18.1.1. Definitions of TBI and Blast-Induced TBI

TBI is a silent epidemic with serious consequences that changes the lives of the victims and their families irrevocably (Thurman et al., 1995). The Centers for Disease Control and Prevention defines TBI as a head injury that is associated with decreased levels of consciousness, amnesia, other neurological or neurophysiological abnormalities, skull fracture, diagnosed intracranial lesions, or death. TBIs can be classified as mild, moderate, or severe.

The American Congress of Rehabilitation Medicine has defined blast-induced mTBI as a head injury resulting in at least one of loss of consciousness (LOC) for 30 minutes or less; posttraumatic amnesia (PTA) for less than 24 hours; any alteration in the mental state immediately after the accident; or focal neurological deficit(s) that may or may not be transient. The emergency room physicians upon presentation of the patient routinely use the Glasgow Coma Scale to assess the state of TBI. The 15-point Glasgow Coma Scale defines the severity of injury as mild (13–15), moderate (9–12), severe (3–8), or vegetative state (<3) based on eye, verbal, and motor responses (Teasdale and Jennett, 1974). A new classification specific to blast-related TBI (bTBI) has been also proposed recently, in which a blast-induced mTBI is characterized by loss of consciousness LOC for less than 1 hour and PTA for less than 24 hours after exposure to an explosive blast. Moderate bTBI is characterized by loss of consciousness LOC for 1–24 hours and PTA for 1–7 days, and severe bTBI is characterized by loss of consciousness for more than 24 hours and PTA for more than 7 days (Ling et al., 2009). Mild to moderate cases of bTBI do not involve skull fracture; further, they are not detectable through current neuroimaging techniques. The Centers for Disease Control and Prevention has reported that up to 75% of TBIs that occur each year are mild, and a World Health Organization task force reported that 70%–90% of all treated TBI were mild (Kraus et al., 2007). mTBI is typically not associated with abnormalities in brain imaging (Okie, 2005), and most patients with mTBI recover fully in 4–12 weeks (Alexander, 1995; Kennedy et al., 2007). However, mTBI patients with more severe injuries, such as those who experienced loss of consciousness lasting more than 10 minutes or PTA lasting more than 4–6 hours, may require months to years to recuperate (Alexander, 1995). In addition, some mTBI patients develop postconcussive syndrome, experiencing persistent cognitive, behavioral, and/or somatic symptoms (Alexander, 1995; Heitger et al., 2009; Zhu et al., 2010). Studies have shown that 15%–35% of patients with mTBI experience onset of long-term disability (Alexander, 1995; Heitger et al., 2009; Thurman et al., 1999).

18.1.2. Classification of TBI: Ballistics, Blunt, and Blast

Head and brain injuries are categorized based on the origin of mechanical forces as blast, blunt, or ballistic injuries.

18.1.2.3. Blast

A field blast is extremely complex and the facets of the injuries include components of both blunt and ballistic impact. In general, blast injuries are classified into four main categories: (1) primary (direct effects of overpressure), (2) secondary (effects of projectiles/shrapnel), (3) tertiary (effects from fall from blast winds), and (4) quaternary (burns, asphyxia, and exposure to toxic inhalants). Usually, depending on the strength of explosive, casing, standoff distance, and scenario (urban, theater, indoor/outdoor), victims suffer a combination of these injuries. An illustration of these injuries is shown in Figure 18.2.

FIGURE 18.2

bTBI classification. In this figure, w is the charge weight and r is the standoff distance.

18.1.2.3.1. Primary Blast Injury

Once a bomb explodes, it results in a sudden rise in the atmospheric pressure, which represents the shock front followed by an exponential decay resulting from the expansion wave depleting the overpressure. This sudden rise followed by an exponential decay comprises the positive phase of the blast wave and is responsible for primary blast injury. Although the sharp rise constitutes the shock wave, the rest of the pressure pulse is referred to as blast wind. Furthermore, depending on surrounding (urban, indoor/outdoor) structure or an enclosure (such as interior of a bunker or a vehicle), the subjects are also exposed to complex blast waves caused by multiple reflections from the walls, floor, and ceiling and their interactions.

18.1.2.3.2. Secondary Blast Injury

IEDs have metal casings and are usually filled with metal fragments. During the explosion, these metal fragments become highly accelerated projectiles that penetrate the body. The pathology associated with secondary blast injuries is similar to that of the ballistic injury discussed previously.

18.1.2.3.3. Tertiary Blast Injury

Tertiary injuries include those with nonpenetrating projectile impacts as well as injuries from a fall. Explosion within a building may result in its collapse, causing impact and crushing of the body, whereas in an outdoor blast, depending on the proximity of the blast epicenter, the whole body might be thrown, resulting in impact with hard surface. When seated inside a vehicle under explosive attacks or in a rollover, the victims are thrown violently inside, leading to an impact-type injury. The pathology of this injury is similar to the blunt impact discussed previously.

18.1.2.3.4. Quaternary Blast Injury

Quaternary blast injuries refer to explosion-related injuries that were not included in the primary, secondary, or tertiary injuries (DePalma et al., 2005). Quaternary injuries include burns (chemical or thermal), toxic inhalation, and exposure to radiation.

18.1.3. Mild/Moderate Blast TBI

When subjects are closer to the epicenter of the blast, they will endure injuries with a degree of severity that in most cases will result in fatalities. However, at a sufficiently longer distance from the source (i.e., in the far field range) only primary or pure blast effects are dominant (Reynolds, 1998). It is conjectured that mild to moderate TBI seen in the theater is due to primary blast effects alone because injuries were nonfatal, with neither head fracture nor open wounds (Dewey, 1971; Kleinschmit, 2011; Skotak et al., 2013; Warden, 2006). In 2005, the U.S. military reported 10,953 IED attacks, at an average of 30 per day. As mentioned previously, these explosions were less fatal with a much higher survivability rate than in previous wars, hence mostly resulting in mild or moderate TBI. Hence, in this chapter, we are only concerned with the primary blast injury wherein the victim is exposed to a pure shock-blast wave that eventually results in a mild or moderate TBI.

18.1.4. Current State of the Research

Creating conditions of blast loading and understanding the physics of the blast are fundamental elements in the study of acute and chronic ailments of primary blast injury. Field-testing and shock tubes are the main methods used for creating a primary blast injury. Among them, shock tubes are more commonly and frequently used in primary blast injury research. Table 1 from a review article by Kobeissy and his colleagues shows the recent major studies on primary blast injury (Kobeissy et al., 2013). Among this list, only 8 cases of 49 used field explosive testing as a method of generating primary blast. In 36 other cases, shock tubes were used. Shock tubes are also classified into three types: (1) compressed gas shock tube, (2) detonation shock tube, and (3) combustion shock tube. Among them, compressed gas shock tubes are predominately used as the source for primary blast (in 33 cases).

Understanding the processes of primary blast inflicting loads (i.e., the mechanism by which the injury occurs) is a prerequisite to the design of proper animal model testing and interpretation of results (Benzinger et al., 2009). Over the past few years, several mechanisms of mechanical insult have been suggested. These mechanisms are (1) a thoracic mechanism in which blast waves enter the brain through the thorax and increase brain pressure via vasculature (Cernak et al., 1997, 2001; Courtney and Courtney, 2009), (2) translational and rotational head acceleration (Courtney and Courtney, 2009), (3) blast wave transmission through cranium (Moore et al., 2009; Nyein et al., 2010; Taylor and Ford, 2009), (4) skull flexure (Bolander et al., 2011; Moss et al., 2009), and (5) cavitation (Panzer et al., 2012). Most of these mechanisms are proposed using numerical models alone and experimental evidence is needed to corroborate these proposed mechanisms. Postmortem human specimen (PMHS) heads, which is closest to humans in terms of anthropometry, can be used to study and validate these mechanisms. However, one main drawback with PMHS is that it is not possible to mimic the material and biological aspects of a living brain.

Animal models are essential for studying the biomechanical, cellular, and molecular aspects of human TBI that cannot be addressed using surrogate models. Furthermore, a validated animal model is required for development and characterization of novel therapeutic interventions. Currently, the majority of this work is done using small mammals; very few studies use large mammals (Kobeissy et al., 2013). Animal models are used to investigate the physiological, neuropathological, and neurobehavioral consequences as well as identify biomarkers that are related to brain injury. However, one main issue with the current research in animal models is the inconsistency in specimen location when doing experiments with a shock tube. This leads to two problems: (1) erroneous loading conditions (in some cases) and (2) comparison of the results between different laboratories is virtually impossible. Therefore, developing a standard experimental model that can be validated across laboratories is essential for the interpretation of the mechanisms of blast injury, the identification of biomarkers, and, eventually, the development of strategies for mitigating blast-induced brain injury (Xiong et al., 2013).

Given the difficulties in understanding TBI associated with conducting experiments on PMHS or human volunteers and translating animal models results to human, computational modeling has always been an easier choice. As a result, there have been a large number of models reported in the literature in the past (refer to Table 2.2 in Ganpule, 2013). Computational modeling is currently used for studying the shock-blast interaction with the head (analysis of reflected pressure fields), for flow field analysis to determine flow separations, and for pressure field distribution in the brain and loading mechanisms (skull flexure, direct pressure transmission) as well as to study the effect of helmets. One main drawback in computational modeling is the lack of consistent and accurate material models. Consequently, it is vital to validate the model against an accepted experimental standard to ensure the validity of the predictions in the model.

It is estimated that since 2005, 77% of soldiers who sustained any type of TBI were wearing their helmets at the time of injury (Wojcik et al., 2010). Although the role of helmets in bTBI is critical, there are very few investigations that have studied helmets in blast mitigation (Brown et al., 1993; Grujicic et al., 2011; Nyein et al., 2010; Sogbesan, 2011). These studies concluded that a padded helmet provides only some degree of protection to the head during blast loading. Reductions of pressure were noticed in coup regions, whereas in other areas (e.g., brainstem) the values of ICP, shear stress, and strain remained unchanged compared with the no-helmet case. A comprehensive study using both experiments and numerical models is essential to elucidate the role of helmets in blast mitigation and to gauge the capabilities of the different helmets currently used by military personnel in the theater.

18.1.5. Organization of the Chapter

In this section, we examine the physics of explosion and the characteristics of shock-blast waves as a function of explosive strength and stand-off distance. Because experiments on biological medium require a large number of samples with variations in key parameters, conducting experiments with actual blasts is not an option in terms of time, cost, or safety. Properly designed laboratory-scale shock tubes can replicate field conditions. Unfortunately, many of the current experiments conducted by different experimental groups around the world have not paid attention to this detail; hence, their results should be viewed with caution. If pure primary blast conditions are not achieved during the test conditions, then the specimens are subjected to a combination of blast as well as the loads associated with tertiary injury. Compressed gas-driven shock tubes are very commonly used to replicate field conditions. In this section, we outline the basic components of the shock tube and how each of them interacts to produce the desired shock wave profile. The location of the test specimen within the length of the tube plays a major role.

In Section 18.2 on PMHS blast testing, we examine the basic question of how a blast wave interacts with the human head and causes mechanical loading on the brain. Both the physics of the problem and preliminary experiments clearly show that the brain loading would depend on the geometry (shape and size) of the head as well as the tissue properties of the head, in addition to the intensity of shock wave itself. To represent humans as accurately as possible, cadaveric heads (PMHS) back-filled with ballistic gels were subjected to field-relevant blast conditions. Deformation was measured both on the surface of the head and by measuring ICP variations. These measurements were repeated when the intensity of the external shock was increased. The results showed that there is an increase in pressure excursion in the brain with an increase in the shock wave intensity.

In Section 18.3 on animal model blast testing, we examine the role of animal models in answering the fundamental question of whether the primary blast causes measurable changes at the cellular and tissue level, and also study behavioral patterns. We discuss the observed physiological changes occurring immediately after exposures in terms of heartbeat, pulse, blood oxygen levels, and weight loss over time. We established a dose-response curve between external blast overpressure from 60 kPa to 450 kPa and mortality rates. The response shows a typical sigmoidal curve. Significant pathophysiological changes like plasma membrane integrity loss, blood-brain barrier damage, and elevated levels of reactive oxygen species are observed. The damage in the deeper brain regions distinguishes bTBI from the coup and contrecoup type damage noticed in blunt impacts. Small, but not statistically significant changes are observed in the behavior.

Computer models play a key role in not only understanding individual test results but results across a range of species and loading conditions. This is the subject of Section 18.4. Here we present the basic mathematical formulations involved, geometric modeling techniques used to convert images to numerical models, and material model parameter issues. In Section 18.5, we address the key issue of protection. Though we do not fully understand the blast injury biomechanics, the adequacy or otherwise of current head protection devices still need to be studied. The role of geometry, material, and external loadings on the design of current helmets are discussed in this chapter.

18.1.6. Replicating Primary Blast in the Laboratory

To study mild/moderate BINT, it is vital to replicate field blast conditions associated with the injury. First, we explore the physics of explosion briefly and establish the shock wave parameters (SWPs) that can be measured and controlled. Further, some of the current techniques used for simulating the mild/moderate blast conditions are also outlined. Attention is paid to the compressed gas-driven shock tube because most researchers tend to use this method. The dos and don’ts of the design, construction, and operation of the tubes are also elucidated.

18.1.7. Characteristics of Field Explosion

Explosives are often used by insurgents as improvised weapons. They are usually present in the form of homemade chemical explosives embedded with a variety of shrapnel and typically called IEDs. They are buried under the ground and detonated from a distance or left on the side of the road to injure a vehicle, occupants of the vehicle, pedestrians, or a gathering. The strength of the explosives is measured by kilogram of TNT-equivalent. When a chemical explosive detonates, enormous energy is released in a very short time. This energy expands very quickly and compresses the surrounding air. For a simplified spherical explosive suspended in air, the expansion is spherical in nature; the compression propagates in the form of a fast traveling wave. When the velocity of the wave front exceeds the sonic velocity of air (corresponding to the temperature, altitude, and humidity) then the wave front is termed a shock wave. The shock wave is a very thin layer and the pressure increases from atmospheric to the peak overpressure within a few atomic distances and in a very short time, along the order of microseconds. Therefore, the shock wave is the front part of the blast wave and can be called the shock-blast wave. Though the terms shock and blast are used interchangeably, the shock is the very small portion of the entire blast wave profile.

Explosives can come from different sources (e.g., chemicals, high-pressure gas tanks, nuclear); their strength is expressed as a single unit in terms of TNT-equivalent. One gram of TNT produces a blast energy of 4610 J; 1 ton of TNT produces about 4.61 million kJ. Explosive strength of a given chemical expresses the effectiveness of a given chemical in producing a blast wave comparable to that of the same weight of TNT. For example, pentaerythritol tetranitrate is 1.5 times more powerful than TNT, whereas C4 (composition B of 60% RDX/40% TNT) is 1.34 times more potent than TNT. When an explosive detonates, the matter of the explosive is converted to the heat of explosion (appearing as heat and fireball) and the energy of expanding gas (leading to blast waves). If the explosive mixture contains metallic products (e.g., nails, metallic sheaths), then part of the expansion energy appears as kinetic energy of the fragments acting as projectiles.

The destructive power of the blast wave increases as a function of its velocity of propagation. If a is the velocity of sound in the undisturbed air medium and u is the velocity of the blast wave, then is defined as the Mach number. For example, the speed of sound at room temperature and pressure at sea level is approximately 343 m/sec (768 mph); a shock wave with M = 2 travels at 686 m/second (1536 mph). The 1-m-long blast wave passes any given point in less than 2 msec, which is such a short time scale the exposed body is unable to react because of its inertial mass.

In the case of an idealized mid-air explosion, the gas products and the fireball expand in a spherical manner and compress the surrounding air. This expanding air continues to increase in velocity at the outer edge and at a certain point, the velocity of the envelope just equals the velocity of sound (i.e., M = 1). When this happens, a shock front is formed. The shock front is extremely small, on the order of a few molecular diameters in width. Across this front, there is a sudden change of pressure of the undisturbed medium from atmospheric to a high pressure. There is a sudden change in pressure, velocity, and density across this narrow thickness. Behind the shock front, there is the blast wind in which the particles move at very high velocities. This shock front blast wind continues to expand outward. However, because the expanding gas occupies an increasing volume (spherical radius r), the shock velocity slowly reduces and eventually becomes the same as the sonic velocity of the medium, at which point the shock wave dies leaving behind a low-velocity blast wind.

The strength of the shock wave is measured in terms of overpressure, termed blast overpressure (BOP). BOP can be expressed as

where M is the Mach number.

Or, equivalently, the shock wave velocity can be found from BOP as

Thus we can see that there is a relationship between shock wave velocity and the blast overpressure. The shock wave velocity decreases as the spherical radius increases. At some distance (standoff distance), this shock degenerates to a sound wave.

18.1.8. Laboratory Reproduction of Shock-Blast Wave

In this section, we consider the characteristics of a spherically expanding blast wave formed by the detonation of an explosive (e.g., an uncased C4 of charge weight W). The effects of the fireball, ground reflections, and other artifacts are not considered because in the far field range of interest these effects are absent (Reynolds, 1998). The impulsive expansion of explosive product initiates first a shock wave propagating spherically outward in the surrounding air and then a family of infinite rarefaction waves propagating in the shock-compressed air. Across the shock front, pressure, density, particle velocity, and temperature rise significantly and rapidly in a few microseconds. As the radial propagation distance r increases, the surface area of the spherical shock front increases as r² and consequently its intensity decreases as 1/r².

The propagation of the shock front is supersonic with respect to the ambient air ahead (upstream) but subsonic with respect to the shocked air behind (downstream); the rarefaction waves remain supersonic with respect to the compressed air ahead (upstream) until the air compression is fully released. The sequential arrivals of rarefaction waves at a given r cause decrease of shock compression. Hence, in the intermediate to far range of r, more and more rarefaction waves catch up the shock front, giving rise to erosion of the shock front intensity, evolving nonlinear decay in overpressure p, and even a period of underpressure (negative overpressure) afterward. The shock front diminishes eventually at large r. Therefore, the p–t profiles of the blast waves of interest have a shock front with a peak overpressure p^* followed by a nonlinear decay during positive overpressure with a duration of t^* and then a period of underpressure. This is confirmed through open-field testing conducted by the authors at the Trauma Mechanics Research Facility at the University of Nebraska-Lincoln in conjunction with the U.S. Army. The laboratory and facility has since moved along with the authors to the New Jersey Institute of Technology and is a part of the Center for Injury Bio-mechanics, Materials and Medicine. The test results are further corroborated with conventional weapons effects (ConWep) (Ganpule et al., 2012) (Figure 18.3).

FIGURE 18.3

Pressure time profiles in a free field explosion: (a) IP profiles measured using a pencil gauge by a free field explosion of 1.81 kg of C4 at a distance of 2.8 m. The testing was conducted by the Trauma Mechanics Research Facility at the University of (more...)

Such p–t curves can be mathematically described by the so-called Friedlander waveform (Kinney and Graham, 1985) given by the following equation:

where b is the decay constant.

The Friedlander waveform is characteristic for an open-field blast. A typical Friedlander waveform profile is shown in Figure 18.4. The integration of p(t) over t^* gives the positive impulse per unit area I^*. p^* (overpressure) and t^* (positive time duration) are the two independent parameters along with a decay constant used for describing the essential characteristics of positive portions of the blast waves of interest. The ranges of p^* and t^* that are of interest for bTBI are postulated to be 50–1,000 kPa and 1–8 ms, respectively (Panzer et al., 2012); however, precise ranges are not available from the field data. Measurements conducted in our laboratory indicate that the upper range can be reduced to 400 kPa and the duration restricted to 7 ms. It should be noted that during an actual explosion, the p–t curve could be more complex because of reflections from ground, surrounding building structures (in an urban setting), and explosives that are cased and/or contain shrapnel. Furthermore, there are charge-to-charge variations possible from packing density, shape of the charge (spherical, cylindrical), the purity of explosives (pure or mixed with shrapnel), the type and material of the cover, and the type and location of the trigger. As far as the target is concerned, not only the distance but also the direction of the wave and environmental parameters (temperature, wind, dust, humidity) determine the loading as well as the shape, size, and material property of the target and its conditions (e.g., if human whether subjects are wearing a helmet). To conduct a fundamental scientific study of bTBI, it is critical to isolate these as extraneous effects and generate the blast waves with the Friedlander type wave profile.

FIGURE 18.4

Mathematical representation of planar Friedlander waveform. The equation in the figure represents instantaneous overpressure p+ at given time t, where p^* is the peak overpressure, t^* is positive phase duration, I^* is positive phase impulse, and b is the (more...)

Another important feature of the open-field blast waves in the intermediate to far range is that the sizes of wave front are much larger than that of a human body. The interactions of such a blast wave with a human body are influenced strongly by the confinement of an effectively edgeless (no flattop or plateau) wave front. This characteristic represents a planar wave and must be re-created in the laboratory testing to realistically simulate field blast loading.

A spherical wave in the near field while expanding becomes more planar at far field ranges. Therefore, a blast wave in the intermediate to far range can be well approximated with a planar Friedlander wave, as shown in Figure 18.4.

18.1.9. Types of Shock Tubes

Although the free-field blast testing closely replicates real-world blast conditions, there are some significant drawbacks: (1) free-field experiments are expensive and (2) time consuming, and (3) repeatability is difficult to achieve. Furthermore, the by-products of HE testing include potential fireball interactions and penetration from high-speed shrapnel (Bass et al., 2012). This can introduce unnecessary confounds to the experiments in which the objective is to understand the effect of primary blast injury and its subsequent biomechanical and biochemical sequela.

18.1.9.3. Compressed Gas Shock Tube

This is the most popular type of shock-blast generation device, which essentially uses a similar technique as a combustion and detonation type shock tube except here, the sudden release of compressed gas is used for generating the shock-blast wave. A detailed description of the theory and work of a shock tube will be discussed in the following sections. Currently, this technique has been adopted by many researchers who investigate bTBI because of its apparent simplicity, repeatability, and experimental control (Bolander et al., 2011; Leonardi et al., 2011; Long et al., 2009; Sundaramurthy et al., 2012; Zhu et al., 2010).

18.1.9.3.1. Theory of Compressed Gas Shock Tubes

Although individual shock tubes for blast wave simulation may have different features, the essential wave physics can be understood by analyzing the wave propagation in a generic shock tube configuration as shown in Figure 18.5. A typical (compressed gas–driven) shock tube consists of a driver section of pressurized gas and a driven section of air at atmospheric pressure with the two sections separated by a set of membranes. When the membranes burst, the driver gas expands rapidly and compresses the atmospheric air (i.e., driven gas) in front of a shocked state, which propagates forward as an air shock wave. Meanwhile, the driver gas expansion initiates a family of infinite rarefaction waves (expansion fan). These rarefaction waves first travel toward the closed end, are reflected at the closed end, and then travel toward the open end. Their sequential arrivals at a given location of driven section produce a nonlinear decay (see wave profiles a-c of Figure 18.5). The wave profile evolves with propagation distance to that of a Friedlander wave (curve c of Figure 18.5) when the fastest rarefaction wave (which is faster than the shock front) catches the shock front at x = x^*, where the shock front intensity is eroded the least by the rarefaction waves. Hence, at x = x^*, peak overpressure p^* has the maximum value with a Friedlander wave profile. The time for the nonlinear decay to reach p = 0 gives overpressure duration t^*, which has the minimum value at x = x^*. Before the initial catchup, x < x^* (curves a and b of Figure 18.5), the blast wave assumes a flattop shape because the rarefaction wave reflected from the closed end has not reached the shock front yet. The flattop duration is given by the difference in the arrival times of the shock front and the fastest rarefaction wave. In the range x^* < x < 0 where x = 0 represents the shock tube exit, more and more rarefaction waves catch up to the shock front causing decreasing p^* and increasing t^* with increasing x. The pressure-time (p-t) profile near (outside) the exit is shown by curve d of Figure 18.5; notice that the waveform is changed significantly (low p^*, low t^*, followed by jet wind).

FIGURE 18.5

Evolution of shock wave in a generic shock tube. (From Chandra, N. et al., Shock Waves. 22:403–415, 2012.)

18.1.9.4. General Features of Compressed Gas Shock Tube Construction

The design for a shock tube is shown in Figure 18.6. The four main components of the shock tube are the driver section or breech, transition, driven section or straight section, and the catch (or expansion) tank. The straight section includes the test section. The driver contains pressurized gas, which is separated from the transition by several frangible membranes. Upon membrane rupture, a shock wave expands through the transition and develops in the straight section. Subjects are placed in the test section, which is strategically placed to produce a desired shock wave profile. Finally, the shock wave exits the shock tube and enters the catch tank, which reduces the noise intensity. Each of the components is described below.

FIGURE 18.6

Photographs of 230 × 230-mm square shock tube used in this work. (From Chandra, N. et al., Shock Waves. 22:403–415, 2012; Sundaramurthy, A. et al. Journal of Neurotrauma. 29:2352, 2012.)

18.1.9.4.1. Driver Section

The driver section contains the compressed gas (e.g., air, helium, nitrogen) used in the generation of the shock-blast wave. It is a cylindrical tube with one end permanently sealed, whereas the other is sealed with frangible membranes such as Mylar. In some designs, thin sheets of metals (aluminum or steel) have been used. It is critical to strictly control the repeatability of burst characteristics by controlling the material and manufacturing processes of the membranes. Driver gas is filled into the driver section, pressurized, and allowed to burst depending on the membrane material and its thickness. Once the pressure reaches the burst pressure, the membranes rupture releasing the high-pressured gas into the transition.

18.1.9.4.2. Transition

The transition is a design element used to change the cross-section of the tube from a circular cylinder (driver section) to a square (driven sections). The transition was designed with a gradual expansion to minimize flow separation and turbulence associated (reflections from the sidewalls) with abrupt changes in a cross-sectional area. If the driver and driven sections remain the same, there is no need for the transition. However, to produce an effective shock front the driven is usually larger and longer than the driver.

18.1.9.4.3. Driven Section

The driven section or the expansion section is the square section where the fully formed shock-blast wave is generated. In our design, we used a square driver section, which has since been followed by other research groups. The use of a square cross-section as opposed to other geometric sections serves two purposes: (1) uniform expansion from the breech, which is vital to have a planar wave, and (2) to observe and record events in the test section with a high-speed camera (where having a circular section may cause image distortion) (Sundaramurthy et al., 2012). The test section is located within this part; while locating the test section, it is vital to note that the test section should be neither close to the breech nor to the exit. The reasons for this will be discussed in details in the coming sections.

18.1.9.4.4. Catch Tank

Depending on the strength of the shock-blast wave produced by the shock tube, the catch tank may or may not be included in a particular design. Its main purpose is to contain and release the large volume of expanded gas generated from a shot, minimizing blast loading of laboratory structures and reducing noise level. The use of a suddenly changing cross-sectional area was studied and found to successfully mitigate energy (Jiang et al., 1999). The inside of the catch tank is lined with sound absorbing material for reducing the noise.

18.1.10. Test Section Location

Modular design of the shock tube allows the flexibility of altering the distance of the test section from the breech. In the following section, we will discuss the important criteria that have to be considered for test section location and their effect on the shock-blast wave profile. Important criteria include (1) contact surface, (2) planarity, and (3) exit rarefaction.

18.1.10.1. Contact Surface

Upon rapid expansion of the driver gas, the temperature and pressure drop causing a density increase in the driver gas. The location of contact between the driver and driven gas is defined as the contact surface. The shock overpressure across a contact surface does not change, but the density difference causes a discontinuity in kinetic energy. This density difference can cause significant changes in reflected pressure across the contact surface. The contact surface can also be avoided by determining the fully expanded volume of the driver gas, and placing the sample beyond that location. The final volume of the expanded gas can be calculated using the isentropic expansion of an ideal gas, given by the following equation (Cengel et al., 2011):

where P₂ is the burst pressure, P₁ is the atmospheric pressure (14.5 psi), V₁ is the driver volume, V₂ is the final volume of the gas after complete expansion, and k (1.4 for nitrogen) is the adiabatic index. By knowing the breech volume, type of driver gas, and burst pressure, we can determine the final volume of the completely expanded driver gas. With this, we can theoretically approximate the minimum distance that is required to avoid any interference from the contact surface. However, this methodology does not consider effects from the mixing of driver and driven gases. Usually, the investigator has to be vigilant when analyzing the first 7–10 ms of p-t profile of the reflected pressure recorded in the test section to determine sudden changes in the reflected pressure. In such a case, either the test section should be moved downstream or the breech volume has to be reduced.

18.1.10.2. Planarity

In the mild/moderate range depending on the explosive strength and the distance from the epicenter, the shock front is planar. When simulating this type of blast in a laboratory shock tube, this feature is essential to replicate. Experiments can be done in the test section to determine the planarity by placing a bar with a linear array of sensors in the shock tube. The shock front planarity is determined by measuring the arrival time at various locations on the bar, and any deviation of arrival time will show the shock front curvature. With this idea, an experiment was conducted to determine the planarity of the shock front produced in our shock tube.

Along the bar, an array of eight sensors was placed with one sensor in the center and six at 4, 8, and 12 inches from the center in either direction. Assuming the bar is not mounted perfectly perpendicular to the shock tube, averaging of sensors equidistant from the center ensures that nonplanarity measurements caused by a skewed bar are corrected. The sensors at 8 and 12 inches were averaged in the same manner. A final sensor was located 13 inches from the center and was adjusted based on the calculated angle of the sensor bar (Figure 18.7a) (Kleinschmit, 2011).

FIGURE 18.7

(a) Sensor bar used for planarity testing in 28-inch shock tube. (b) Shock wave planarity measurements in 28-inch tube taken at 48, 98, and 136 inches from the exit of the transition for 18-psi overpressure shock waves.

After the sensor bar was mounted, shock waves were generated using a 11.75-inch nitrogen driver and 10 membranes. This produces shock waves with peak overpressures between 105 and 160 kPa, depending on the burst pressure and distance from the breech. The sensor bar was mounted 48, 98, and 136 inches from the transition exit. The planarity results shown in Figure 18.7b demonstrate nonplanarity with a shock front leading at the edges by approximately 0.15 inches at a location 48 inches from the transition exit.

The curvature corresponds to an approximate diameter of 47 feet, which can be considered planar. The results became increasingly planar to the location 136 inches from the transition exit. The general trends show increasing planarity with increase in the distance from the transition.

18.1.10.3. Exit Rarefaction

Figure 18.8 shows the untailored shock blast wave produced by a shock tube (Ritzel et al., 2011). The secondary decay and the following upstream recompression are the artifacts created by the rarefaction wave from the open end of the shock tube. Therefore, the subject within the test section, if not placed appropriately, will be subjected to a blast wave that does not simulate a field blast. This problem can be resolved by two methods: either having a reflector plate attached to the exit or placing the test section deep inside the driven section. In the case of a reflector plate, the venting gases when impinging on the plate create a compression wave, which essentially nullifies the rarefaction wave generated from venting (Coulter et al., 1983). In the second case, by having the specimen deep inside the driven section the exit rarefaction wave enters the test section only after the passage of the actual shock-blast wave.

FIGURE 18.8

Untailored shock blast wave profile. (From Ritzel, D. et al., Proceedings HFM-207 NATO Symposium on a Survey of Blast Injury across the Full Landscape of Military Science, 2011).

18.1.11. Effects of Placing Specimen outside the Shock Tube

Studies on the evolution of the shock wave at the exit or open end have attracted researchers over the years because of numerous flow phenomena occurring at the exit (Arakeri et al., 2004; Honma et al., 2003; Jiang et al., 1999, 2003; Kashimura et al., 2000; Onodera et al., 1998; Setoguchi et al., 1993). It is shown in these studies that at the exit of the shock tube, the shock wave evolves from planar to three-dimensional (3D) spherical with other effects like vortex formation, secondary shock formation, Mach disc, subsonic jet flow, shock-vortex interaction, and impulsive noise. All of these effects may or may not be present depending on the shock wave strength and geometry of the exit. Most of these studies, however, have focused on the flow dynamics aspects with no emphasis on qualitative or quantitative analysis of shock/blast-wave profiles (e.g., p-t profiles). This becomes particularly important as many researchers do primary blast-injury testing using animal models and surrogates outside the shock tube. Therefore, to understand the evolution of the shock-blast wave outside the shock tube, and whether it truly represents all aspects of the field blast, a detailed study of the blast wave exiting the shock tube was conducted using numerical simulations. Furthermore, an experimental study was conducted using a cylindrical model as well as an animal model comparing the mechanical responses from outside with the responses obtained inside the shock tube (Chandra et al., 2012; Sundaramurthy et al., 2012)

18.1.11.1. Evolution of Shock-Blast Wave outside the Shock Tube

The flow field at the exit of the shock tube is studied using numerical simulations. Figure 18.9 shows the pressure and velocity (vector) fields at the exit of the shock tube without a cylinder. As the blast wave exits the shock tube, the flow changes from planar to 3D spherical (Figure 18.9a). Rarefaction waves and vorticities at the corners mix with blast and the remaining air ejects as subsonic jet wind, which is evident from the velocity vector field of Figure 18.9b. This jet wind effect is not present deep inside the tube. Further, to clearly demonstrate this, Figure 18.10a shows the nodal velocities at various locations inside and outside the shock tube. Because fixed Eulerian mesh is used for modeling, velocity at a given mesh node corresponds to the instantaneous velocity of the material point coincident at given time ‘t’ with the considered node. High-velocity jet wind is recorded in nodal history for locations outside the shock tube. Particle velocity associated with this jet is higher than particle velocity associated with the shock (Figure 18.10b). Locations inside the shock tube that are closer to the exit also show a second peak in velocity because of rarefaction waves moving into the tube, but the magnitude of this second peak is lower than the particle velocity associated with the jet for outside locations. In addition, the magnitude of this second peak gradually reduces as we move inside the shock tube away from the exit (open end). Deep inside the shock tube (x = −3048 mm), a second peak is completely absent.

FIGURE 18.9

(a) Pressure field near the exit of the shock tube. Three-dimensional expansion of shock wave along with vortex formation is seen at the exit. (b) Velocity vector field near the exit of the shock tube. Jet wind is clearly visible in velocity vector field. (more...)

FIGURE 18.10

Nodal velocities at various locations inside and outside the shock tube. Because a fixed Eulerian mesh is used for modeling, velocity at a given mesh node corresponds to the instantaneous velocity of the material point coincident at given time ‘ (more...)

To clearly exhibit transition of the blast wave from planar to 3D spherical, Figure 18.11 shows the pressure distribution at the exit of the shock tube for a sequence of times. The black arrows indicate the (velocity) vector field. In each figure an outer surrogate contour indicates the primary shock wave and an inner green portion indicates the primary vortex loop. The primary shock wave at first appears to be a square shape with rounded corners as shown in Figure 18.11i. These corners become significantly rounded and straight parts at the shock tube walls are shortened (Figure 18.11ii–iv). This indicates that the primary shock wave is planar at the exit (open end) of the shock tube and evolves three dimensionally into a spherical one as time elapses. This process is called a shock wave diffraction and affects the flow expansion behind it (Jiang et al., 2003). Similar arguments can be used to show the 3D nature of the primary vortex loop that is evident from green color of Figure 18.11.

FIGURE 18.11

Flow fields illustrating physics of shock wave diffraction. Row 1 shows the axial view and row 2 shows the top view. Arrival of shock wave at the exit is marked as t = 0. (From Chandra, N. et al., Shock Waves. 22:403–415, 2012.)

18.1.11.2. Comparison of Inside and Outside Mechanical Response

18.1.11.2.1. Animal Model

In this experiment, five male Sprague Dawley rats of 320–360 g weight were sacrificed and instrumented with a surface mount Kulite sensor (LE-080-250A) on the nose to measure the reflected pressure, and two Kulite probe sensors (XCL-072-500A, diameter: 1.9 mm, length: 9.5 mm) were implanted in the thoracic cavity and in the brain, respectively. For the brain sensor implantation, the tip of the sensor was backfilled with water to ensure good contact with tissue, and the sensor was inserted through the foramen magnum 4–5 mm into the brain tissue. Then the instrumented rat was placed at four different locations twice inside the shock tube and twice outside. Complete details on the experimental setup can be found elsewhere (Sundaramurthy et al., 2012).

Figure 18.12 shows incident pressure as well as pressure in the brain and thoracic cavity corresponding to various locations along the length of the shock tube. At placement locations a and b, incident pressure profiles follow the Friedlander waveform (Figure 18.4) fairly well. Pressure profiles in the brain and thoracic cavity also have similar profiles (the shape is almost identical) to that of the incident pressure profiles. At these locations, peak pressures recorded in the brain are higher than the incident peak pressure and the peak pressure recorded in the thoracic cavity is equivalent to the incident peak pressure. It is clear, at location c, the incident pressure profile differs significantly from the ideal Friedlander waveform; the overpressure decay is rapid and the positive phase duration is reduced from 5 ms at a to 2 ms at c (Figure 18.12a and c, respectively). The pressure profile in the brain shows a similar trend. The pressure profile in the thoracic cavity shows a secondary loading with higher pressure and longer duration. The pressure profile in d is similar to the pressure profile recorded in c, except the value of the peak pressure reported in the brain is lower than the incident peak pressure.

FIGURE 18.12

Measured pressure-time profile in the brain and thoracic cavity with their corresponding incident pressures at all animal placement locations (APLs). At APL (a) and (b), both intracranial and thoracic pressures follow the same behavior as incident pressure; (more...)

The biomechanical response of the animal significantly varies with the placement location. Inside the shock tube (a, b, Figure 18.12), the load is due to the pure blast wave, which is evident from the p-t profiles (Friedlander type) recorded in the thoracic cavity and brain. For animal placement locations (APLs) at the exit, c and d, p-t profiles show sharp decay in pressure after the initial shock front. This decay is due to the interaction between the rarefaction wave from the exit of the shock tube, eliminating the exponentially decaying blast wave, which occurs in a and b. This has two consequences: first, the positive blast impulse (area under the curve) reduces drastically. Second, because the total energy at the exit is conserved, most of the blast energy is converted from supersonic blast wave to subsonic jet wind (Haselbacher et al., 2007). This expansion of blast wave at the exit (subsonic jet) produces an entirely different biomechanical loading effect compared with the blast wave. Consequently, the thoracic cavity experiences secondary loading (i.e., higher pressure and longer positive phase duration). When the animal is constrained, this high-velocity subsonic jet wind exerts severe compression on the tissues in the frontal area (head and neck) that in turn causes pressure increase in the thoracic cavity (lungs, heart). To further illustrate the effect of subsonic jet wind on the rat, experiments a and c without any constraint were performed. Figure 18.13 shows the displacement (motion) of the rat at various time points starting from the moment the blast wave interacts with the animal. At a, the displacement is minimal; however, at c, the rat is tossed away from the bed (motion) because of jet wind. This clearly illustrates the effect of high-velocity subsonic jet wind on the rat when placed outside the shock tube. Consequently, the animal is subjected to extreme compression loading when constrained and subjected to high-velocity (subsonic) wind when free, both of which are not typical of an IED blast. This in turn changes not only the injury type (e.g., brain vs. lung injury) but also the injury severity, outcome (e.g., live vs. dead), and mechanism (e.g., stress wave vs. acceleration).

FIGURE 18.13

Motion of unconstrained rat under blast wave loading (a) inside and (b) outside. (i−iv) Time points t = 0, 20, 40, and 60 msec, respectively; the rat is thrown out of the bed when placed outside. (From Sundaramurthy, A. et al. Journal of Neurotrauma (more...)

18.1.11.2.2. Cylindrical Model

The evolution of the shock-blast wave along the length of the shock tube was measured using an aluminum cylinder (length 230 mm, diameter 41.3 mm). In this case, a cylinder was placed along the longitudinal axis of the shock tube at various offset distances from the exit (open end) both outside (+x) and inside (−x) (Figure 18.14). Seven holes were drilled and tapped to locate seven Dytran model 2300V1 piezoelectric pressure sensors used in conjunction with Dytran model 6502 mounting adapters. The location labeled t0 was centered between the two end surfaces of the cylinder, and the rest of the holes were evenly spaced for a total span of 84 mm. The cylinder was mounted (i.e., firmly secured) using brackets made out of flat steel bar. In addition to the gauges mounted on the cylinder, there were a set of gauges (PCB pressure sensor model 134A24) mounted at various locations on the shock tube (along the length) that measure the incident (side-on) pressures. The experiment was repeated three times at each location along the length of the shock tube (N = 3).

FIGURE 18.14

Experimental setup to measure evolution of the shock wave along the length of the shock tube. Placement of the cylinder at two representative locations along the length of the shock tube is shown. (From Chandra, N. et al., Shock Waves. 22:403–415, (more...)

Figure 18.15a and b shows the reflected pressure profiles for cylinder placement locations inside and outside of the shock tube. The reflected pressure measures total pressure (both kinetic and potential energy components) at a given point. The reflected pressure profiles for placement locations inside the shock tube show gradual decay in pressure and pressure profiles follow the Friedlander waveform. A small secondary peak in pressure profiles is due to reflection from the walls of the shock tube; however, these wall reflections do not significantly affect pressure profiles. The reflected pressure profiles for placement locations outside the shock tube show rapid pressure decay that does not conform to the Friedlander waveform; the shock front and pressure decay instead look like a delta function. This is followed by a long-duration, relatively constant low-pressure regime (starting points of which are demarcated by cross symbols). This long duration, relatively constant low-pressure regime is referred to as subsonic jet wind in this work. This jet wind is an artifact of the shock tube exit effect and does not occur in free field blast conditions.

FIGURE 18.15

(a) Experimentally measured p-t profiles at various x locations inside the shock tube. p-t profiles follow the Friedlander waveform fairly well. (b) Experimentally measured p-t profiles at various x locations outside the shock tube. In these profiles, (more...)

Figure 18.16a and b shows the impulse profiles for cylinder placement locations inside and outside of the shock tube. The total impulse is reduced significantly for outside placement locations when compared with inside placement locations. The shape of impulse profiles for placement locations inside the shock tube is relatively constant (i.e., gradual increase) as opposed to nongradual (i.e., with slope changes) for outside placement locations. The contribution of subsonic jet wind to the impulse is high starting points, which are demarcated by cross symbols.

FIGURE 18.16

(a) Impulse profiles at various x locations inside the shock tube obtained by integration of experimentally measured p-t profiles. (b) Impulse profiles at various x locations outside the shock tube obtained by integration of experimentally measured p-t (more...)

18.1.12. Tailoring the Shock-Blast Wave Profile

In previous sections, we learned about the design, theory, and working of a compressed gas shock tube and then critically analyzed the factors that influence the test section locations for simulating a primary shock-blast wave. In the following sections, we will study the methods to tailor a shock-blast wave profile. To attain this goal, an extensive experimental analysis was carried out to show the methods in which the blast wave profile can be tailored to replicate the field conditions. We study the effects of shock tube–adjustable parameters (SAPs) such as breech length, type of gas, membrane thickness, and measurement location on the SWPs. Furthermore, we characterize the flattop or plateau wave and determine the influence driver gas and breech length has on this phenomenon. Finally, we compare the shock-blast wave profile from the shock tube with the field explosion profiles generated in ConWep.

Figure 18.17 shows the experimental variables and the sensor locations. The length of the breech is varied between 67 and 1,210 mm. The membrane thickness is varied by varying the number of membranes between 1, 5, and 10 (each membrane is 0.254 mm thick). In this work, both nitrogen and helium were used as the driver gas, and the driven gas was air at ambient laboratory conditions (temperature range of 23 ± 20°C). The evolution of the blast wave along the length of the shock tube was measured using a PCB pressure gauges (model 134A24) mounted on the wall of the shock tube at locations A1, A2, X, B1, and B2 (Figure 18.17). Burst pressure in the driver just before the rupture of the membranes was also recorded.

FIGURE 18.17

Experimental variables and sensor location; here A1, A2, X, B1, and B2 are the side-on pressure sensors. (From Sundaramurthy, A. et al. Journal of Neurotrauma. 29:2352, 2012.)

18.1.12.1. Burst Pressure

Burst pressure is the pressure in the driver section (breech) during the membrane rupture. This highly compressed gas when allowed to expand compresses atmospheric air in the transition section generating a shock front. Burst pressure for different membrane thicknesses and breech lengths were noted and presented in Figure 18.18a. From Figure 18.18a, it can be seen that the burst pressure increases with an increase in the membrane thickness. Furthermore, there is no discernible difference in the burst pressure with respect to increase in breech length for any of the three membrane thicknesses. This is because the membrane rupture is pressure-dependent, and this critical pressure is not influenced by breech volume (i.e., burst pressure that can be achieved at a minimum breech length L₁ can also be achieved at L₂, L₃, or L₄). Therefore, the quantity of membranes used and their thickness is directly proportional to the burst. This result corroborates with the findings from the study conducted by Payman and Shepherd in which they used copper as their membrane. They determined that for the same thickness, the burst pressure does not vary more than ±3%. Similarly, they also determined that membrane thickness has a linear relationship with burst pressure (Payman and Shepherd, 1946).

FIGURE 18.18

(a) Relationship between the number of membrane used and burst pressure produced with respect to different breech lengths. (b) Relationship between shock front Mach number and burst pressure; there is a linear relationship between Mach number and burst (more...)

18.1.12.2. SAPs and Their Influence on the SWPs in the Test Section

By changing the SAPs such as membrane thickness (burst pressure) and breech length, we were able to determine the effect on SWPs such as Mach number, overpressure, and positive time duration (PTD) in the test section. Figure 18.18b and c shows the relationship between overpressure, Mach number, and burst pressure for different breech lengths. Both of these variables have a strong positive relationship with the burst. Furthermore, with increase in the burst pressure both Mach number and overpressure for L₂, L₃, and L₄ increases with a higher rate than L₁.

PTD is the period within which the shock overpressure reaches the atmospheric pressure. Figure 18.18c shows the relationship between PTD and membrane thickness for different breech lengths. For a given membrane thickness, PTD increases with an increase in the breech length. Furthermore, there is an increase in PTD between membrane thicknesses 1 and 5 for breech lengths L₁, L₂, and L₃; however, such an apparent difference is not observed between membrane thicknesses 5 and 10. Finally, for breech length L₄, there is no apparent difference in PTD for different membrane thicknesses.

Controlling overpressure and PTD is essential when replicating a field blast. It is possible in a laboratory shock tube to control the aforementioned variables by manipulating breech length, burst pressure (membrane thickness), type of gas, and test section location (by varying the test section within expansion section). It can be seen within the test section that, with an increase in burst pressure, both the overpressure and the Mach number (strength of shock wave) increases, which implies that both these variables can be increased by increasing the membrane thickness. Similarly, PTD increases with increase in breech length for any given burst pressure. However, at lower breech lengths, both overpressure and PTD are affected by expansion waves (also known as rarefaction waves) that arise from the end of the breech. Furthermore, type of driver gas used also plays a role in the PTD and overpressure. The effect expansion waves and driver gas have on the overpressure and PTD is explained in detail in the following section.

Figure 18.19 shows the x-t wave propagation diagram with shock front, rarefaction head, and tail for two breech lengths. When membranes burst, the driver gas expansion initiates a family of infinite expansion waves toward the closed end. Once the expansion head encounters the closed end, it travels backwards toward the transition. Because of increased density and pressure from the traverse of the shock front, the expansion head accelerates further and catches the shock front. However, because of the expansion, the temperature and pressure in the breech reduces, which reduces the velocity of the successive expansion waves, resulting in a series of waves, which yields the nonlinear decay and ultimately shapes the shock-blast wave (Ritzel et al., 2011). Once the waveform attains the shape shown in the Figure 18.4, expansion waves start to deplete the overpressure and PTD. Similarly, the breech length and driver gas also play major roles in the evolution and interaction of the expansion wave. The expansion wave while traveling toward the closed end of the driver section travels at least the velocity of sound in that medium, which is the driver gas (helium C_He = 972 m/s and nitrogen C_N = 353 m/s). Therefore, for a given breech length and membrane thickness, having helium as a driver gas increases the expansion wave velocity resulting in a Friedlander type wave at an earlier point compared with nitrogen gas. Consequently, by varying the length of the breech in conjunction with using the appropriate driver gas, we can optimize PTD and overpressure.

FIGURE 18.19

Ideal breech length x-t diagram for explosive shock wave replication. (From Sundaramurthy, A. et al. ASME 2013 International Mechanical Engineering Congress and Exposition, 2013.)

18.1.12.3. Evolution of the Shock-Blast Wave along the Expansion Section

Figure 18.20(a–c) shows the evolution of the overpressure along the length of the expansion section. From Figure 18.20a, it can be seen that for one membrane there is no discernible change in overpressure for breech lengths L₂, L₃, and L₄. For all cases with breech length L₁, there is a continuous decay in the overpressure downstream of the shock tube. For all the other breech lengths, unique points of overpressure decays are identified along the expansion section, which is illustrated in the following section.

FIGURE 18.20

Variation of shock-blast profile parameters along the length of the shock tube expansion section. All of these experiments were performed for breech lengths 66.68 (black), 396.88 (red), 803.28 (blue), and 1,209.68 (green) mm; (a−c) show the variation (more...)

For L₁ (66.68 mm), L₂ (396.88 mm), L₃ (803.28 mm), and L₄ (1,209.68 mm), we observe the following: (1) for any membrane thicknesses, an obvious difference in overpressure is observed between L₁ and other breech lengths (Figure 18.20a–c); (2) beyond 3,000 mm from the breech, for 5 and 10 membranes and breech length L₂, overpressure starts to decay (Figure 18.20b, c); and (3) beyond 5,000 mm from the breech, for 10 membranes and breech lengths L3, the overpressure starts to decay (Figure 18.20c). Finally, for L₄, there is no unique decay point, which implies a flattop wave throughout the expansion section.

In a typical free field blast, the overpressure decreases rapidly with respect to increase in distance from the blast epicenter (Mott et al., 2008). However, overpressure in a shock tube does not show a drastic reduction because of its constant cross section. There is a considerable difference between the overpressures for L₁ and all the other breech lengths. As discussed earlier, this difference arises from the interaction of expansion waves that comes from the breech. This suggests that the expansion waves from the breech for breech length L₁ arrives earlier than all other breech lengths. With an increase in the breech length and burst pressure, distinct points at which the shock blast starts to decay are shown in Figure 18.20b and c, which implies that beyond this point the shock-blast wave has a Friedlander form. Consequently, longer breech lengths that tend to produce a flattop wave in the test section will produce a Friedlander type wave at some point beyond the test section. As discussed in Section 18.1.3, when moving closer to the exit, the rarefaction waves from the exit starts to interact with blast wave creating artifacts, which results in inaccurate blast simulation.

Figure 18.20d–f shows the evolution of the PTD along the length of the shock tube expansion section. For any given breech length and membrane thickness, the PTD remains reasonably constant along the length; however, it decreases drastically toward the exit of the shock tube. Positive impulse (PI) is the area under the shock-blast wave profile. Figure 18.20g–i shows the evolution of the PI along the length of the shock tube expansions section. PI is a function of both overpressure and PTD; hence, it increases with an increase in both membrane thickness and breech length. Because of its relationship with the PTD, the impulse drastically reduces near the exit of the shock tube.

As a result, PTD reduces drastically near the exit of the shock tube because of the interaction between shock front and exit expansion waves. This has two consequences: first, the PI (energy of blast wave) reduces drastically (Figure 18.20g–i). Second, because the total energy at the exit is conserved, all the blast energy is converted from supersonic blast wave to subsonic jet wind, which produces erroneous results. The effect of jet wind and specimen placement location along the expansion section for blast simulation using a shock tube is illustrated in Section 18.1.2.2.

18.1.12.4. Flattop or Plateau Wave

A flattop or plateau wave is usually witnessed in a gas-driven shock tube (Reneer et al., 2011). In this case, the shock-blast wave profile, once reaching the peak overpressure, maintains its peak value for a certain period before decay. Longer breech lengths in combination with the use of nitrogen as a driver gas seem to have a strong influence on this phenomenon. Figure 18.21 shows the comparison between the shock-blast wave profile with nitrogen and helium as the driver gas. In both cases, 10 membranes with a breech length of 1,209.68 mm were used. It can be seen that only in the case of nitrogen as the driver gas is a flattop wave observed, whereas in the case of helium, a pure Friedlander wave is witnessed.

FIGURE 18.21

Comparison of the shock-blast profile for helium and nitrogen with 10 membranes and breech length of 1,209.68 mm; clearly, the wave profile corresponding to helium gas is a Friedlander wave and the wave profile corresponding to nitrogen is a flattop wave. (more...)

There is an inherent relationship between SAPs, such that an optimization of one variable might have a deviating effect on the other variables, resulting in having an nonoptimal shock blast wave (in this case a flattop wave). This particular problem arises depending on (1) breech length and (2) type of driver gas. In the current study, the breech length L₁ is low enough for the expansion waves to reach the shock front and create a Friedlander type wave somewhere before the test section. Therefore, when it arrives at the test section, it has already gone through some overpressure and PTD reduction. Once the breech length is increased, the time taken by the expansion wave to reach the shock front increases. Consequently, for breech lengths L₂, L₃, and L₄, there is no change in the overpressure and PTD at the test section, which implies that it is a flattop wave.

Although nitrogen, because of its low sound speed, has a tendency to produce a longer PTD, using a longer breech length results in a flattop wave. Conversely, helium produces a lower PTD compared with nitrogen but has a sharp decay to the atmospheric pressure. Similar findings are reported by Reneer and Bass in their separate works comparing the wave profiles generated from air (which has a speed of sound close to nitrogen) and helium. They found that using air as a driver gas produces a flattop wave (Bass et al., 2012; Reneer et al., 2011). One technique used for avoiding a flattop wave when using long breech length is to place the test section downstream of the expansion section, so that the expansion waves would eventually catch up and produce a Friedlander type wave; nevertheless, this method has its own limitations.

18.1.12.5. Comparison between Field and Laboratory Profiles

Table 18.1 shows the range of IEDs and mines typically used in the field; their explosive capacity is expressed as TNT strength (AEP-55, 2006; Department of Homeland Security, 2009). An important requirement for studying BINT is the ability to produce accurate and repeatable blast loading, which can be related to strengths mentioned in Table 18.1. Using ConWep, we were able to determine the pressure profiles for TNT explosives within the range of strengths described in Table 18.1. A comparison was made between the profiles generated from ConWep with those generated from a shock tube device (Figure 18.22a–c, d). Clearly, there is good match in the results, which indicates that the wave profile generated by the compressed gas shock tube is directly related to relevant field conditions.

TABLE 18.1

Explosive Capacity of the Currently Used IEDs and Mines in the Field

FIGURE 18.22

Comparison of the shock blast profiles from a UNL shock tube device and ConWep simulation software. (a) Comparison between shock blast profile from a 10-membrane, 66.68-mm breech length shot with nitrogen as driver gas, and 2.56 kg of TNT at 5.18 m. (b) (more...)

18.2.1. Experiments

18.2.1.1. PMHS Testing in the 28-Inch Shock Tube

The PMHS heads were used in conjunction with the Hybrid III neck in these experiments. The head assembly was placed in the test section of the shock tube, as shown in Figure 18.23, and was subjected to frontal blast loading. Experiments were also conducted with helmets mounted on PMHS heads. The shape, overpressure, and duration of the incident blast wave at a given location are known a priori. This is achieved through sample trials in the shock tube, conducted without the surrogate head and the neck.

FIGURE 18.23

(a) PMHS heads with hybrid III neck placed in the test section of the shock tube. (b) Schematic showing sensor locations on PMHS head. (c) CT images of instrumented PMHS showing sensor locations. Anthropometric data were also obtained from these CT images. (more...)

18.2.1.1.1. PMHS Preparation

Three PMHS heads were used for this purpose. The specimens had no record of osseous disease and preexisting fractures were not present as confirmed by computed tomography (CT) imaging. The age, gender, and basic anthropometry of the specimens are listed in Table 18.2.

TABLE 18.2

Characteristics of the Three PMHS Heads Tested in This Study

PMHS heads that are kept frozen for a substantial period tend to have severe tissue degradation after thawing. Thus, the brain was removed from each PMHS head and intracranial space was backfilled with ballistic gelatin. The brain tissue and dura mater were removed through the foramen magnum using a flat head screwdriver. Twenty percent ballistic gelatin (ballistic gel, from here on) was prepared by dissolving 2 parts of 250 bloom gelatin into 9 parts of warm (40°C) water (by mass), stirring the mixture while pouring in the powdered gelatin. The gelatin was obtained from Gelita USA Inc. (Sioux, IA) in the bloom form. The ballistic gel was poured in the intracranial cavity through the foramen magnum and allowed to settle at room temperature. After the ballistic gel had settled, the entire head was put inside plastic bags and air bubbles were removed using a vacuum cleaner. The foramen magnum was sealed using filler material (Bondo). A Hybrid III neck was attached to the head using a base plate. A base plate was screwed to the bottom of the head.

18.2.1.1.2. Instrumentation

Each PMHS head was instrumented to measure surface pressures, surface strains, and ICPs. Eleven sensor measurements were made on each PMHS head. Surface pressures were measured at two locations, surface strains were measured at four locations, and ICPs were measured at five locations within the head, as shown in Figure 18.23b. CT images of the instrumented head were also taken. CT images were used to verify locations of the sensors inside the head (Figure 18.23c). In addition, CT images were also useful in identifying the precise geometry of skull and the face in the vicinity of the sensor; by geometry, we implied anatomical features such as the air sinus, eye socket, and nasal cavity. For example, a huge air sinus was present in front of certain sensors (see Figure 18.23c, nose ICP). Surface pressures were measured using Kulite surface mount sensors (model LE-080-250A) and ICPs are measured using Kulite probe sensors (XCL-072-500A). Surface strains were measured using Vishay strain gauges (model CEA-13-250UN-350). In addition to these sensors, a PCB pressure gauge (model 134A24) was used to measure incident (side-on) pressure of the blast wave. Incident pressure was measured just before (distance = 200 mm) the blast wave encountered the PMHS head. An incident pressure gauge was mounted on the wall of the shock tube.

18.2.1.2. Blast Wave Exposure

All PMHS heads were subjected to blast waves of three different incident intensities or overpressures (70, 140, and 200 kPa). This was achieved by changing the number of membranes, which resulted in different burst pressures. The head was oriented in a frontal direction to the blast. Each intensity was repeated three times for each head, so there were a total of 27 (3 × 3 × 3) shots. The PMHS head was placed in the test section of the shock tube located approximately 2,502 mm from the driver end; the total length of the shock tube was 12,319 mm.

18.2.1.2.1. Sample Pressure-Time Profiles

Figure 18.24a–h shows sample incident pressure (a), surface (reflected) pressure (b, c), and ICP (d–h), respectively. Sample pressure-time profiles presented here are based on the mean of three shots (experiments) for head 1. While presenting sample pressure-time profiles, the time axis was shifted so that arrival of the blast wave at a given sensor location corresponded to t = 0. Incident blast intensity for these profiles was 200 kPa. Raw (pressure) data showed oscillations; thus, the profile was smoothed by performing simple moving average (Kraus et al., 2007). The number of data points selected for moving average varied from 5 to 20.

FIGURE 18.24

(a) Incident pressure profile. Incident pressure profile is measured 200 mm (upstream) from the PMHS head. (b, c) Sample pressure profiles on the surface of the head: (b) forehead, (c) temple. (d–h) ICP profiles: (d) forehead ICP, (e) nose ICP (more...)

Figure 18.24a shows the incident pressure profile. The incident pressure was measured at 200 mm (upstream) from the PMHS head. The incident profile showed a sudden rise in pressure (rise time is 10 µs) followed by nonlinear decay; the peak pressure and positive phase duration were 190 kPa and 5.4 ms, respectively. Secondary peaks were also seen in the incident profile. The incident pressure profile was highly repeatable; shot-to-shot variations in the pressure profile were less than 5%.

Figure 18.24b and c shows pressure profiles on the surface of the head. Surface pressures were measured at the forehead and temple locations. The profile for forehead location showed a sudden rise in pressure (rise time was 30 µs) followed by nonlinear decay. The rate of decay was much faster than that observed in the incident pressure profile. The peak pressure and positive phase duration for this profile were 592 kPa and 5 ms, respectively. The peak pressure was amplified 3.11 times the peak incident pressure but the positive phase duration remained approximately similar at 5 ms. The pressure profile for temple location also showed a sudden rise in pressure (rise time was 30 µs) followed by another spike. This was followed by a sudden (instantaneous) decay in pressure with huge oscillations (from t = 0.25 to t = 0.5). An oscillating profile with much smaller peaks was observed after this time. The peak pressure and positive phase duration for this profile were 295 kPa and 5 ms, respectively. The ratio of peak pressure to peak incident pressure was 1.55.

The shape and positive phase duration of the surface pressure profile at the forehead location was similar to that of the incident pressure profile. However, peak overpressure was significantly higher because of aerodynamic effects and the rate of decay was much faster compared with the incident pressure profile. Temple location had a much smaller peak pressure (50% decrease) compared with the forehead location. These effects can be explained by studying blast wave–head interactions. The flow field around the head is illustrated using numerical simulations. At the beginning of the interaction, as the shock front impinges on the forehead at its most upstream region, a reflected shock propagating in the opposite direction starts to develop. The incident shock starts to propagate around the surface of the head. At the same time, regular reflections occur that propagate radially both in the upstream and downstream directions. These reflections continuously interact with the incoming tail part of the blast wave. The reflections are tensile in nature and hence a compressive pattern of decreasing strength develops as a result of the incident shock reflection over the surface and the forward motion of the shock tail (also known as blast wind) as shown in Figure 18.25a. Thus, the forehead surface gauge recorded a faster decay compared with the incident pressure profile. As the shock wave traversed the head, shock wave diffractions occurred and as a result geometry-induced flow separation took place (Figure 18.25b). This geometry-induced flow separation increased as we moved away from the leading edge (or incident blast site) downstream. Because of this flow separation, the temple location showed a decrease in peak pressure with respect to the forehead location. A similar phenomenon of shock wave diffractions and flow separation over cylindrical objects was seen in studies involving shock wave propagation over cylindrical objects (Bass et al., 2012; Brun, 2009; Gould and Tempo, 1981).

FIGURE 18.25

Blast wave−head interactions as blast wave traverses the head. (a) Blast wave−head interaction at leading edge or incident last site. (b) Illustration of Mach reflection and flow separation as blast wave traverses the head. In all figures, (more...)

Figure 18.24d–h shows the ICP profiles for sensor locations shown in Figure 18.23b. For ICP profiles, wave action (dynamic events) played out in very short time. Positive phase duration was 2.5 ms with initial (or majority of) intracranial dynamics playing out within 0.5 ms. The forehead (FH) ICP (Figure 18.24d) sensor showed a sharp rise (rise time = 40 µs) in pressure profile. This was followed by a decay in pressure until t = 0.15 ms; an abrupt pressure increase was also seen during this decay at t = 0.07 ms. The decay was not sustained and there was another rise in the pressure. A secondary (loading) pulse, which was similar to a pulse observed in impact loading, was seen after 1.0 ms. The nose ICP sensor (Figure 18.24e) also showed a sharp rise (rise time = 70 µs) in the pressure profile. This was followed by decay in pressure until t = 0.5 ms; during this decay, a distinct pressure increase (like another peak) was seen at t = 0.2 ms. After this decay, the pressure remains approximately constant with small oscillations until t = 1 ms. Secondary pulse, which is similar to the pulse observed in the FH ICP profile, was seen after 1 ms. Figure 18.24f shows the pressure profile for the center ICP. The pressure rise (rise time = 390 µs) is not as sharp compared with the FH and nose ICP profiles. The pressure pulse seemed to repeat itself with damping until the pressure equilibrates. Figure 18.24g shows the pressure profile for temple ICP. The profile has similar features as that of center ICP. It appears that the center and temple ICPs experience several waves that are emanating from different sources during the rise. Rise time for the temple ICP was 115 µs. Figure 18.24h shows the pressure profile for the back ICP. The back ICP shows the negative phase followed by the positive phase. The negative phase had a rise time of 100 µs and duration of 0.25 ms. The positive phase of the back ICP has features similar to the FH and nose ICP profiles. This is followed by decay in pressure after which pressure equilibrates.

Analysis of the peak pressures of ICP profiles has led to some interesting observations. Peak pressures are marked with a black cross on each ICP profile. The highest peak pressure is observed behind the forehead. The peak pressures were decreased as we move away from the coup (impact) site toward the contrecoup (opposite to impact) site. It should also be noted that a significant difference in peak pressure is observed for FH (430 kPa) and nose (310 kPa) ICPs that are in same coronal plane.

The pressure profiles in the brain drastically deviated from surface pressure profiles. ICP dynamics play out in much shorter duration compared with incident or surface pressure profiles. For ICP profiles, the positive phase duration was 2.5 ms with initial (or the majority of) intracranial dynamics playing out within 0.5 ms; this was due to wave propagation in skin–skull–brain parenchyma. The impedance mismatch of the layered system (skin–skull–brain) was such that the magnitude of the pressure wave (or input signal) in the skin–skull–brain parenchyma was either amplified or attenuated as it reflected and transmitted through these layers. The wave traversal times through the skin and the skull were 52.9 µs and 3.62 µs assuming 10-mm thickness for both skin and skull. Within 506 µs (i.e., 0.506 ms) two wave traversals occurred in the skin and 29 wave traversals were possible in the skull. Thus, wave action (reflections and transmissions) happens at a much shorter time scale, and a sharp decay in ICP profiles was seen for FH and nose ICPs. This aspect is further elaborated using a one-dimensional model of skin–skull–brain parenchyma as shown in Figure 18.25c, d. Figure 18.25c shows response of the skull–brain parenchyma when a loading pulse (Heaviside function) of intensity P is applied to the skull. Pressure in the brain was 0.62 (transmission coefficient) times the applied pressure after first transmission from the skull into the brain. Transmission coefficient increased with each transmission and pressure in the brain equilibrated after the fifth transmission from the skull into the brain. Figure 18.25d shows response of the skin–skull–brain parenchyma when a loading pulse of intensity P was applied to the skin. In this setup (or model), thickness of the skull and skin were designed such that (t_skull = 14 t_skin) reflection/transmission occurs at the same time for skin–skull and skull–brain interfaces. The transmission coefficient was 1.16 after the first transmission from the skull into the brain. However, transmission coefficient drastically decreased to 0.15 after the second transmission. Subsequent transmissions had transmission coefficients of 1.39, 0.42, 1.32, and 0.7. In reality, the thickness of the skin and the skull were almost similar and hence the wave traversal in the skull was ~14 times faster than the skin based on the wave speeds. The transmitted wave from the skull into the brain equilibrated after the fifth transmission (Figure 18.25c). Hence, for wave propagation through skin–skull–brain parenchyma in real scenarios skull (thickness) should not play a major role in the wave amplification or attenuation and brain should ideally see the pressure that is seen by the skull at the skin–skull interface. Even in that case, the transmission coefficient in the brain after the second transmission was 0.25. This explains the sharp decay in ICP profiles for FH and nose ICPs that were closest to the incident blast site wherein initial wave propagation obeys one-dimensional theory fairly well (at least in terms of qualitative trends). For FH ICP, the pressure decreased from 430 kPa (first transmission) to 138 kPa (second transmission) in 117 µs; the ratio of these pressures is 0.32, which is consistent with the one-dimensional theory.

The FH and nose ICPs showed a second peak during the initial decay; the second peak was abrupt in the FH ICP and distinct in the nose ICP. This second peak was due to a delayed wave transmission from the eye socket/eyebrow region as illustrated in Figure 18.26. After these initial phases, wave reflections from the head boundaries dominated the response and it was not possible to delineate these effects because of the complex and highly dynamic nature of the problem. The center and temple ICPs, which are probably located deep inside the brain, experienced many waves emanating from different sources. By the time the wave reached the center ICP, the pressure was attenuated. This attenuation could be due to material damping, wave dispersion over a larger area, and reflections from geometric boundaries and material interfaces. The pressure pattern of the center ICP can be best described as follows. At any given point, the brain experiences a complex set of direct and indirect loadings emanating from different sources (e.g., blast wave transmission, reflections from tissue interfaces, skull deformation) at different time points. These disturbances continuously propagate into the brain as waves. Constructive and destructive interferences of these waves control the pressure history deep inside the brain. The back ICP shows the contrecoup effect (negative pressure) initially because the wave velocity in the skull is higher than the wave velocity in the ballistic gel. Because of this, the skull moves forward; displacement of the brain lags displacement of the skull and hence tension or negative pressure is generated in the brain. If tensile loads/forces are not allowed to transfer, then separation will take place at the skull–brain interface.

FIGURE 18.26

Comparison of surface and ICP profiles near the eye socket regions.

18.2.1.2.2. Pressure Response as a Function of Incident Blast Intensity

Figure 18.27 shows the positive phase impulse for various sensors as a function of intensity. The positive phase impulse increased as the incident blast intensity increased. The differences were statistically significant (p < 0.05) at all sensor locations except the back ICP sensor for all heads and temple ICP for head 1. Head-to-head variations in the positive phase impulse were also significant; impulse variations up to 77% were observed between the heads.

FIGURE 18.27

Positive phase impulse for various sensors as a function of intensity.

As mentioned earlier, one of the lingering questions facing the medical and scientific communities is whether blast waves cause TBI. We exposed PMHS heads to pure primary blasts of varying intensities and observed statistically significant differences in the peak ICP and total impulse. This finding supported our hypothesis that intracranial response changes with change in incident blast intensity. Thus it is clear that primary blast waves acting alone can cause mechanical insult to the brain. The potential of this mechanical insult in causing the BINT will be assessed in the following section. Over the past few years, several mechanisms (see the previous section) have been proposed based on numerical simulation without experimental backup. Our experimental measurements categorically indicated that the blast wave transmission through the cranium induced mechanical insult to the brain of varying degrees that changed with variations in incident intensity. Thus we propose direct transmission of the blast wave into the intracranial cavity as an essential loading pathway to the brain.

18.2.1.2.3. Sample Strain Profiles

Figure 18.28 shows sample (circumferential) strain profiles for strain gauge locations shown in Figure 18.23b. Incident blast intensity for these profiles was 200 kPa. Negative strain indicated compression and positive strain indicated tension. The front strain gauge showed a compressive phase up to 1 ms; this was followed by small (equilibrium) oscillations for 0.2 ms, followed by another compressive pulse. The right temple strain gauge showed initial tension followed by compression. This compressive phase was sustained for 0.2 ms only. This was followed by a tension-compression phase with higher magnitudes and longer durations. Top and back strain gauges showed several compressive phases; each compressive phase was followed by equilibrium oscillations (i.e., small oscillations around zero). The highest circumferential strain was observed for the front location at 0.06%. The shape of the strain profiles was not consistent across the heads and incident intensities or in some cases, even from experiment to experiment for a given head. The magnitude of the strain however was on the same order of magnitude. Head-to-head variation of ±40% was observed in the peak strain.

FIGURE 18.28

Strain profiles for various strain gauge locations.

18.3.1. Experiment

Adult 10-week-old male Sprague Dawley rats weighing 320–360 g were used in all studies. Three studies with separate groups of rats were performed. In the first study, five cadaver rats were used to record the pressure on the surface of the nose (reflected pressure), in the lungs, and in the brain. In the second study, rats were anesthetized and exposed to the blast and were sacrificed immediately after blast exposure for gross lung pathology and histology evaluation (27 rats). In the third study, monitoring of physiological vitals such as heart rate, blood oxygen saturation, and perfusion index was performed for two periods of 30 minutes, before and after the blast injury (50 animals). The 14 rats in this group were transcardially perfused 24 hours after the blast exposure, and the brain sections were evaluated by histological and immunohistological methods.

The rats were anesthetized with mixture of ketamine and xylazine (10:1 [100 mg/10 mg/kg], 0.1 mL/100 g) administered via intraperitoneal injection. Rats were exposed to the blast wave in the test section located inside of the shock tube (i.e. 2800 mm from the breech). The tests were performed at five discrete incident peak overpressures detailed in Table 18.3. The incident pressure was controlled by adjusting the number of Mylar membranes, and keeping the breech length constant at 447.7 mm. An aluminum bed was designed and fabricated for holding the rat during the application of blast wave. The aerodynamic riser is attached to the bed to hold the sample in the center of the shock tube. Typically, rats are tested in a prone position and are strapped securely to the bed with a thin cotton cloth wrapped around the body (Figure 18.29a). Sham control rats received anesthesia and noise exposure, but without blast exposure, and naive control rats were given anesthesia only.

TABLE 18.3

The Peak Overpressure and Impulse Average Values Measured by Side-On Sensor (Incident Pressure)

FIGURE 18.29

Schematic representation of the shock tube facility (a): left inset illustrates the side-on pressure sensor location (arrow) and the position and strapping of the animal in the holder; right inset: close-up high-speed video still images demonstrate the (more...)

18.3.1.1. Biomechanical Loading Evaluation in Cadaver Rats

Five male Sprague Dawley rats were sacrificed using carbon dioxide exposure. After the sacrifice, they were instrumented in the same manner as discussed in previously (Figure 18.30a). Cadaver rats were exposed to five shots per animal at 130 kPa (two rats) and 190, 230, 250, and 290 kPa (two rats). One rat with lungs filled with ballistic gel to ameliorate inconsistent pressure readings was exposed at five blast overpressures (130–290 kPa). To study the harmonic effects in the ICP signal, fast Fourier transformation (FFT) analysis was used.

FIGURE 18.30

The diagram of sensor locations (a) used in the cadaver rat experiments. The rightmost inset shows the brain remains intact after the experiments. Representative pressure profiles of the 190-kPa group (b), and average peak overpressures recorded by side-on, (more...)

18.3.1.2. Live Animal Testing

18.3.1.2.1. Mortality Evaluation

To study the mortality, the anesthetized animals were subjected to various levels of blast, characterized by the peak overpressure ranging from 130 to 290 kPa (Table 18.3). No resuscitation was performed after trauma.

18.3.1.2.2. Physiological Evaluation

Body weight measurements were performed before and 24 hours postexposure. Multiple vital signs were measured noninvasively using clinically validated MouseSTAT, a rodent physiological monitoring system (Kent Scientific Corp., Torrington, CT). Briefly, the anesthetized rat was placed in a supine position on the warming pad attached to the control unit to maintain natural body temperature and prevent hypothermia. The temperature was measured rectally at a rate of 1 Hz over the 30-minute period before and after the blast exposure. Simultaneously, heart rate, blood oxygen saturation, and perfusion index were recorded by a y-clip sensor clamped to the back paws of the rat.

18.3.1.2.3. Injury Evaluation

Immediately after the blast exposure, the lungs were extracted from the thoracic cavity, placed in 30–40 mL of freshly prepared solution of 10% formalin, and stored at 4°C for further evaluation. Lung injury severity associated with the primary blast exposure was evaluated using the Pathology Scoring System for Blast Injuries developed by Yelverton (Yelverton, 1996). The extent of injury is defined by the elements if injury severity (IS) according to the equation

where E is the extent of injury to the lungs (range 0–5); G is the injury grade including the surface area of the lesions (range, 0–4); ST represents the severity type elements, which classify the type of the worst-case lesions (range, 0–4); and SD is the severity depth element, indicating the depth or the degree of disruption of the worst-case lesion (range, 1–4).

18.3.1.2.4. Transcardial Perfusion and Histological Procedures

At 24 hours postexposure, rats were anesthetized with ketamine/xylazine mixture and transcardially perfused with wash solution (0.8% NaCl, 0.4% dextrose, 0.8% sucrose, 0.023% CaCl2, and 0.034% sodium cacodylate), followed by fresh fix solution (4% paraformaldehyde solution supplemented with 4% of sucrose in 1.4% sodium cacodylate buffer, pH = 7.3). Immediately, after perfusion, heads were decapitated and stored for additional 16–18 hours in the excess of fix solution. The brains were subsequently extracted the following day, and stored refrigerated (4°C) in a sodium cacodylate storage buffer. Brains were shipped in phosphate-buffered saline (pH = 7.2) to Neuroscience Associates (Knoxville, TN) for sectioning and staining.

The brains were embedded into one gelatin matrix block using MultiBrain Technology. The block was then allowed to cure and subsequently rapidly frozen in isopentane chilled with crushed dry ice (−70°C). The block was mounted on a freezing stage of a sliding microtome and coronal sections with a thickness of 40 μm were prepared. All sections were collected sequentially into an array of 4 × 6-inch containers that were filled with antigen preserve (buffered 50% ethylene glycol). At the completion of sectioning, each container held a serial set of one of every twelfth section, i.e., single section at every 480-μm interval). These sections were stained with an antibody against rat IgG to assess the BBB damage, and with amino cupric silver staining to detect neurodegeneration.

18.3.2. Results and Discussion

18.3.2.1. Biomechanical Loading Evaluation and ICP Analysis

Typical pressure profiles for all four sensors used (denoted as side-on, nose, brain, and lung) are presented in Figure 18.30b. The intensity of the reflected pressure measured on the rat’s nose was higher than the intensity of side-on pressure, and both were characterized by gradually increasing values (Figure 18.30c). The reflected (nose sensor) and ICPs were higher than the incident pressure (p < 0.001) as observed in Figure 18.30c, but there are no clear differences between reflected pressure and ICP. Moreover, the ICP not only was higher than the incident pressure, but also showed an oscillatory tendency. Although the reflected pressure showed a monotonic increase with side-on pressure for all of the pressure groups (130, 190, 230, 250, and 290), the same cannot be said for the ICP: there were no statistically significant differences between the 190- and 230-kPa groups (p = 0.70) (Figure 18.30c). The impulse values for the brain sensor indicated there were statistically significant correlations between the nonlethal group at 230 kPa and the two lethal groups (i.e., 190 and 250 kPa). This indicates the outcome (mortality) is not a function of the total energy transferred to the brain, but depends on the specific response of the cranium. The mortality rates among animals exposed to different levels of blast overpressures are shown in Table 18.3. Exposure to 130 kPa (n = 18) and 230 kPa (n = 10) resulted in no animal death, whereas exposure to 190 kPa (n = 20), 250 kPa (n = 17), and 290 kPa (n = 6) peak overpressure resulted in 30%, 24%, and 100% mortality, respectively.

To elucidate various modes of blast loading, we performed a set of analyses on the data reported by the brain pressure sensor. Using peak–trough analysis, we correlated biomechanical loading at the early, most violent stages of blast wave–cranium interactions (initial 0.4 ms, Figure 18.31a, b), with the blast-induced mortality (Table 18.3). The monotonically increasing peak overpressure and impulse values measured outside (side-on sensor) or inside of the cranium did not follow the same trend as mortality data in the Table 18.3. However, there is a strong correlation between the sum of the first two peak-to-peak amplitudes and mortality (Figure 18.31c). The ICP data subjected to FFT analysis indicated the oscillations have harmonic characteristics, and the high frequency component (10–20 kHz) is present exclusively in the case of the lethal groups (190, 250, and 290 kPa, Figure 18.31d). Moreover, the maximum ICP frequency increases with increasing blast intensity: peaks at 12.6 kHz (190 kPa), 16.4 kHz (250 kPa), and 19.3 kHz (290 kPa) marked with arrows in Figure 18.31d. In the 290-kPa group, two characteristic frequencies are noted (16.4 and 19.3 kHz). The ICP pressure oscillations are caused by the specific material response of the rat skull, which was also demonstrated in a simplified numerical model of in vivo response to blast proposed by Moss (Moss et al., 2009).

FIGURE 18.31

Representative pressure traces for five exposure groups (a). The signal was smoothed using a FFT band-pass filter for clarity of presentation. (b) Schematic representation of peak−trough analysis on overpressure recorded by brain sensor: Δpt (more...)

The intrathoracic pressures (lungs) were substantially lower than the incident overpressure and ICPs. However, there were considerable difficulties during pressure measurements in lungs caused by uncontrolled interactions between the pressure sensor and walls of internal organs in the abdomen. It is clearly seen in the experimental data as high variability of peak overpressure values between data sets and relatively high experimental errors (Figure 18.30c). In an attempt to mitigate these issues, we filled up lungs with ballistic gel (20% gelatin), but the recorded pressure profiles in this configuration were of much higher magnitude (i.e., comparable with pressures recorded in the brain).

18.3.2.2. Physiological Parameters

We noted statistically significant weight loss in both 190- kPa and 250-kPa injury groups 24 hours after blast exposure (Figure 18.32). Body weight loss proportional to the blast intensity was reported in mice (Cernak et al., 2011; Wang et al., 2011), and this is typically a manifestation of impairment of the neuroendocrine system (hypothalamic-pituitary-target organ axes) (Brown et al., 1993) associated with the damage to the hypothalamic region of the brain.

FIGURE 18.32

Exposure to different intensities of primary blast results in body weight decrease in groups 190 and 250 kPa. In these groups, statistically significant differences (marked with asterisk) were established immediately before and 24 hours after the exposure (more...)

Heart rate monitoring was performed over the 30-minute period before and after the blast exposure (Figure 18.33a, b). Observed changes in the heart rate difference between control groups were nonsignificant (Figure 18.33c, p > 0.05). The blast groups, as expected, showed an onset of bradycardia occurring immediately after the blast exposure (Figure 18.5c, p < 0.05): the heart rates decreased by 40 ± 9 (130 kPa), 62 ± 7 (190 kPa), 62 ± 15 (230 kPa), and 62 ± 10 (250 kPa) bpm. It appears that blast-induced bradycardia (i.e., expressed as the difference of average heart rate measured during 30-minute intervals before and after exposure) follows a simple dose-response model, with an upper value of 62 ± 4 bpm, and inflection point at 126 kPa (Figure 18.33d). However, all the groups of rats exposed to the blast are correlated, with the 130- and 190-kPa groups being borderline correlated (p = 0.065). Animals surviving blast exposure had a significantly decreased heart rate, compared with control (Figure 18.33). At high peak overpressure values (i.e., 190 kPa or higher), the heart rate is independent of the blast intensity (Figure 18.33c, d). The blast-induced bradycardia in the current model is similar to that reported in other studies using rats (Guy et al., 1998; Irwin et al., 1999; Knöferl et al., 2003). However, in blast trauma mouse models, the opposite relationship was observed: exposure to the mild blast resulted in the heart rate increase immediately after the blast (Cernak et al., 2011). Bilateral cervical vagotomy and intraperitoneal injection of atropine methyl-bromide completely prevented the bradycardia, thus confirming the vasovagal reflex is responsible for bradycardia in the acute phase after the blast exposure (Irwin et al., 1999).

FIGURE 18.33

Representative experimental data demonstrating blast-related bradycardia in the rat model: (a) naive control, (b) rat exposed to 190-kPa peak overpressure (sampling rate 1 Hz in both cases). (c) The heart rate monitoring in the acute phase postexposure (more...)

18.3.2.3. Shock Wave–Induced Acute Lung Injury

No animals showed external signs of trauma. At necropsy (performed immediately after the blast exposure), animals subjected to a blast wave showed typical evidence of moderate pulmonary hemorrhage (Figure 18.34), associated with vascular damage, direct alveolar injury, and edema, and generally described as “blast lung” (Brown et al., 1993; Pizov et al., 1999; Sasser et al., 2006; Tsokos et al., 2003). As illustrated in Figure 18.34, animals subjected to a higher blast overpressure (190–290 kPa) were found to have statistically significant pulmonary hemorrhage compared with sham control. The onset of injury in our model takes place at 130 kPa peak overpressure. Modeling of the pulmonary injury as a function of peak overpressure with the dose-response function revealed the inflection point is located at 184.2 ± 15.7 kPa, with a plateau (the highest score) at 14.32 ± 1.85 (Figure 18.34g). This value is rather low: the maximum possible value in Yelverton’s blast injury index for lungs is 64 (Yelverton, 1996). Interestingly, pathological evaluation of pulmonary injuries in one rat from the 290-kPa group (100% acute mortality) revealed virtually no signs of lung injury. In the areas most severely affected by hemorrhage, approximately 15%–60% of the alveoli were filled with acute pools of hemorrhage (data not shown). Near these pools, there were respiratory bronchioles, which contained small amounts of edema fluid. The more normal sections of lung have moderate atelectasis and very sparse to absent emphysema.

FIGURE 18.34

Blast-induced lung injury immediately postexposure: (a) control, (b) 130 kPa, (c) 190 kPa, (d) 230 kPa, (e) 250 kPa, and (f) 290 kPa. The extent of injury was quantified using the Pathology Scoring System for Blast Injuries (Yelverton, 1996). The extent (more...)

18.3.2.4. BBB Damage and IgG Uptake by Neurons

Staining to reveal IgG in the brain parenchyma was performed for every twelfth section (each section was 40 μm thick) equally spaced across the entire rat brain (Figure 18.35). The purpose of this staining was twofold: (1) to measure the extent of BBB damage and (2) to visualize the accumulation of IgG in the intracellular space of neurons. Figure 18.35a presents coronal sections where elevated levels of IgG with respect to control samples are obvious. The slides with mounted sections were first digitized, and then we integrated the staining intensity to quantify of the amount of IgG. The olfactory bulb region (initial 12 sections, 5.76 mm from the brain anterior, see inset in Figure 18.35b) was omitted in this analysis. The optical density (OD) of each section was divided by the average OD of three control samples resulting in relative OD (ROD). Figure 18.35b presents RODs of 40 sections as a function of section number for a single rat from the 190-kPa group. At lower blast intensities (190 and 230 kPa), in some rats the ROD had a higher value in the rostral than in the dorsal brain region (Figure 18.35). The ROD gradient was not a general pattern and typically the variations were not substantial. However, in one rat exposed at 250-kPa blast intensity, the highest RODs (2.3–2.7) are noted in slide numbers 28–35 (not shown). The RODs pertaining to the same brain were averaged and expressed as average ROD. The results for rats exposed at three different blast intensities are presented in Figure 18.35c. In the 190 kPa group, two rats have higher average ROD than the control group (average ROD > 1), but two others are comparable. Similar results are noted for the 230-kPa group, but only one rat had a higher average intensity of IgG staining than the rats in the 190-kPa group. The level of BBB damage appears to be correlated with the blast intensity (Figure 18.35c). In animals with compromised BBB, numerous IgG-positive cells of neuronal morphology were noted across brain parenchyma. The IgG accumulated predominantly in the cells localized in various parts of the cerebral cortex (Figure 18.36) and the hippocampus (Figure 18.37). We have also evaluated brain sections of rats exposed to 190-, 230-, and 250-kPa blast intensity using amino cupric silver staining. These tests gave negative results, indicating 24 hours is an insufficient period to induce neuro-degeneration.

FIGURE 18.35

(a) Immunostaining for rat IgG as an indicator for a compromised BBB. (b) Relative OD of full coronal sections from the brain of a single rat exposed at 190 kPa. The OD of each section was divided by the average OD of three controls (two naive and one (more...)

FIGURE 18.36

The IgG uptake in cortical neurons 24 hours after the blast. Coronal sections of the brain were collected from brains of sham control (a) and surviving rats exposed to blast with (b) 190-kPa, (c) 230-kPa, and (d) 250-kPa peak overpressure. The scale bar (more...)

FIGURE 18.37

The IgG-positive cells in hippocampus of sham control (a) and in rats exposed to blast overpressure of 190 (b), 230 (c), and 250 kPa (d−f). The scale bar (300 μm) is the same for all samples. (From Skotak, M. et al., Journal of Neurotrauma (more...)

18.4.1. Development of Human Head Model

The 3D human head model was generated from segmentation of high-resolution MRI data obtained from the Visible Human Project (2009). The MRI data consisted of 192 T1-weighted slices of 256 × 256 pixels taken at 1-mm intervals in a human male head. The image data are segmented into four different tissue types: (1) skin, (2) skull, (3) subarachnoid space (Sasser et al., 2006) and (4) brain (for which material properties are available). Brain included all important sections: frontal, parietal, temporal, and occipital lobes, cerebrum, cerebellum, corpus callosum, thalamus, midbrain, and brain stem. It was not possible to separately segment cerebrospinal fluid and structures such as membranes and bridging veins because of the resolution of the MRI data; as such, they are considered a part of the subarachnoid space (SAS). The segmentation uses 3D image analysis algorithms (voxel recognition algorithms) implemented in Avizo. The segmented 3D head model was imported into the meshing software HyperMesh and was meshed as a triangulated surface mesh. The volume mesh was generated from this surface mesh to generate 10-noded tetrahedrons. Tetrahedron meshing algorithms are more robust than hexahedral meshing algorithms and can model complex head volumes like brain and SAS faster and easier (Baker, 2005; Bourdin et al., 2007; Schneiders, 2000). The modified quadratic tetrahedral element (C3D10M) available in Abaqus is very robust and is as good as hexahedral elements (Abaqus user’s manual) as far as accuracy of results is concerned (Cifuentes and Kalbag, 1992; Ramos and Simões, 2006; Wieding et al., 2012). In addition, hexahedral elements can suffer from the problem of volumetric locking for highly incompressible materials like brain. The problem of volumetric locking is not present for a modified quadratic tetrahedral element (C3D10M) (Abaqus user’s manual). For these reasons, we chose the modified quadratic tetrahedral element. The use of specialized 3D image processing (Avizo) and meshing software (HyperMesh) allowed for the development of a geometrically accurate finite element (FE) model. FE discretization is schematically shown in Figure 18.38.

FIGURE 18.38

FE discretization.

18.4.2. Development of Animal Head Model

A very similar approach as described for the human head model was used for developing a 3D rat head model. For development of this model, MRI/CTs of a Sprague Dawley rat was used. Two different T2-weighted MRI scans (one for the muscle skin and other for the brain), and one CT scan (for the skull and the bones) were used. These three different scans were necessary to achieve proper contrast and segmentation of various tissues (i.e., muscle, skin, brain and skull, and bones). The brain MRI has an isotropic resolution of 256 × 256 × 256 pixels, for a field of view of 30 mm in all three directions. The MRI for muscle and skin has an anisotropic resolution, with a pixel size of 512 × 512 × 256, for a field of view of 30, 30, and 50 mm. The three data sets were overlapped, registered, segmented, and triangulated using Avizo 6.2 software. The triangulated mesh (i.e., surface mesh) is imported into HyperMesh and a volume mesh is generated from this surface mesh to generate 10-noded tetrahedrons (C3D10M). The skull, skin, and brain share the node across the interface. These elements are treated as Lagrangian elements.

18.4.3. General Theory of Material Models and Material Parameters Used in the Head Models

In both human and animal models, skin and skull are modeled as linear, elastic, isotropic materials with properties adopted from the literature. Similarly, SAS in the human model is considered elastic. In general, elastic properties are sufficient to capture the wave propagation characteristics for these tissue types and this approach is consistent with other published works (Chafi et al., 2010; Chen and Ostoja-Starzewski, 2010; Grujicic et al., 2011; Moore et al., 2009; Moss et al., 2009; Nyein et al., 2010). For elastic material, stress is related to strain as

where σ is a Cauchy stress, E is Green strain (also known as Green-Lagrange strain), λ and μ are Lame constants, and δ is a Kronecker delta.

The brain is modeled with an elastic volumetric response and viscoelastic shear response. Viscoelastic response is modeled using the Maxwell model. The associated Cauchy stress is computed through

where F is a deformation gradient, J is a Jacobian, and S is the second Piola-Kirchhoff stress, which is estimated using following integral:

where E is the Green strain and G_ijkl is the tensorial stress relaxation function. The relaxation modulus for an isotropic material can be represented using a Prony series:

where G_∞ is the long-term modulus and β is the decay constant.

18.4.3.1. Material Models for Human Head Model

For material parameters of the brain tissue, the widely accepted bulk modulus value of 2.19 GPa is used in this work. This value is motivated from the works of Stalnaker (Stalnaker, 1969) and McElhaney (McElhaney et al., 1973). The shear properties of the brain tissue are adopted from Zhang et al. (Zhang et al., 2001), who derived the shear modulus from the experimental work of Shuck and Advani (Shuck and Advani, 1972) on human white and gray matter. For material parameters, we relied on widely accepted values in the literature for base simulations. In addition, parametric studies are conducted to account for reported variations in the brain material properties. The material properties of the human head model are summarized in Table 18.4.

TABLE 18.4

Material Properties

18.4.3.2. Material Models for Rat Head Model

Skin and skull are modeled as a homogenous linear elastic isotropic material with properties adopted from the literature (Willinger et al., 1999). Brain tissue is modeled as elastic volumetric response and viscoelastic shear response with properties adopted from previous work (Zhang et al., 2001). Air is modeled as an ideal gas equation of state (EOS). The Mach number of the shock front calculated from our experiments is approximately 1.4 and hence the ideal gas EOS assumption is acceptable; the ratio of specific heats does not change drastically at this Mach number value. The material properties along with longitudinal wave speeds are summarized in Table 18.5.

TABLE 18.5

Material Properties for Rat Head

18.4.4. Computational Framework for Blast Simulations

18.4.4.1. Numerical Approach for Shock Tube

In this section, a detailed description of the computational framework for blast simulations using the Euler-Lagrangian coupling method is discussed. In this method, an Eulerian mesh is used to model shock wave propagation inside the shock tube and a Lagrangian mesh is used for the head. This computational environment allows accurate concurrent simulations of the formation and propagation of the blast wave in air, the fluid–structure interactions between the blast wave and the head models, and the stress wave propagation within the head. The computational framework is shown in Figure 18.39.

FIGURE 18.39

Computational framework for blast simulations.

The shock tube used in the modeling is based on our experimental shock tube. The head is placed inside the shock tube and subjected to the blast wave profile of interest. The Eulerian domain (air inside the shock tube) is meshed with eight-noded brick elements, with appropriate mesh refinement near the regions of solid bodies to capture fluid– structure interaction effects. Parametric studies on mesh size have been performed and it has been found that mesh size of 3 mm is appropriate to capture the flow field around the head (i.e., pressures, velocities) and fluid–structure interaction effects. For Eulerian elements, mesh convergence is achieved at this element size; thus an element size of 3 mm is used near the regions of solid bodies and along the direction of blast wave propagation. Air is modeled as an ideal gas EOS (see Equation 18.7) with the following parameters: density, 1.1607 kg/m³; gas constant, 287.05 J/(kg-K); and temperature 27°C:

where P is the pressure, γ is the constant-pressure to constant-volume specific heat ratio (= 1.4 for air), ρ₀ is the initial air mass density, ρ is the current mass density, and e is the internal volumetric energy density. The Mach number of the shock front from our experiments is approximately 1.4, and hence the ideal gas EOS assumption is acceptable because the ratio of specific heats do not change drastically at this Mach number.

18.4.5. Solution Scheme

The finite element model is solved using the nonlinear transient dynamic procedure with the Euler-Lagrangian coupling method (Abaqus). In this procedure, the governing partial differential equations for the conservation of momentum, mass and energy (Equations 18.8 through 18.10) along with the material constitutive equations (Equations 18.3 through 18.6) and the equations defining the initial and boundary conditions are solved simultaneously.

Conservation of mass (continuity equation):

Conservation of momentum (equation of motion):

Conservation of energy (energy equation):

where ρ is a density; x, v, and a are displacement, velocity, and acceleration of a particle, respectively; σ is a Cauchy stress; b is a body force; e is an internal energy per unit mass; q is a heat flow per unit area; and q_s is a rate of heat input per unit mass by external sources.

In the Eulerian–Lagrangian method, we are actually solving the whole model (i.e., both the Eulerian and Lagrangian domains) with the same Lagrangian equations. The notion of a material (solid or fluid) is introduced when specific constitutive assumptions are made. The choice of a constitutive law for a solid or a fluid reduces the equation of motion appropriately (e.g., compressible Navier–Stokes equation, Euler equations). For the Eulerian part/domain in the model, the results are simply mapped back to the original mesh with extensions to allow multiple materials and to support the Eulerian transport phase for Eulerian elements. The Eulerian framework allows for the modeling of highly dynamic events (e.g., shock) which would otherwise induce heavy mesh distortion. In Abaqus, the Eulerian time incrementation algorithm is based on an operator split of the governing equations, resulting in a traditional Lagrangian phase followed by an Eulerian, or transport, phase. This formulation is known as “Lagrange-plus-remap.” During the Lagrangian phase of the time increment, nodes are assumed to be temporarily fixed within the material, and elements deform with the material. During the Eulerian phase of the time increment, deformation is suspended, elements with significant deformation are automatically remeshed, and the corresponding material flow between neighboring elements is computed. As material flows through an Eulerian mesh, state variables are transferred between elements by advection. Second-order advection is used in the current analysis. The Lagrangian (solid) body can be a deformable body and can deform based on the forces acting on it; the deformation of the Lagrangian solid influences the Eulerian part/domain.

A typical 3D simulation requires about 7 hours of CPU time on 48 dedicated Opteron parallel processors (processor speed 2.2 GHz, 2 GB memory per processor), for an integration time of 2.5 msec. The simulation time is selected such that the peaks resulting from stress wave action have been established. A time step of the order of 5 × 10^–7 seconds is used to resolve and capture wave disturbances of the order of 1 MHz, which increases the overall computational effort for the total simulation time of interest.

18.4.6. Model Validations

In this section, both human and animal models will be validated against the experimental blast results. This is a necessary step in developing a finite element model to show that the model mimics the reality to as close as possible.

18.4.6.1. Validation of the MRI-Based Human Head Model against Blast Experiments

The head model is validated against PMHS experiments, which were previously discussed in this chapter. PMHS specimens (N = 3) were subjected to primary blasts of incident intensities 70, 140, and 200 kPa, respectively. The numerical model is validated against surface (reflected) pressures, surface/skull strains, and ICPs obtained from these experiments. In the PMHS experiments, dura, subarachnoidal spaces, and brain were removed and intracranial contents were backfilled with ballistic gel whose wave speed is calculated at 1,583 ± 118 m/s, which was close to the longitudinal wave speed of water. Thus for head model validation purposes only, SAS and brain were assigned the same bulk modulus value of 2.19 GPa. The validation results are presented for incident intensity of 200 kPa, but similar agreement in the simulation and experimental results are seen for other intensities. The arrival time of the experiment at each sensor location is shifted to match the arrival time of the numerical simulation for ease of comparison of the different features of the p-t profile. The experimental profiles are based on the average of three shots (experiments) for head 1.

Figure 18.40a shows a comparison of the surface pressure profile for the front location. For the front location (i.e., FH SM), good agreement is seen between the experiment and the simulation both in terms of peak pressure and nonlinear decay. Figure 18.40b shows the comparison of ICP profiles between the experiment and the simulation. Both experimental and simulation data are filtered using a 10-kHz four-pole Butterworth filter. From the figure, it can be seen that there is a reasonably good agreement between the experiment and numerical ICP profiles both in terms of peak values (maximum% difference in peak ICP value is 17%) and shape of the profiles. In general, simulation ICP profiles show more oscillatory behavior than experimental ICP profiles. This can be attributed to the following: (1) lack of material characterization for ballistic gel. It is possible that the response of ballistic gel to shock loading is much more complex (because of heterogeneity, rate dependent behavior, effect of curing, wave propagation, and dispersion in 3D setting) than assumed here and (2) frequency response of ICP pressure probes (gauges) embedded in the ballistic gel is not known; it is possible that frequency response of the gauge is slower because of which behavior is less oscillatory in the experiments. Because of these artifacts, discrepancies in the oscillatory pattern are to be expected between the simulation and the experiment. It is also common that the repeated experiments will show small variations in the oscillatory pulse patterns from shot to shot. For these reasons, it is improbable if not impossible to match every aspect of experiment with the computational simulation. With these considerations in mind, some aspects of comparison between experimental and simulation ICP profiles are discussed.

FIGURE 18.40

(a) Comparison of surface pressure profile between experiment and simulation for FH location. (b) Comparison of ICP profiles between experiment and simulation.

For FH ICP, simulation is able to capture major trends including the initial sharp rise (rise time = 40 µs) associated with the shock front, initial decay (till t = 0.15 msec), abrupt pressure increase (second peak) during initial decay (at t = 0.07 msec), and subsequent pressure pattern. The secondary (loading) pulse seen in the experiment after 1.0 ms is not seen in the simulation profile for FH ICP. For nose ICP, fair agreement is seen between the experiment and the simulation. The distinct secondary peak seen during the decay (at t = 0.25 msec) is delayed in the simulation. This secondary peak is due to wave transmission from the eye socket. The delay in this transmission between simulation and experiment is attributed to the difference in geometry of the eye socket between the simulation and the experiments. The center ICP also shows fair agreement between the simulation and experiment. The rise time is sharper in the simulation than the experiments. In addition, the peak ICP value is slightly higher (20% difference) in the simulation. These discrepancies are expected considering the complex set of direct and indirect loadings experienced by the center location. The center location experiences a complex set of direct and indirect loadings emanating from different sources (e.g., blast wave transmission, reflections from tissue interfaces, skull deformation) at different points of time. These disturbances continuously propagate into the brain as waves. The constructive and deconstructive interferences of these waves control the pressure history deep inside the brain. We believe the obtained match between experiment and simulation for the center ICP is reasonable considering these complexities. The back ICP also shows reasonably good agreement between experiment and simulation. The contrecoup phase seen in the experiments is replicated in the simulation.

During PMHS experiments (circumferential) skull strains were also measured at four locations, front, temple, top, and back, as shown in figure. Table 18.6 compares the peak skull strains obtained from the simulation with that of experimentally measured skull strains at various locations. The maximum peak strain is seen at the front location and the strain value at this location is less than 0.1%. It should be noted that the standard deviations in experimentally obtained skull strain values are huge. Standard deviations up to 100% of mean values are seen in certain strain measurements. Such standard deviations are normal during strain measurements, especially when strain values are very small. Obtained strain values from the simulations fall within the range of experimentally obtained strain values. The experimental and numerical strain–time profiles are not compared, as the obtained profiles were not highly repeatable during the experiments.

TABLE 18.6

Comparison of Peak Circumferential Strains at Various Skull Locations

18.4.6.2. Validation of the MRI-Based Rat Head Model against Blast Experiments

The rat head model is validated against rat head experiments, were which previously discussed. In this experiment, five male Sprague Dawley rats of 320–360 g weight were sacrificed and instrumented with a pressure sensor on the nose to measure reflected pressure, and two additional sensors were implanted in the thoracic cavity and brain, respectively. To test the robustness of the model it was validated against experiments both inside and outside of the shock tube. Input loading is simulated using partial model technique in Section 18.4.4.2.

Figure 18.41a, b shows comparisons of p-t profiles for the surface mount sensor on the nose inside and outside of the shock tube, respectively. Overpressure and the subsequent decay recorded within 1 ms is slightly lower in the case of simulation in the simulation is slightly lower than that of the overpressure recorded in the experiment. Beyond 1 ms there is an almost perfect match between experiment and simulation. This deviation in the initial part of the pressure profile is due to a coarse Eulerian mesh relative to the size of the Lagrangian body (rat head). A tradeoff was made to get reasonably accurate results for higher model efficiency (time for model simulation) and this did not affect the intracranial measurement drastically, which will be seen in the following section.

FIGURE 18.41

Comparison between experiments and numerical models both inside and outside the shock tube. (a) Surface pressure measured on the nose. (b) Surface pressure measured on the nose. (c) ICP inside the brain. (d) ICP inside the brain. (From Sundaramurthy, (more...)

Figure 18.41c, d show the ICP comparisons between experiment and simulation for inside and outside, respectively. There is a good agreement between experiment and simulation. Both experimental and simulation data are filtered using a 10-kHz four-pole Butterworth filter. In general, the simulation inside has less oscillations while outside has more oscillations after 0.5 msec. This model also predicts the effect of exit rarefaction depleting a overpressure and resulting in a low-pressure, high-velocity wind (after 0.5 msec).

18.5.1. Flow Field on the Surface of the Head with the Helmet

Figure 18.42 shows the pressure-time history on the surface of the head with and without the helmets. In general, rise time and time to peak is increased and rate of pressure decay is decreased with the helmet. Peak pressures are reduced with the padded helmet at all locations compared with the no-helmet case. For the suspension helmet, peak pressure is reduced at the incident blast site (sensor FH); on the contrary, peak pressures are increased on the side away from the incident blast side compared with the no-helmet case. The positive phase impulse either remained equivalent or increased with the suspension and padded helmets as time to peak is increased and rate of pressure decay is decreased with the helmets.

FIGURE 18.42

Pressure-time history on the surface of the head with and without the helmets.

18.5.2. Comparison of Experiments and Numerical Simulations for Helmeted Cases

Numerical simulations are used to understand the mechanics of the flow field around the head with the helmet. Before using the numerical simulations for this purpose, numerical results are compared and validated against helmeteds experiments. Figure 18.43a, b shows a comparison of surface pressures on the surrogate head with suspension and padded helmets, respectively. For the suspension helmet, there is a reasonably good agreement between the experiment and numerical simulation, in terms of peak pressures (maximum difference 26%, minimum difference 0.3%) and nonlinear decay. The simulation is able to capture the majority of the features well, including the arrival of the blast wave at a given location, shock front rise time, and underwash (explained in detail in the next section) beneath the helmet. For the padded helmet, fair agreement is obtained between the experiment and simulation. The huge difference is seen in the values of peak pressure and total impulse. This is because it is very difficult if not impossible to know precise placement of the padded helmet on the surrogate head in the experiments. The blast wave can enter through small gaps, if any, created during the mounting of padded helmet on the surrogate head. In addition, porosity of the foam pads is not modeled in the simulations, which may contribute to the surface pressures.

FIGURE 18.43

Comparison of surface pressures from experiments and numerical simulations on the realistic explosive dummy (RED) head with (a) suspension and (b) padded helmet respectively.

18.5.3. Underwash Effect of the Helmet

As indicated in previous sections, surface pressures are increased under the suspension helmet on the side away from the incident blast side. This is due to the underwash effect of the helmet. This effect is illustrated using numerical simulations (Figure 18.44). The blast front after encountering the head–helmet assembly divides into two fronts: one front travels around the outer perimeter of the helmet whereas the other front penetrates the gap between the head and the helmet (i.e., head–helmet subspace) and travels underneath the helmet toward the back of the head as shown in Figure 18.44a. The blast front traveling outside the helmet reaches the rear of the helmet before the blast front traversing through the gap (Figure 18.44b-i), and eventually when these two blast fronts meet they focus at a region on the back side of the head (Figure 18.44b-ii). This focusing produces higher pressures on the head, away from the incident blast side when the location is shielded by the helmet. After this high pressure is generated, the high pressure air in the head-helmet subspace expands in all directions (Figure 18.44b-iii).

FIGURE 18.44

Underwash effect of the suspension helmet: (a) schematic explaining underwash effect of the helmet; (b) flow field inside and outside of the head–helmet subspace.

To understand how the underwash influences both the local peak pressure and the impulse, it is postulated that the pressure intensification depends on the shape of the helmet (curvature) and the head–helmet subspace gap size with respect to the oncoming pressure wave and its characteristics (e.g., pressure, velocity, rise/fall time). These aspects are studied in the following section using simplified two-dimensional head models. It should be noted that local peak pressures in the head-helmet subspace and impulse transmitted to the head are analyzed because these quantities determine the effective load on the head.

18.5.3.1. Effect of Curvature, Head–Helmet Gap Size, and Incident Peak Pressure Intensity

To examine the effect of geometry, three different cases are considered. In the first case, the head and the helmet are modeled as cylinders, in the second case the head is cylindrical and the helmet flat, and in the third case both the helmet and the head are flat (Figure 18.45a). In all these cases, there is constant gap of 13 mm between the helmet and the head. Figure 18.45b-i and b-ii shows the pressure and impulse profiles at the back of the head–helmet subspace where the focusing occurs. It is clear from Figure 18.45b that the pressure and impulse are increased when both the shapes are cylindrical in comparison with the other two cases. This trend is the same when the incident overpressure is increased from 0.18 MPa to 0.52 MPa.

FIGURE 18.45

Effect of curvature of the helmet and the head. (a) Modeling setup for studying curvature effect of the helmet and the head. (b-i) Average pressure in the back region of the head−helmet subspace and (ii) total impulse transmitted to the back region (more...)

Having identified that the cylindrical case offers the most severe loading conditions, this case is used to study the effect of head–helmet gap size and incident peak pressure intensity on the underwash. Figure 18.45c shows the P_max/P* (normalized peak maximum overpressure) in the head–helmet subspace as a function of gap size for different incident peak pressure intensities P*. As the gap is reduced, pressure in the gap increases (P α 1/V, V-volume). Thus, the P_max/P* ratio increases as the gap size is reduced until a certain critical gap size. Thereafter, the boundary effects become dominant and P_max/P* ratio decreases because of these boundary effects. The P_max/P* ratio increases as incident peak pressure intensity P* increases. Numerical simulations indicate that for the ranges tested, the angle θ at which P_max occurs is between 140 and 155°.

Another quantity of interest is the transmitted impulse, I, which depends on the maximum peak pressure, P_max and rate of pressure decay (i.e., rate of expansion) once P_max is established. The higher P_max and lower the rate of pressure decay, the higher is the impulse transmitted. As shown earlier, P_max increases as the gap size is reduced until a critical gap size. The rate of pressure decay, however, decreases continuously (no critical gap size) as the gap size is decreased as shown in Figure 18.46a. This is because as the gap size is reduced, there is not enough space for expansion and boundary reflection effects become dominant. Similar observations are reported by Rafaels et al. (2010) from their blast experiments on a helmeted head. From our simulations it was found that, for a given incident peak pressure intensity P*, rate of pressure decay contributes more to impulse transmitted to the head than P_max. Hence, for a given incident peak pressure intensity P*, impulse transmitted to the head continuously increases as the gap size is reduced as shown in Figure 18.46b.

FIGURE 18.46

(a) Rate of pressure decay in head−helmet subspace. (b) Impulse transmitted to the head as a function of gap size for different incident blast intensities P*. (From Ganpule, S. et al., Computer Methods in Biomechanics and Biomedical Engineering (more...)

18.5.4. Effect of Orientation on Blast Wave–Head Interactions with and without Head Protection

As mentioned earlier, to study the effect of orientation on blast wave–head interactions experiments were conducted on the surrogate head with four different orientations to the blast. These orientations are front, back, side, and 45°. The experiments were conducted with and without helmets and each scenario was repeated three times. Blast wave–head interactions are studied by monitoring surface pressures on the surrogate head and experimental observations are elucidated with the help of validated numerical models. The role of head orientation is studied by understanding the mechanics of the blast wave head interactions for the no helmet, suspension helmet, and padded helmet cases.

18.5.4.1. No–Helmet Case

Figure 18.47a shows the experimentally measured peak pressures for the no-helmet case for each head orientation. Pressure at the incident blast site is amplified (Λ = p_R / p_I) by 2.40, 2.79, 2.38, and 1.39 times the incident pressure for front, back, side, and 45° orientations, respectively, because of aerodynamic effects. This amplification factor (Λ) is based on the mean values of the incident and reflected pressures for each orientation.

FIGURE 18.47

Peak pressures recorded by the sensors for various orientations: (a) no helmet, (b) suspension helmet, and (c) padded helmet. (Experiment).

For front orientation, pressure gradually decreases from sensor FH to T1 to T2. Sensors T2 and T3 record equivalent pressures. There is a slight increase in pressure from sensor T3 to sensor RH, which is located on the side opposite to the incident blast side. A similar trend is observed for back orientation but in reverse order (i.e., from sensor RH to sensor FH). For these orientations, sensor R records pressure equivalent (within ±2%) to sensor T2. For the side orientation,a sensor R (i.e., the sensor facing the blast) records the highest pressure; and all sensors in the midsagittal plane record equivalent pressures. Sensor L (i.e. the sensor opposite to the blast side) records marginal pressure. For the 45° orientation, a trend similar to the front orientation is observed, but the flow reunion takes place near sensor T3 (as indicated by increase in pressure) because of tilt.

18.5.5. Total Impulse around the Head for Various Head–Helmet Configurations

The total impulse (positive phase, I⁺) is obtained by integrating pressure over time (≡P dt). The total impulse shows similar trends as peak surface pressures for all head helmet configurations and for all orientations. The total impulse plots are not shown for brevity.

18.5.6. Results from Numerical Simulations

Numerical simulations are conducted to understand and explain some of the experimental observations. Figure 18.48 shows the flow (pressure) field around the head for various head orientations for the no-helmet case. The flow field around the head is complex. Orientation of the head to the blast wave governs the flow mechanics around the head.

FIGURE 18.48

Flow mechanics around the head. (a) Flow separation on the top and sides of the head for front orientation. (b) Flow reunion on the back of the head for front orientation. (c) Flow separation along the midsagittal plane for side orientation. (d) Flow (more...)

Figure 18.49 shows the pressure field in helmet–head subspace (at the incident blast site) for each orientation for suspension helmet case. The lowest pressures in the helmet–head subspace at the incident blast site are observed for the back orientation.

FIGURE 18.49

Pressure contours in the helmet–head subspace at the incident blast site for each orientation. The lowest pressures in the helmet–head subspace at the incident blast site are observed for the back orientation because of the shock wave (more...)

Figure 18.50 shows pressure field in the helmet–head subspace (away from the incident blast site) for each orientation for the suspension helmet case. From the pressure field it can be seen that pressures are increased under the suspension helmet on the side away from the incident blast side. This also confirms the presence of the underwash effect for back and side orientations. Varying degrees of pressure intensification are observed depending upon the orientation of the head and the helmet to the blast wave. The simulation results are consistent with experimental observations.

FIGURE 18.50

Pressure intensification on the side away from the incident blast side for the suspension helmet. Varying degrees of intensification are observed for various orientations from the geometric effects that govern the flow field within the head−helmet (more...)

18.5.7. Discussion on Orientation-Dependent Blast Wave–Head Interactions

The present results validate both the hypotheses postulated: (1) the external pressure field on the surface of the head depends on whether the wearer has a suspension or padded helmet or no helmet and (2) orientation of the head to the blast wave governs the pressure field experienced by the head, for a given head–helmet configuration. In this section, the results are discussed in the context of these hypotheses.

The blast wave–head interactions are quite complex, as evident from the surface pressure patterns and the values of Λ at the incident blast site for various orientations for various head–helmet configurations (Figure 18.47). For the no-helmet case (Figure 18.47a), statistically similar amplification at the incident blast site is observed for the front and side orientations (p = 0.81) and a higher amplification for the back orientation (p_max = 0.019). The amplification for 45° orientation is lower (p_max = 0.006) because of flow separation at the face and because a sensor is not present at the exact incident site due to the 45° tilt (Figure 18.48d). The Λ depends on the incident blast intensity, the angle of incidence, the mass and the geometry of the target, and the boundary conditions, and can vary from 2 to 8 (Anderson, 2001; Ganpule et al., 2011). By geometry, we imply geometrical features, such as topology and area of exposure. At the plane of specimen blast wave interaction, the different geometrical features have different effects. For the suspension and padded helmet cases (Figure 18.47b, c), Λ at the incident blast site for each orientation is statistically different. For these cases, Λ is governed by the geometry of the helmet, the head–helmet configuration, and its orientation to the blast.

To better understand surface pressure patterns and hence the flow fields around the head, numerical simulations are used. First, the flow field around the head for the no helmet case is presented. Once the blast wave impinges the head, flow separation occurs, as is evident from the values of the recorded pressure for the sensor next to the incident blast site (Figure 18.47a). For example, pressure reduction of 53.89%, 25%, 67.91%, and 43.94% are observed for front, back, side, and 45° orientations respectively for the sensor next to the incident blast site. Flow separation causes low-pressure zones (e.g., top and sides of the head [Figure 18.48a]); thus, pressures are further reduced as we move away from the incident blast side (Figure 18.48a). The blast wave traversing the head and blast wave traversing the neck reunite on the side opposite to the incident blast side (Figure 18.48b). This flow reunion causes an increase in pressure (e.g., sensor RH for front orientation, sensor FH for back orientation, sensor T3 for 45° orientation) (Figure 18.47a). For the side orientation, flow separation occurs before the blast wave reaches the midsagittal plane (Figure 18.48c), thus all the sensors on the mid-sagittal plane record similar pressures (Figure 18.48a). This flow separation is further enhanced as the blast wave reaches the side opposite to the incident blast side; hence, the corresponding sensor (sensor L) records very low pressure. Flow separation for the side orientation is attributed to a larger area facing the blast. Numerical simulations clearly show that the surface pressures and the flow field around the head are strongly governed by the geometry of the head. Several other studies (Chavko et al., 2010; Ganpule et al., 2011; Mott DR, 2008; Taylor and Ford, 2009; Zhu et al., 2012) have also shown that the geometry of the head plays an important role in the blast wave head interactions and hence in the biomechanical loading of the brain.

Figure 18.51 shows the peak pressure plots for no-helmet, suspension helmet, and padded helmet cases superimposed on each other for each head orientation. Table 18.7 shows the percentage reduction in peak pressures at the incident blast site for the suspension and padded helmet cases compared with the no-helmet case. By comparing the values of peak pressures (Figure 18.51 and Table 18.7) at the incident blast site, it can be seen that varying degrees of pressure reduction at the incident blast site are observed for suspension and padded cases compared with the no-helmet case. Back and side orientations show statistically significant reduction (p < 0.05) in pressure under the helmet, but only marginal reduction is seen under the helmet for front orientation (p > 0.05). For front orientation, part of the oncoming blast wave contributes to the pressure as the helmet does not cover the forehead completely (Figure 18.49). A 45° orientation does not show reduction in peak pressure at the incident blast side (i.e., sensor FH). In contrast, for 45° orientation, peak pressures are increased by 95.13% and 73.94% for the suspension and padded helmet cases respectively compared with the no-helmet case. This is because (1) there is flow separation at the face for the no helmet case (Figure 18.48) and (2) in contrast to flow separation for the no-helmet case, the blast wave is directed to the head–helmet subspace for the suspension and padded helmentt cases (Figure 18.49). Of all orientations, the maximum reduction in pressure (65.18% and 77.98%, respectively, for suspension and padded helmet cases) at the incident blast site compared with no-helmet case is observed for the back orientation (Figure 18.51 and Table 18.7). The helmet has a larger area and height on the back than the front or the side. Thus the helmet diffracts and blocks the oncoming blast wave, offering maximum protection as shown in Figure 18.49. Zhang and Makwana (2011) have also found maximum reduction in peak ICP for the back orientation from their numerical simulations.

FIGURE 18.51

Peak pressures for no helmet, suspension helmet, and padded helmet cases superimposed on each other for various orientations: (a) front, (b) back, (c) side, and (d) 45°.

TABLE 18.7

Percent Reduction in Peak Pressures at Incident Blast Site for Suspension and Padded Helmet Cases as Compared with a No-Helmet Case

Pressures are increased under the suspension helmet on the side away from the incident blast side (Figure 18.51). This is due to the underwash effect of the helmet. For the orientations studied, the maximum underwash (i.e., pressure intensification) under the suspension helmet is observed for the 45° orientation followed by the front orientation (see sensor RH, Figure 18.51). This is mainly for two reasons: (1) orientation of the head–helmet configuration to the blast and (2) the geometry of the helmet. For the 45° orientation, the blast wave penetrates the head–helmet subspace (i.e., gap) more effectively from both sides (Figure 18.50). The blast wave is continuously directed in the head–helmet subspace from the face because of tilt. As mentioned earlier, the back of the helmet has a larger area and height than the front. Thus, for the front orientation, the blast wave traversing the neck and blast wave traversing outside the helmet, after reaching the head–helmet subspace, engulfs the head–helmet subspace (due to geometric effects) and causes higher intensification. For other orientations, this engulfment is less intense (Figure 18.50) because of the shorter height of the helmet in the corresponding regions. It should be noted that, for front and 45° scenarios, the maximum surface pressure recorded by the surrogate head with the suspension helmet (P_max = 0.76 and 0.87 for front and 45° orientations, respectively) exceeds the maximum surface pressure recorded in the no-helmet case (P_max = 0.56 and 0.34 for front and 45° orientations, respectively).

The underwash effect is not seen for the padded helmet case, as evident from Figure 18.51. However, equivalent pressures are seen on the top region of the head (sensors T1–T3) compared with the no-helmet case (Figure 18.51). This indicates that additional pathways/modes of energy transfer exist under the padded helmet. Thus, performance of the foam pads under the blast-loading conditions needs further investigation to identify these pathways/modes.

18.5.8. Role of Helmet in Mechanics of the Blast Wave–Head Interactions: Effect on ICP Response

In published work (Ganpule et al., 2011; Ropper, 2011), the role of the helmet in the mechanics of the blast wave–head interactions has been studied in terms of pressure field experienced on the surface of the surrogate dummy head. From the previous study done on surrogates, it was shown that peak surface pressures were reduced at the incident blast site (i.e., site closest to impact) for both padded and suspension helmet configurations and these reductions were statistically significant. At other locations, surface pressures were marginally reduced for the padded helmet configuration and increased for the suspension helmet configuration. Increases in surface pressures under the suspension helmet were due to focusing of the blast wave under the suspension helmet. Impulse values were either equivalent or were marginally reduced for both padded and suspension helmet configurations. In addition, for the suspension helmet, a statistically significant increase in impulse was seen for locations that were directly below the focused region. How the external pressure field measured on the surface of the head translates to the intracranial contents is currently not known. Some of the interesting questions include the following: What dose increase in surface pressures under a suspension helmet imply in terms of brain injury? Does blast wave focusing affect injury outcome and injury severity? In this section, an attempt is made to address these questions by studying ICP response with and without helmets.

For these experiments, the cadaver heads used in Section 18.2 were used with the same array of instrumentation used for the bare head tests. Padded helmets were used for heads 1 and 2 and a suspension helmet was used for head 3. Figure 18.52 shows the comparison of peak ICPs between no-helmet and helmeted cases for FH, nose, and center ICP. Peak ICP values are reduced with the padded helmet for FH and nose locations compared with the no-helmet case; these reductions are statistically significant for most of the cases (p < 0.05). Center and back (not shown) locations show equivalent peak ICP values with the padded helmet, compared with the no-helmet case. With the suspension helmet, peak ICP value were increased (statistically significant increase) for the FH location at incident blast intensities of 70 kPa and 140 kPa and remained equivalent for an incident blast intensity of 200 kPa. With the suspension helmet, a statistically significant reduction in peak nose ICP is seen at all intensities. Center and back locations show equivalent peak ICP values with the suspension helmet.

FIGURE 18.52

Comparison of peak ICPs between no helmet, padded helmet, and suspension helmet cases.

Figure 18.53 shows the ICP impulse comparison between the no-helmet and helmeted cases for FH, nose, and center ICP. Equivalent ICP impulse values are seen with the padded helmet at all locations compared with no helmet, but there is no fixed pattern of variation across the heads and/or intensities, making it difficult to draw any concrete conclusions. Increased ICP impulse values are seen with the suspension helmet at all locations compared with no-helmet scenarios. This increase in impulse under suspension helmet is statistically significant for most of the scenarios.

FIGURE 18.53

Comparison of peak ICPs between no helmet, padded helmet, and suspension helmet cases.

For the padded helmet, peak ICP values at the FH, nose, and temple locations are reduced and these reductions are statistically significant; center and back locations do not show significant reduction in peak ICP values. ICP impulse values are either equivalent or marginally reduced with the padded helmet (Figure 18.54a). This suggests that the face is an important pathway of load transfer to the brain. The area of the face is ~25%–30% that of the area of the forehead. Thus, a significant amount of impulse/energy is transferred through the face; this is further confirmed using numerical simulations. The face has been identified as an important pathway of load transfer by Nyein et al. (2010) and a face shield has been proposed for blast mitigation; peak pressure reductions up to 80% were proposed with the face shield. However, some of the major limitations of Nyein et al.’s study were total simulation time and military/field relevance of their input data. The authors simulated the entire blast event for 0.76 msec only; this time period is not sufficient to truly evaluate the role of a face shield in blast mitigation as a face shield delays the blast wave transmission into the intracranial cavity. In addition, incident blast pressures and durations used as an input for numerical simulations were not representative of a realistic blast. In addition to the role the face plays in blast wave transmission, marginal impulse reduction seen with the padded helmet from our experiments also highlight the role that structure/geometry of the head plays in governing the ICP response.

FIGURE 18.54

Pressure time histories with and without the helmet. (a) Padded helmet at incident intensity of 200 kPa. A similar trend is seen for intensities of 70 kPa and 140 kPa. (b) Suspension helmet at incident intensity of 200 kPa. (c) Suspension helmet at incident (more...)

With the suspension helmet, peak ICP values are reduced for nose ICP, and these reductions are statistically significant. On the contrary, peak values are increased for FH ICP at incident blast intensities of 70 and 140 kPa and remained equivalent for blast intensity of 200 kPa because of the focusing/underwash effect as shown in Figure 18.54b, c. In addition, temple, center, and back locations show equivalent peak ICP values with the suspension helmet. Statistically significant increases in impulse values are seen with the suspension helmet at all locations compared with no-helmet scenarios. Thus it is clear that the focusing effect seen under the suspension helmet translates to the intracranial contents and adversely affects the ICP response (Figure 18.55). This does imply that for a suspension helmet configuration, wearing a helmet can be worse than not wearing the helmet for primary blasts in which blast waves can potentially focus under the helmet. However, this does not preclude the use of helmets that provide critical protection against blunt and penetrating conditions.

FIGURE 18.55

Focusing effect seen under the suspension helmet translates to the intracranial contents. Intracranial regions with increased pressure with respect to the no-helmet counterpart are highlighted with an ellipse.