacts of various energies, and interpolating linearly for the impact energy yielding an acceleration of 300 g s on the acceleration-energy curve. Impacts of increasing energy were first generated, until a value above 300 g s was obtained. This was then followed by a final bracketing around this acceleration level. Only one drop was performed at each of the 5 energy levels tested. Each helmet was impacted only once at each impact site, in accordance with the NIJ and CSA requirements. For the face-shield deflection test, instead of measuring the acceleration of the headform, it was verified whether an impactor dropped on the nose area (worst-case scenario, since this represents the smallest headform-visor gap) from a given height would cause the face-shield to come in contact with the nose of the headform. Therefore a different apparatus, which also complies with the NIJ requirements, was used. The helmet was attached, using its retention system, to a stationary anthropomorphic headform, with the visor facing up. A 1.0-kg impactor was dropped from various heights corresponding to different impact energies. A 2.0-kg impactor with the same impact area was used to generate larger energies. The impact energy was then calculated from the same equation as above, except that m represents the mass of the impactor only. Contact sensors allowed for the detection of contact between the visor and the nose of the headform during the impact, in which case the helmet failed the test. The impact energy for nose contact E c was calculated by taking the average of the highest pass energy and the lowest fail energy. In order to determine an overall performance criterion, taking these three impact tests into account, a global score system is proposed. All impact results were first normalized with respect to the NIJ requirements (111 J for frontal and side impacts, 7.85 J for face-shield deflection), used as a reference. Since it was estimated, through interaction with end users, that about 85% of the threats in a

Assessing Riot Helmets 249 Figure 2 Experimental setup for (a) front impact, (b) side impact, and (c) face-shield deflection. In (a) and (b), a hemispherical anvil is used. In (c), the impacting surface is hemispherical.

250 crowd management environment are frontal, a combined weight of 85% was given to the frontal impact and face-shield deflection tests, and the remaining 15% was given to the side impact tests. The 85% was then subdivided according to the percentage of frontal area occupied by the visor relative to the total frontal area. The global score is thus obtained by the following formula: E 300 E 300 Score = ( ) 15% + [( ) ( ) + ( )( )] 85% 111 J 111 J Front A frontal 7.85 J A frontal Side A shell E c A visor where 111 J and 7.85 J correspond to the NIJ requirements, A shell is the frontal area occupied by the helmet shell, A visor is the frontal area of the visor, and A frontal is the total frontal area of the helmet (including the shell and visor). This global system is thus based on the impact performance of the various helmets with respect to the NIJ requirements, in objective and well-defined impact tests, and on the direction probability of the incoming threats in crowd management operations. Finally, the measured E 300 and E c values were compared to the energies corresponding to typical riot threats. These typical threat energy values were obtained in a parallel experimental study in which three healthy male participants had to throw objects from a distance of 10 meters, typical of rioters/police standoff. The maximum and minimum energies were recorded, through high-speed video, just prior to impact for a golf ball, a billiard ball, a full beverage can, a half brick, and the blow of a baseball bat (see Table 1). The use of energy as the selected impact parameter makes it possible to compare impacts from impactors and projectiles of different masses, and in different configurations. Table 1 Measured Energy of Several Riot Threats Mass Minimum Maximum Object (g) energy (J) energy (J) Golf ball 46 10.3 19.6 Half brick 936 134.4 214.1 Billiard ball 170 33.1 82.3 Beverage can 372 68.3 132.3 Baseball bat 470 96.6 174.1 Results For all helmets, a steep increase was observed in the acceleration-energy curve, after the relatively flat initial portion of the curve. This may indicate that the protective foam liner became completely compressed and ended up transmitting most of the force to the head. This behavior was also observed with football headgear (McIntosh & McCrory, 2000).

Assessing Riot Helmets 251 For frontal impacts, Helmet D performed the best with the highest E 300 value of 171 J, closely followed by Helmet A, with 162 J. Helmet B was found to perform poorly based on a low E 300 of 69 J, which clearly indicates that it fails the NIJ requirement of 111 J. All helmets were found to perform somewhat better on the side than on the front (higher E 300 values). The best performer was Helmet E with an E 300 value of 231 J, much higher than the front impact results. This can be explained by the reinforcement of the helmet near the side location where the visor is attached. Helmet B performed poorly, with a value of 102 J, again below the NIJ requirement of 111 J. The best performer for the face-shield test was Helmet F. Due to physical limitations of the drop tower used (maximum height), a 2.0-kg impactor had to be used to determine the value of E c for Helmet F (95 J). Impacts with the 2.0-kg impactor are more severe since the linear momentum for the same impact energy is larger. The result for Helmet F is therefore probably conservative (its E c value with a 1.0-kg impactor may be higher). The next two best performers are Helmet E and Helmet D, with E c values of 58.2 J and 48.0 J, respectively. Table 2 Summary of Impact Testing Results (in J) Impact attenuation Face-shield deflection Helmet E 300 E 300 Highest Lowest E c model front side pass fail visor A 162 176 10.1 15.2 12.6 B 69 102 17.7 18.7 18.3 C 144 165 18.7 19.9 19.3 D 171 194 45.5 50.6 48.0 E 140 231 55.6 60.7 58.2 F 152 190 90.0 100.0 95.0 The summary of all quantitative results for front impacts, side impacts, and face-shield deflection (including the highest pass and the lowest fail) is shown in Table 2. The E 300 and E c results are illustrated in Figure 3, together with the energy levels associated with the various riot threats, shown on the right side of the graph, with the arrows indicating the range of possible energies and the dotted lines representing the average values. The global scores obtained by each helmet are given in Table 3, along with the percentage of frontal area occupied by the visor. The global scores are also illustrated in Figure 4, where the performance of the helmets in each of the three impact tests is illustrated with different colors. Based on this global scoring scheme, the best overall performer is Helmet F with a score of 8.54. The next best helmet is Helmet E with a significantly lower score of 5.19. Sensitivity analysis of the selected percentage for side impacts (15%) indicates that the ranking of the 4 best performing helmets would only be affected for weights of side impacts in excess of 90%, which is not realistic.

252 Lateral Impact Frontal Impact Faceshield Impact Energy [J] Beverage Can Half Brick Billiard Ball Baseball Bat Golf Ball Helmets Figure 3 Energy levels for 300-g acceleration thresholds (E 300 ) and impact energies causing contact between the nose and the face-shield (E c ) compared with impact energies of typical riot threats for front and side impacts. Table 3 Global Score for Impact Performance of the 6 Helmets Helmet Visor area (%) Global score A 68 1.56 B 71 1.70 C 75 2.07 D 67 4.18 E 73 5.20 F 78 8.54 Discussion It was observed that the shell of Helmet B is only capable of protecting the wearer against golf balls and billiard balls. All other helmets could protect the wearer against more energetic threats such as a full beverage can, and an average baseball bat blow (although some may fail in the higher range). According to these results, it seems that no helmet can safely protect against the average impact energy of a half brick. However, the impacts generated using a drop tower, on which the E 300 values are based, are more severe than real life threats, as the impacted

Assessing Riot Helmets 253 Lateral Impact Frontal Impact Face-shield Deflection Global Score Helmets Figure 4 Global score assessed to each helmet based on the weighted balance of the three experiments carried out. specimens are constrained (i.e., no specimen motion is allowed). Moreover, the impacts are more localized, normal to the surface, and deflection of the impactor is not possible (i.e., worst case scenario). Therefore, care must be taken when comparing the values from Table 2 with the E 300 values. Only the energies from the golf ball, the billiard ball, and the beverage cans appear in that range. Based on this observation, the face-shield is not as impact-resistant as the helmet shell, as expected due to the material constraints for a transparent visor. However, direct comparisons cannot be made, since accelerations are not measured for the face-shield. Although Helmets E and D were outperformed by Helmet F for the faceshield deflection test, they exhibited relatively high values of E c, given that their respective visors could be readily deformed or bent manually. The relatively good performance of these two helmets may be attributed to their good retention systems, since the face-shield deflection tests often failed due to the pivoting of the helmet over the headform in the forward direction, rather than being caused by the deflection of the face-shield, which yields the same injurious effect on the face of the wearer. The global scoring presented in the present paper provides an objective and quantitative method with which to rank the performance of riot helmets, with respect to the main impact requirements from the NIJ and CSA standards. References CSA Riot Helmet Standard. (1986). CSA Z611-M86. Toronto: Canadian Standard Association. McIntosh, A.S., & McCrory, P. (2000). Impact energy attenuation performance of football headgear. British Journal of Sports Medicine, 34, 337-341.

254 Naunheim, R.S., Standeven, J., Richter, C., & Lewis, L.M. (2000). Comparison of impact data in hockey, football and soccer. The Journal of Trauma, 48, 938-941. NIJ Standard for Riot Helmets and Face Shields. (1984, Oct.). NIJ Standard 0104.02. Washington, DC: U.S. National Institute of Justice. Riot Helmet Testing Report. (2001, Jan.). Maximum energy testing of various riot helmets. Document R2001-002/bp. Biokinetics and Associates, Ltd. Ryan, G.A. (1992). Improving head protection for cyclists, motorcyclists, and car occupants. World Journal of Surgery, 16, 398-402. Stretch, R.A. (2001). The impact absorption characteristics of cricket batting helmets. Journal of Sports Science, 18, 959-964. Statement on financial interest in the research Most of the requirements from the commonly used riot helmet standards are GO/ NO-GO types of tests, making it difficult to quantitatively compare the impact performance of helmets based solely on conformity with these standards. Moreover, there is no standard method for quantitatively evaluating and comparing the overall impact performance of riot helmets, taking into account the performance of both the helmet shell and visor. For this purpose, Med-Eng Systems carried out the experimental study discussed in the present paper to generate a convenient comparative tool. This study was financed by the company s R&D funds.