Design of ventilated helmets: computational fluid and impact dynamics ...

International Journal of Crashworthiness Vol. 13, No. 3, June 2008, 265–278

Design of ventilated helmets: computational fluid and impact dynamics studies Praveen K. Pinnoji, Zafar Haider and Puneet Mahajan∗ Department of Applied Mechanics, Indian Institute of Technology Delhi, New Delhi, India (Received 13 December 2007; final version received 14 January 2008) The existing motorcycle helmets are thermally uncomfortable as there is no provision for air flow inside the helmet. A new design of helmet, with grooves in the liner foam and slot in the outer shell and liner foam to improve the ventilation, is proposed. Computational fluid dynamics studies show considerable improvement in air velocities inside the helmet in the presence of grooves and slot. Impact dynamics and biomechanics studies on various motorcycle helmets with deformable and rigid head show that the proposed design meets the requirements of the standards in the drop test. The optimum size of the groove and number of grooves for a motorcycle helmet are decided on the basis of the above studies. Keywords: helmets; design; fluid flow; impact; finite element; biomechanics

1. Introduction A well-designed motorcycle crash helmet has proved to be a very good protection device for the rider to prevent or minimise the head injuries in road accidents. If a helmet is not worn, the head impact with any object would cause localised high pressure on the skull, which leads to brain injury. The helmet design can be divided into functional (like shockabsorbing capability, penetration resistance, retention and reliability) and non-functional (like low cost, good aesthetics, comfort, light weight and good thermal characteristics) categories. Though a helmet is well-designed for functional characteristics, because of weak non-functional characteristics drivers sometimes dislike wearing it while riding. In South Asia, excessive sweating and resulting discomfort due to hot and humid weather conditions discourage motorcycle riders from using helmets unless it is mandatory by law. The space between the head and helmet is small, and both mass flow and air velocities in this gap are also low; as a result, the sweat is unable to evaporate making the rider uncomfortable. The discomfort caused by sweating can be reduced by increasing the air velocities inside the helmet so as to enhance the sweat evaporation rate. Air flow in helmet can be improved by large ventilation openings as in bicycle helmets, but unfortunately such ventilation openings may be detrimental to the safety and structural integrity of the helmet. Most of the studies on motorcycle helmets are based on the material and biomechanics aspects, and few studies exist which investigate the effect of ventilation on air flow inside the helmet or effect of this (or ventilation openings) on the dynamic performance of helmet. It is possible that helmets with ventilation are ∗

Corresponding author. Email: [email protected]

ISSN: 1358-8265 C 2008 Taylor & Francis Copyright DOI: 10.1080/13588260801933626 http://www.informaworld.com

available in the market, but systematic studies on these are not available in the literature. In first part of the paper, we investigate air velocities inside the helmet, with and without ventilation using computational fluid dynamics (CFD) techniques. The fluid flow study was carried out to examine the possibility of improving the ventilation in motorcycle helmets. In second part of the paper, the biomechanics characteristics of head impact were studied for helmets with and without ventilation using finite element (FE) analysis. In a motorcycle helmet, the comfort foam apparently helps in fitting the helmet on heads of different sizes, although it is rarely used in bicycle helmets. Because the comfort foam always rests on the head and gives resistance to air flow, it has not been included in the helmet designs studied here. We found that, without comfort foam and if tied properly, the helmet sits on the head and does not move. Even without comfort foam, however, there are regions in the central plane of the head, where this helmet rests on the head, and there is no space for air to circulate. It was decided to support the helmet on the head by comfort foam of 2 mm thickness provided on the sides of the helmet only (but not on the top). A groove was made in the central plane to provide space for air to flow, and a slot was provided in front of the helmet for air to enter. The groove and slot in the helmet are shown in Figure 1. Flow velocity inside the helmet was determined by varying the depth and width of the groove keeping slot dimensions fixed. Four different sizes of groove, listed in Table 1, were investigated. The dimensions of the slot present in the outer shell and liner foam were fixed at 48 mm × 7 mm. These helmets did not have

266

P.K. Pinnoji et al.

Figure 1. Three parallel grooves of 14 mm × 7 mm with slot in liner foam.

a visor. Studies were also performed on helmets without a slot. However, velocities inside the helmet were generally lower than those observed with a slot, and therefore, only the latter case is discussed here. 2. Fluid flow analysis Reischl [20] carried out an investigation of helmet ventilation designs for firefighter helmets and found that a helmet with side ventilation holes was cooler than unventilated helmet, and increasing the gap between helmet and head also enhanced the cooling action due to the improved air circulation. Abeysekera and Shanavaz [2] investigated the potential benefits of helmets with ventilation holes for industrial workers both in laboratory and field settings. Bruhwiler et al. [4] studied the heat transfer variations of bicycle helmets. They carried out experiments in water channel to study the efficiency and placement of vents in bicycle helmets. They concluded that there is significant potential Table 1. Various ventilation designs. Design 1 2 3 4 5

No. of grooves

Groove width (mm)

Groove depth (mm)

Slot (mm)

1 1 1 1 3

14 28 42 14 14

7 7 7 14 7

48 × 7 48 × 7 48 × 7 48 × 7 48 × 7

within the basic helmet structure for improving the heat transfer. The above studies except that of Bruhwiler et al. [4] were mainly in conditions where the velocities of air were very small. All the above studies were experimental, and no CFD analysis had been done for fluid flow inside the helmet. Here we study air flow inside a motorcycle helmet using CFD where the relative velocity of air outside the helmet is about 15 m s−1 . Inside the helmet the air gap varies from approximately 0 mm (where the helmet touches the head) to 10 mm, and the velocities are small. One of the ways of improving these velocities and resulting heat transfer is by providing grooves in the central plane of the helmet. This leads to increased evaporation of the sweat. The present study used CFD to investigate air flow inside the gap between the helmet and head. To get an idea of how well CFD is able to simulate the air flow in a gap of 10 mm, experiments were performed in a wind tunnel on two concentric cylinders with a gap of 10 mm between the cylinders and then simulated with CFD. Such experiments with a helmet can be performed, but measurements in small gap regions would be difficult with the Pitot tube used here.

2.1.

Experiments in wind tunnel and validation of CFD

Experiments were performed on concentric cylinders in the wind tunnel of cross section 450 mm × 750 mm. The cylindrical model was mounted with axis of the cylinders

International Journal of Crashworthiness

267

Figure 2. Schematic view of two concentric cylinders model.

perpendicular to the direction of flow as seen in Figure 2. The diameter of the inner cylinder is approximately the same as that of the actual head, i.e. 140 mm, and the diameter of the outer cylinder was 160 mm. Six pressure tabs were placed on the surface of the inner cylinder, and velocity within the gap at various points along the y-axis was measured using a three-hole velocity probe connected to the micro-manometer. Velocity of the wind at inlet was measured by a Pitot-static tube. Experiments were carried out at three different speeds of the wind, 11 m s−1 , 15.7 m s−1 and 23.5 m s−1 although only results for 15.7 m s−1 are shown here. Fluid flow analysis for the above concentric cylindrical model was carried out through CFD using FLUENT [6]. The upstream length was four times the diameter of the inner cylinder, and downstream length was 20 times the diameter of inner cylinder. The downstream length was kept large to ensure that the flow was fully developed. Fluid inside the gap was air with density of 1.225 kg m−3 and viscosity of 1.789 × 10−5 kg (m s)−1 . For numerical simulations the inlet velocity was taken as 15.7 m s−1 , and outflow condition available in FLUENT was used at the outlet. No slip condition was assumed at the walls. Fluid flow was assumed as steady and incompressible. The standard k − ε model [7] was used as the turbulence model with standard wall functions. Here ‘k’ is the kinetic energy of the particle and ‘ε’ is its dissipation rate. Segregated solver, which solves the non-linear equation set sequentially, was used. Convergence criteria were defined by specifying a tolerance on all FLUENT residuals such as velocities, turbulence kinetic energy and dissipation rate, which appear while solving the transport equations. The air velocities along a vertical section on the central plane of the cylinder (i.e. from top to bottom of the gap

at the central plane) in CFD simulations and wind tunnel experiments are compared and depicted in Figure 3. The air velocities in this gap are relatively lower on top and bottom of the gap compared to the centre. It can be observed from this figure that the experimental and numerical results show a reasonable match, thus validating the CFD results.

3.

CFD analysis for air flow inside the helmets with ventilation

FLUENT was next used to study the flow between the head and various ventilated helmets. Geometry of the head was

Figure 3. Comparison of air velocities with computational fluid dynamics and wind tunnel experiments.

268

P.K. Pinnoji et al.

the same as the one used later in impact dynamics analysis. The helmet geometry was taken from a commercially available helmet and had a liner foam thickness of 32 mm. The computational domain was 4750-mm length and 600-mm high. The upstream and the downstream lengths were 750 and 4000 mm, respectively. It was assumed that helmet and head are symmetrical with respect to the central plane. On account of symmetry one half of the domain was modelled. Tetrahedral grid was used for meshing and the minimum grid size was 1.2 mm. The total number of elements in different models was approximately 750,000. Boundary conditions, turbulent models and convergence criteria, here, are the same as used in the Section 2. The helmet without ventilation had regions touching the head where the air velocity is zero as air could not circulate. Figure 4 shows the velocity contours in three dimensions within the helmet–head gap with 14 mm × 14 mm groove and 48 mm × 7 mm slot. Here air velocities in the central plane are improved to 8 m s−1 near the slot, 6 m s−1 on top and 5 m s−1 at the end of the groove. In a helmet with a groove of 42 mm × 7 mm in the central plane and a slot of 48 mm × 7 mm, air velocities in the central plane which can be seen in Figure 5 are further improved to 11 m s−1 near the slot, 5 m s−1 on top and 5 m s−1 at the end of the groove. With these ventilation models, air velocities are enhanced not only in the central plane of the groove but also in other regions such as C and D in Figures 4 and 5, respectively. The air velocities in other regions are higher in the 42 mm × 7 mm groove helmet than in the 14 mm × 14 mm groove helmet. A comparison with Figure 4 shows

Figure 4. Velocity contour values within the foam-head gap with half symmetry in a helmet of 14 mm × 14 mm groove and a slot of 48 mm × 7 mm.

Figure 5. Velocity contour values within the foam-head gap with half symmetry in a helmet of 42 mm × 7 mm groove and a slot of 48 mm × 7 mm.

that a larger groove in the helmet leads to higher air velocities inside the helmet–head gap. In a helmet with three 14 mm × 7 mm parallel grooves with spacing of 35 mm between them and a slot of 48 mm × 7 mm in the front, air velocity is higher only in the central plane (i.e. the groove region) and almost zero adjacent to the central plane (region G in Figure 6) because eddies are formed in this region. It, therefore, seems that for ventilation purposes providing three grooves is not very advantageous.

Figure 6. Velocity contour values within the foam–head gap with half symmetry in a helmet of three 14 mm × 7 mm parallel grooves with spacing of 35 mm and a slot of 48 mm × 7 mm.


269

Figure 7. Streamlines on the central x–z plane in and outside the gap with 14 mm × 14 mm groove and 48 mm × 7 mm slot (slot at 30◦ to the horizontal).

In ventilated helmets, the orientation of the slot is an important parameter that affects the flow inside the helmet. The slot is provided to increase the mass flow of air in the helmet. Two slot orientations, at angle 30◦ to the horizontal and tangential to the head surface, were studied. Higher air

velocities were observed when the slot is tangential to the head. For better fluid flow visualisation, the streamlines, which show the direction of fluid flow, on the central plane for the 30◦ slot and tangential slot are outlined in Figures 7 and 8, respectively. For a slot at 30◦ to the horizontal, a vortex zone near the slot (region E in Figure 7) is formed

Figure 8. Streamlines on the central x–z plane in and outside the gap with 42 mm × 7 mm groove and 48 mm × 7 mm slot (slot at tangential to the head surface).

270

P.K. Pinnoji et al.

Figure 9. Comparison of air velocities in various helmet ventilation models.

because of obstruction to the flow, but when the slot is tangential to the head surface, the vortex zone is not formed and the flow is smooth as seen in region F in Figure 8. Air velocities along the head surface on the central plane in different ventilated helmet models are compared and shown in Figure 9. Angle θ , in Figure 9, is 0◦ at the entrance to the helmet, 90◦ at the top and 210◦ at the rear end. It can be observed from Figure 9 that there is lot of fluctuation in air velocities at the entrance. There is a decrease in velocity up to 20◦ followed by an increase up to 40◦ , along the head surface. This decrease in velocity of air in the beginning is due to the interaction of air which is entering through the slot and from below through the groove. The direction of air entering from both the sides is different, and hence the resultant magnitude of air is less. The ventilated helmet model with 42 mm × 7 mm groove and slot has a maximum air velocity of 12 m s−1 and a minimum of 3.8 m s−1 . With 14 mm × 14 mm groove and slot in the helmet, air has maximum velocity of 10.4 m s−1 and minimum of 2.0 m s−1 . Air flow is improved and similar trends are observed for all the ventilated helmet models. In the conventional helmet without ventilation, air velocities decreased in the beginning of helmet–head gap and are almost constant at the top at 1 m s−1 .

4.

Dynamics of motorcycle helmets with deformable head model

In the second part of the work, dynamics of helmet impact are studied by focusing on the biomechanic characteristics of the head. In the past, the FE analysis of drop test of helmet used a rigid head, and results were reported in the form of head injury criterion (HIC) values and accelerations of the headform. Lately, there has been an increased

interest in the FE modelling of the human head to study the various injuries [11,15,21,23]. Ruan et al. [21] studied the dynamic response of the human head impact with 50 percentile of the actual human head model. Kleiven and Hardy [15] studied the localised brain motion, intracerebral acceleration, intracranial pressures and HIC. Horgan and Gilchrist [11] constructed an FE model of the human head and predicted the brain motion and intracranial pressure changes in head impact in pedestrian accidents. Zong et al. [25] used a structural intensity approach to study the power flow distribution inside the head in frontal, rear and side impacts. The results using human head models are presented in the form of pressures, stresses and strains in the brain although a clear relation between stresses and brain injury is still to be fully established. Gilchrist and Mills [8] used a one-dimensional (1-D) analytical model to examine the dynamic response of helmets, and Brands et al. [3] carried out numerical simulation for predicting the head injury using a 3-D FE model of a motorcycle helmet. Yettram et al. [24] performed an FE parametric study of the impact response of the motorcycle helmets and used HIC for judging the crashworthiness performance. Pinnoji and Mahajan [19] investigated the ventilation in motorcycle helmets by considering the helmet as a hemisphere and the head as a cylinder. They showed that the presence of grooves improved the air velocities in the gap between helmet and head without having a detrimental effect on the dynamic performance of the helmet. Deck et al. [5] simulated a frontal impact for helmet optimisation against the biomechanical criteria using FE modelling of helmet–head. Their parametric study was based on the dynamic behaviour of helmet–head components, and the values of brain pressures exceeded the tolerance limits proposed in the literature. Otte et al. [18] found that many of the injuries in motorcycle crashes are a result of rotation of the head in oblique impacts. Aare et al. [1] investigated the oblique impacts with FE model of Hybrid III dummy head and FE models of human head and helmet. Kleiven [14] studied the effect of load directions and durations with FE model of the human head and evaluated Head Impact Power (HIP), HIC, peak accelerations and maximum principal strain in the brain. Halldin et al. [9] developed a new oblique test method for motorcycle helmets. They measured the linear and rotational accelerations by impacting the helmet at 28◦ to the horizontal axis and found helmet deformation to be larger than the oblique impacts in British Standard (BS) 6658 and Economic Commission for Europe (ECE) Regulation 22/05. Safety standards for helmets prescribe a drop test. All three impact situations – front, side and oblique – were considered here. The drop test with a flat anvil was simulated in LS-DYNA [10] using the FE models of helmet and head. LS-DYNA is an explicit FE code for non-linear and dynamic analysis. The FE analysis of helmet–head impact required input consisting of geometry, initial and

International Journal of Crashworthiness Table 2. Material properties of head [23].

Part Skin Skull (outer table) Skull (inner table) Cerebrospinal fluid Face Falx Tentorium

271

Table 3. Material properties of brain [23].

Density (kg m−3 )

Elastic modulus (N m−2 )

1200 1800 1500 1040

16.7 × 106 15.0 × 109 4.5 × 109 12.0 × 103

0.42 0.21 0.0 0.49

7 2 3

3000 1140 1140

5.0 × 10 31.5 × 106 31.5 × 106

0.21 0.23 0.23

5 2 1

Poisson’s Thickness ratio (mm)

9

Part Brain

1040

Bulk modulus (N m−2 )

G0 (N m−2 ) G∞ (N m−2 ) β (s−1 )

1.125 × 109 49.0 × 103

16.7 × 103

4.2. Constitutive model for helmet The motorcycle helmet is made of a shell, liner foam, comfort foam and strap. Outer shells are made either from a moulded thermoplastic like acrylo-butadiene styrene, polycarbonate or from a composite material with glass or carbon or Kevlar fibres. The outer shell resists the penetration of any foreign object and distributes the localised forces to a wider area causing the large volume of the liner foam to deform thus increasing its energy-absorbing capacity. A full-face helmet without visor is used in helmet–head impact analysis as it covers a larger part of the head. A nylon strap is attached to the outer shell and is used to tie the helmet to the head. The strap was assumed as linear elastic, and the properties are defined in Table 4. In the dynamics part, shell thickness and material, liner foam thickness and density are kept as constant but groove sizes are varied to study their effect on the helmet impact performance. Comfort foam is not modelled in this analysis. Material model 3 (*MAT PLASTIC KINEMATIC) available in LSDYNA was used for the outer shell in FE analysis to model the acrylo-butadiene styrene material, and the properties defined are given in Table 4. In motorcycle helmet impact, the liner foams are subjected to plastic deformations. The energy-absorbing liner foam considered here is made of expanded polystyrene (EPS), which is elasto-plastic in nature. The stress– strain behaviour of EPS foam with 44 kg m−3 density is taken from Yettram et al. [24]. Material model 63 (*MAT CRUSHABLE FOAM) available in LS-DYNA was used to model the EPS foam. This material model is an isotropic foam model and crushes one-dimensionally with zero Poisson’s ratio. This model transforms the stresses into the principal stress space where the yield function is defined. If the principal stresses exceed the yield stress they are scaled back to the yield surface and transformed

4.1. Constitutive model of head The biomechanical response of the head, by considering it as deformable, is studied in terms of forces, pressures and stresses. A 3-D FE model of human head developed by Willinger et al. [23], which was validated against Nahum’s experiments [16], is used here. The FE model of head has skin, face, skull, cerebrospinal fluid (CSF), falx, tentorium and brain. The various layers of the head are generally nonhomogeneous, anisotropic, non-linear and viscoelastic. However, for modelling purposes here, they are assumed as homogeneous, isotropic and linearly elastic, except for the brain, which is assumed as viscoelastic in nature. The properties of the various head parts are listed in Table 2. The shear characteristics of viscoelastic behaviour of the brain are expressed by (1)

Here G∞ is the long-term shear modulus, G0 is the shortterm shear modulus and β is the decay factor. The values of G0 , G∞ and β suggested by Willinger et al. [23] are used and listed in Table 3. Skull was modelled by a three-layer sandwich material, which represents the outer and inner table along with a soft dipole layer. Shell elements are used to model the face, skull, falx and tentorium; solid elements are used to model the skin, brain and CSF. Mass of the FE Table 4. Material properties of outer shell and strap in helmet.

Part Outer shell (acrylo-butadiene styrene) Strap (nylon)

145

model of head is 4.5 kg, and there are 11,939 nodes and 13,193 elements in the FE mesh.

boundary conditions, interface conditions and material properties. The output was in the form of stresses for the deformable head. To evaluate the HIC, however, a rigid head model was used.

G(t) = G∞ + (G0 − G∞ ) e−βt

Density (kg m−3 )

Density (kg m−3 )

Elastic modulus (N m−2 )

Yield stress (N m−2 )

Poisson’s ratio

Thickness (mm)

1200 1100

2.0 × 109 3.0 × 109

34.3 × 106 –

0.37 0.42

3 1

272

P.K. Pinnoji et al.

back to the original stress space. The yield surface and its evolution are defined by the Equations (2) and (3), respectively, given below. ft = |σi | − Y = 0

(2)

Y = Y +H (ev ) 0

Yt = Yt0

(3)

Here Y is the yield stress, Y ◦ is initial compressive yield stress, Yt is tensile cut off stress, σ i is the principal stresses and H is strain hardening, which is a function of the volumetric strain, ev , defined by natural logarithm of relative volume. An associative flow rule for a flow surface, which is same as yield surface, is assumed and the plastic strains are derived from Equation (4) ·

εij·p = λ

∂F ∂ σij

(4) ·

εij·p is the plastic flow rate tensor, and λ is the plasticity consistency parameter. In LS-DYNA, the stress versus volumetric strain data for the liner foam is given in tabular form and it fits the material model 63 to this curve. The EPS foam has Young’s modulus of 1.8 × 107 N m−2 with zero Poisson’s ratio and a compressive yield stress (Y ◦ ) of 0.6 × 106 N m−2 . 4.3.

Finite element mesh and injury criteria

To carry out the complete impact analysis of helmet–head FE simulations were performed for front, side and oblique impacts against a flat anvil. Figure 10 shows the FE model of helmet–head used in front, side and oblique impacts. Fournoded Belytschko-Tsay shell elements were used to model the outer shell of motorcycle helmet with 3 mm thickness. Belytschko-Tsay shell element has five integration points through the thickness and is computationally efficient. The EPS liner foam and the Nylon strap were modelled with eight-noded brick elements. The model had 2130 elements for outer shell, 7360 elements for the liner foam and 858 elements for the strap. The mass of the helmet was 0.8 kg. Surface-to-surface contact based on penalty formulation with low coefficient of friction was modelled between the

head–helmet and between the helmet–flat anvil to prevent interpenetration of these surfaces. Numerical simulations were performed with initial velocity of the helmet–head system varying from 7 to 10 m s−1 . Head injury criterion characterises the injury of the head under the impact by not only involving the peak acceleration but also the distribution and duration of the acceleration over the time of impact [17]. For finding HIC, the head is considered as rigid. For numerical simulations in LSDYNA, the geometry of the head is the same as used in the deformable head model. The HIC is calculated as   HIC = 

1 (t2 − t1 )

t2

2.5  a(t)dt 

∗ (t2 − t1 )

(5)

t1

where a is the resultant acceleration at the centre of gravity of the rigid head in units of acceleration of gravity (g = 9.81 m s−2 ). t1 and t2 are the time points in seconds during the crash for which HIC is maximum. For a deformable head to predict the damage in brain during head impact, von Mises stress has been used by Kang et al. [13] and Shuaeib et al. [22]. Lower the von Mises stress in the brain under helmet impact, the better helmet it is. Intracranial pressure, which is the pressure exerted by skull on the brain tissue and CSF, is used by Ruan et al. [21], Zong et al. [25], Klieven [14] and Kang et al. [13] for validation of human head models. One of the most damaging aspects of the increase in intracranial pressure is brain trauma. Normally, the intracranial pressures are increased due to the brain swelling and the blockage of CSF outflow in the brain ventricles at the base of brain. In the head impact, skull may undergo deformation, and change in volume of the skull is compensated by change in volume of the CSF and brain. The skull deformation compresses the CSF, which accumulates in the brain and causes an increase in the intracranial pressure. 5. Results and discussion Helmet standards use flat anvils as impact surface for determining the performance of helmet. For Indian road conditions, the IS 4151standard [12] prescribes four impact points for impact tests with one at front, one on the rear and

Figure 10. Finite element model of helmet–head in front, side and oblique impacts.


273

Figure 11. Force on the head with and without helmet (front impact).

Figure 12. Intracranial pressures at coup and contra-coup with and without helmet (front impact).

two on the either sides of the helmet and is compatible with ECE R22:03 and SNELL motorcycle helmet standards. It recommends that the peak acceleration of the head should not exceed 300 g for an impact velocity of 7 m s−1 . First, the numerical results were compared for front impact with a bare head and with the helmet over it.

pressures at coup and contra-coup sites when head impacts the flat rigid anvil without and with helmet. Without helmet the intracranial pressures are 1.4 and 0.78 MPa at coup and contra-coup, respectively. With helmet on head, the intracranial pressures are reduced to 0.21 MPa at coup and 0.12 MPa at contra-coup site and closely match with the values given by Deck et al. [5]. When the helmet–head impacted the flat rigid surface, initially the polystyrene foam deformed from outside as the outer shell came in contact with the rigid surface and came to rest whereas liner foam continued to deform. Later, at 1.5 ms the head crushes the polystyrene foam permanently from inside and the compression is highly local. At 10 m s−1 impact velocity, the liner foam in the helmet without ventilation has almost bottomed-out and the stresses in the foam are high. Figure 13 shows the deformation in motorcycle helmet under head impact at 10 m s−1 velocity. Force on the head at different impact velocities during the frontal impact is depicted in Figure 14, and it can be observed that the sharpness of the force versus time curve rises with impact velocity. It means that the peak force on the head acts

5.1.

Front impact

Figure 11 shows the force versus time curve at 7 m s−1 impact velocity for head impact with and without a helmet. In head impact without a helmet, the contact duration between the head and rigid surface was approximately 2 ms. The contact force increased sharply to 47,000 N at 1.8 ms and dropped thereafter. For the head with a helmet, the maximum force on head was found to be 7230 N at 6 ms, and the contact duration between the head and helmet was approximately 7 ms. This force on the head matches closely with the results of Deck et al. [5], who predicted a force of 8000 N on head for front impact of helmet–head at 7.5 m s−1 velocity. Figure 12 shows the intracranial

Figure 13. Deformation in helmet under head impact at 10 m s−1 velocity (front impact).

274

P.K. Pinnoji et al.

Figure 14. Force on the head without helmet ventilation at different velocities (front impact).

Figure 15. Force on the head with ventilated helmet at 7 m s−1 velocity (front impact).

for longer duration at low-velocity impacts and for smaller duration at high-velocity impacts. The force on the head creates positive pressures at the coup site in brain because of compression, and negative pressures are developed at the contra-coup site because of tension. With all impact velocities, the maximum force on head was reached between 5 and 6 ms time after the first contact between the helmet and rigid surface. In front impact at 10 m s−1 velocity, the helmet without ventilation predicted a maximum force of 11,496 N on the head and is higher than the predicted fracture force of skull 9900 N by Willinger et al. [23]. So, this helmet would not protect the human head at 10 m s−1 impact velocity. Figure 15 shows the force on the head with different ventilation models in the helmet when it impacts a flat rigid surface at 7 m s−1 velocity. It is observed that the trend in the dynamic force with time is qualitatively same for all the ventilated helmet models but the peak value differs. Forces on the head were lower with some ventilated helmet models compared with the non-ventilated helmet models.

In the region with grooves, the liner foam behaves locally as low-density foam. Moreover, during the deformation under the impact, the foam was deformed towards these grooves and slot as there was some space to displace. These probably account for reduction of forces on the head with ventilated helmets. The minimum peak force on the head was 6574 N for the helmet with a 14 mm × 7 mm groove and 48 mm × 7 mm slot, and the maximum peak force was 7364 N for the helmet with 42 mm × 7 mm groove. All the helmet ventilation models except 42 mm × 7 mm groove gave lower peak force than the helmet without ventilation at 7 m s−1 impact velocity. The biomechanical parameters such as force, intracranial pressure, von Mises stress, HIC and resultant acceleration of head with all ventilated helmet models for 7 m s−1 velocity in front impact and the maximum value for each case are listed in Table 5. The first five parameters are calculated by considering the head as deformable, whereas for last two, the head is considered as rigid. Higher forces were predicted for a helmet model with a groove of size

Table 5. Biomechanical parameters with different helmet ventilation models in front impact at 7 m s−1 initial velocity. Intracranial pressure (N m−2 ) Helmet type No ventilation 14 mm× 7 mm groove 28 mm × 7 mm groove 42 mm × 7 mm groove 14 mm × 14 mm groove 14 mm × 7 mm – 3 grooves

Force on the helmet (N)

Force on the head (N)

Coup

7441 6846 7363 7739 7057 7137

7230 6574 7058 7364 6862 6850

2.1 × 105 1.87 × 105 2.1 × 105 2.36 × 105 1.94 × 105 1.95 × 105

Contra-coup

Von Mises stress in the brain (kPa)

Head injury criterion

Peak acceleration (g)

−1.19 × 105 −1.1 × 105 −1.2 × 105 −1.28 × 105 −1.15 × 105 −1.14 × 105

47.4 45.7 50.9 55.6 46.1 46.9

867 774 868 1051 744 691

170 158 168 183 155 160


275

Figure 16. Von Mises stress in brain with helmet of 14 mm × 14 mm groove and slot (front impact).

42 mm × 7 mm as compared with other ventilation models. At 7 m s−1 impact velocity, the variation of intracranial pressures at coup and contra-coup sites for various helmet models with ventilation was very similar to the one for helmet without grooves and slots (Figure 12), and the peak values are almost the same with all the ventilation models of helmet. Figure 16 shows the von Mises stress distribution in the brain when helmet with 14 mm × 14 mm groove and slot impacts the rigid surface at 7 m s−1 velocity. In front impact, the maximum von Mises stress occurred in brain stem in all models of helmet whether with or without ventilation. The maximum von Mises stresses are in the range of

45–55 kPa and are similar to the values observed by Deck et al. [5] but are higher than the 20-kPa limit proposed for the neurological injuries. The HIC and peak acceleration are lowest for the helmet with 14 mm × 14 mm groove although force on the head and von Mises stresses are lowest for the helmet with 14 mm × 7 mm groove. The difference in the values of these quantities between the two helmets is quite small. The resultant acceleration of head versus time for various ventilation models of helmets is shown in Figure 17. The magnitude of acceleration (170 g) for the non-ventilated helmet is lower than that in Brands et al. [3] but closely matches

Figure 17. Resultant acceleration of the head with and without helmet ventilation (front impact).

Figure 18. Force on the head without helmet ventilation at different velocities (side impact).

276

P.K. Pinnoji et al.

Figure 19. Force on the head with ventilated helmet at 7 m s−1 velocity (side impact).

with that in Deck et al. [5]. In all the helmet models with grooves and slot, the liner foam was crushed totally for an initial impact velocity of 9 m s−1 , and the forces on the head increased, whereas in the helmet with no grooves and slot, the liner foam was crushed at an impact velocity of 10 m s−1 . 5.2.

Side impact

The forces on the head, HIC and resultant acceleration of the head are higher in side impact than in front impact for same impact velocity, but the intracranial pressures and von Mises stresses are almost same. Figure 18 shows the force on the head under helmet impact on side at different velocities. The trend of force versus time is almost similar to the front impact although the magnitude is almost 33% higher. The gap between head and liner foam is almost 10 mm on side compared with 4 mm on front. So, the head took some time to come into contact with liner foam, and it can be observed from the figure that the forces rise after 3 ms. In side impact, the liner foam was totally crushed at 10 m s−1 velocity.

Figure 20. Force on the head with various ventilation models (oblique impact).

Figure 19 shows the force on the head with various ventilation helmet models at 7 m s−1 impact velocity. The variation in forces, intracranial pressures, von Mises stresses, HIC and acceleration of the head between different ventilation models is much smaller in side impact as compared with the front impact. The biomechanical parameters for side impact with various helmet ventilation models are given in Table 6. 5.3. Oblique impact The study of biomechanics of helmet–head in oblique impact is also needed as it involves rotation and results in large shear strains in the brain, which can cause traumatic brain injuries like diffuse axonal injuries. Rotation occurs in helmet–head impact because the impact point is not situated straight under the centre of gravity of the helmet–head. For the oblique impact test, the mean angle between the helmet direction and the horizontal axis considered in this study was 45◦ . In FE analysis, the helmet–head impact was simulated at 7 m s−1 impact velocity.

Table 6. Biomechanical parameters with various helmet ventilation models in side impact at 7 m s−1 initial velocity. Intracranial pressure (N m−2 ) Helmet type No ventilation 14 mm × 7 mm groove 28 mm × 7 mm groove 42 mm × 7 mm groove 14 mm × 14 mm groove 14 mm × 7 mm – 3 grooves

Force on the helmet (N)

Force on the head (N)

Coup

10,202 10,132 10,207 10,074 10,187 10,126

9513 9504 9552 9384 9531 9512

2.2 × 105 2.2 × 105 2.2 × 105 2.2 × 105 2.2 × 105 2.19 × 105

Contra-coup

Von Mises stress in the brain (kPa)

Head injury criterion

Peak acceleration (g)

−1.43 × 105 −1.43 × 105 −1.43 × 105 −1.43 × 105 −1.43 × 105 −1.43 × 105

63.5 62.3 62.1 63.4 62.2 63.2

1680 1693 1701 1690 1730 1682

229 231 235 234 223 230


277

Table 7. Biomechanical parameters with various ventilation models in oblique impact at 7 m s−1 initial velocity. Intracranial pressure (N m−2 ) Helmet type No ventilation 14 × 7 groove 28 × 7 groove 42 × 7 groove 14 × 14 groove 14 × 7 – 3 grooves

Force on the Force on the helmet (N) head (N) 7248 7663 7596 7483 7467 8043

6339 6657 6474 6506 6502 6691

Coup

Contra-coup

1.76 × 105 1.81 × 105 1.82 × 105 1.78 × 105 1.85 × 105 2.17 × 105

−1.57 × 105 −1.56 × 105 −1.56 × 105 −1.55 × 105 −1.57 × 105 −1.43 × 105

Figure 20 shows the force on the head with various ventilation models of helmet. It is observed that there is not much variation in force on the head between various ventilation helmet models and non-ventilated helmet models. During oblique impact the helmet seems to undergo both compression and bending, and the forces on the helmet and head are not as high as in front and side impacts. The compression in oblique impact is highly localised than in front and side impacts. This could be due to the higher curvature of helmet and head at 45◦ angle. The maximum values for various biomechanical parameters in helmet–head oblique impact are listed in Table 7. As in front impact the intracranial pressures are almost the same in oblique impact with all the helmet models, and a maximum compressive value of 0.21 MPa in the region of the impact site and a maximum tension value of 0.15 MPa are predicted at the opposite point. The HIC and the resultant acceleration of the head are higher compared with front impact for helmet without ventilation. The maximum principal strain (Almansi) occurred in the central region of the brain and did not vary much with the type of ventilation in helmet. The maximum strain, 0.36, in the brain tissue and peak acceleration of the head, 198 g, at 7 m s−1 impact velocity for helmet without ventilation closely matches with Aare et al. [1]. In oblique impact, the liner foam in the helmet model with three grooves of 14 mm × 7 mm and a slot crushed totally near the impact site at 7 m s−1 velocity, and the forces were found to be 8043 N on the helmet and 6691 N on the head. The HIC was not calculated for this helmet model as the liner foam crushed totally after 6.5 ms, and the simulation was stopped. The EPS foam in all the ventilation models of this helmet was crushed totally at 9 m s−1 , and the forces on the head increased to approximately 12,000 N, which is higher than the skull fracture force.

6.

Conclusions

The air flow inside a motorcycle helmet and impact dynamics of the helmet are studied. The CFD is used to determine air velocities inside the gap between the head and helmet. Air velocities inside the conventional helmet are very low.

Von Mises Peak in stress Max. strain Head injury acceleration the brain (kPa) in brain criterion (g) 33.9 33.5 33.3 33.6 33.5 35.4

0.361 0.367 0.366 0.366 0.366 0.363

1524 1195 1264 1314 1550 –

198 184 177 201 199 –

These velocities can be improved by providing a groove in the central plane of the liner and a slot in front. All the ventilated helmet models gave improved higher air velocities than the helmet without ventilation. This improvement in air velocities is observed not only in the central plane of the groove but also in other regions of the helmet–head gap. For a helmet model with three parallel grooves and slot (of fixed size used here), results show that air velocities vary at each groove entrance and did not show increase in regions other than the central plane. The slot orientation affects air velocities, and the slot tangential to the head gives higher values of air velocity in the helmet–head gap. In the dynamics study, the various head biomechanical parameters are estimated for helmet with ventilation through the FE analysis. The human head model is considered as deformable for finding the forces, intracranial pressures, stresses and strains and as rigid for finding HIC and peak acceleration. The analysis is carried out for front, side and oblique impacts with all the ventilated helmet models along with head. The analysis predicts higher stresses in the head than the tolerance limits given in the standards although they match well with values reported by other researchers. In front impact at 7 m s−1 velocity, the von Mises stresses for helmets with 14 mm × 7 mm and 14 mm × 14 mm groove dimensions are lower as compared with the conventional helmet without ventilation. In side and oblique impacts, the von Mises stresses are almost same for all the helmet models. All the models of helmet with ventilation passed the standards in terms of HIC and peak acceleration at 7 m s−1 impact velocity but failed at higher velocities. During front impact, the liner foam in helmets with ventilation bottomed out at approximately 9 m s−1 as compared to the 10 m s−1 observed in the conventional helmet. A similar phenomenon was also observed in side and oblique impacts. Though the air velocities are higher in the gap between head and helmet with 42 mm × 7 mm groove and forces on the head are lower in side impact, the stresses in front impact exceed those of stresses in the conventional helmet and helmet with 14 mm × 14 mm groove. For 42 mm × 14 mm groove, the values in front impact are still higher and, therefore, this size of groove is not recommended. It is seen that

278

P.K. Pinnoji et al.

higher groove size gives higher air velocities inside the helmet but at the same time results in higher stresses in the brain if the groove width and depth exceed a certain size. The compromise design is to minimise the injury or harm to the helmet wearers while providing ventilation. Among the various helmet ventilation models, the one with a groove of 14 mm × 14 mm along with a slot of 48 mm × 7 mm is preferable at 7 m s−1 impact velocity as it enhances the average air velocities in the helmet–head gap, and the head experiences lower forces and stresses. As no literature is available on the optimisation of ventilation in motorcycle helmets, it is recommended that further studies should be carried out to improve the ventilation as well as dynamics performance of ventilated helmet impact. Acknowledgment The authors express gratitude to the Transportation Research and Injury Prevention Programme, IIT-Delhi, India, for assistance and financial support provided through a sponsored research project.

References [1] M. Aare, S. Kleiven, and P. Halldin, Injury tolerances for oblique impact helmet testing, Int. J. Crashworthiness 9 (2004), pp. 15–23. [2] J.D.A. Abeysekera and H. Shanavaz, Ergonomic evaluation of modified industrial helmets for use in tropical environments, Ergonomics 31 (1988), pp. 1317– 1329. [3] D.W.A. Brands, J.G.M. Thunnissen, and J.S.H.M. Wismans, Modelling head injury countermeasures: A 3D helmet model, AGARD Specialists’ Meeting on Impact Head Injury: Responses, Mechanisms, Tolerance, Treatment and Countermeasures, New Mexico, 7–9 November 1996. [4] P.A. Bruhwiler, M. Buyan, R. Huber, C.P. Bogerd, J. Sznitman, S.F. Graf, and T. Rosgen, Heat transfer variations of bicycle helmets, J. Sports Sci. 24 (2006), pp. 999–1011. [5] C. Deck, D. Baumgartner, and R. Willinger, Helmet optimization based on head-helmet impact modeling, 8th International Conference on Computer Aided Optimum Design of Structures, OPTI VIII, MNP, Michigan, USA, 2003, pp. 319–328. [6] FLUENT 6.2 User’s Guide, FLUENT Inc., 2005. [7] J.H. Ferziger and H. Peric, Computational Methods for Fluid Dynamics, 3rd ed., Springer, Berlin, 2002. [8] A. Gilchrist and N.J. Mills, Modelling of the impact performance of motorcycle helmets, Int. J. Impact Eng. 15 (1994), pp. 201–218. [9] P. Halldin, A. Gilchrist, and N.J. Mills, A new oblique impact test for motorcycle helmets, IRCOBI Conference, Isle of Man, UK, 2001, pp. 343–345.

[10] J.O. Hallquist, LS-DYNA Theoretical Manual, Livermore Software Technology Corporation, 1998. [11] T.J. Horgan and M.D. Gilchrist, Influence of FE model variability in predicting brain motion and intracranial pressure changes in head impact simulations, Int. J. Crashworthiness 9 (2004), pp. 401–418. [12] IS 4151, Indian standard specification for motorcycle helmet users, Bureau of Indian Standards, New Delhi, 1993. [13] H. Kang, R. Willinger, B.M. Diaw, and B. Chinn, Validation of a 3D anatomic human head model and replication of head impact in motorcycle accident by finite element modeling, SAE Transactions, Paper No. 973339, 1997, pp. 329– 338. [14] S. Kleiven, Evaluation of head injury criteria using an FE model validated against experiments on localized brain motion, intra-cerebral acceleration, and intra-cranial pressure, Int. J. Crashworthiness 11 (2006), pp. 65–79. [15] S. Kleiven and W.N. Hardy, Correlation of an FE Model of human head with experiments on localized motion of brain – consequences for injury prediction, Society for Automotive Engineers, 46th Stapp Car Crash Journal, 2002, SAE Paper No. 02S-76. [16] A.M. Nahum, R. Smith, and C.C. Ward, Intracranial pressure dynamics during head impact, Proceedings of 21st Stapp Car Crash Conference, New Orleans, USA, 1977, pp. 339–366. [17] J.A. Newman, Head injury criteria in automotive crash testing, Society of Automotive Engineers, Warrendale, PA, 1970, SAE 700426. [18] D. Otte, B. Chinn, D. Doyle, S. Makitupa, K. Sturrock, and E. Schuller, Contribution to Final Report of COST 327 Project, University of Hanover, Germany, 1999. [19] P.K. Pinnoji and P. Mahajan, Impact analysis of helmets for improved ventilation with deformable head model, IRCOBI Conference, Madrid, 2006, pp. 159–170. [20] U. Reischl, Fire fighter helmet ventilation analysis, Am. Ind. Hyg. Assoc. J. 47 (1986), pp. 546–551. [21] J.S. Ruan, T.B. Khalil , and A.I. King, Dynamic response of the human head impact by three-dimensional finite element analysis, J. Biomech. Eng. Trans. ASME 116 (1994), pp. 44–50. [22] F.M. Shuaeib, A.M.S. Hamouda, R.S. Radin Umar, M.M. Hamdan, and M.S.J. Hashmi, Motorcycle helmet Part 1: Biomechanics and computational issues, J. Mater. Process. Technol. 123 (2002), pp. 406–421. [23] R. Willinger, B.M. Diaw, and H.-S. Kang, Finite element modeling of skull fractures caused by direct impact, Int. J. Crashworthiness 5 (2000), pp. 249–258. [24] A.L. Yettram, N.P.M. Godfrey, and B.P. Chinn, Materials for motorcycle crash helmets – a finite element parametric study, Plastics Rubber Compos. Process. Appl. 22 (1994), pp. 215–221. [25] Z. Zong, H. Lee, and C. Lu, A three-dimensional human head finite element model and power flow in a human head subjected to impact loading, J. Biomech. 39 (2006), pp. 284–292.