The paper aims to theoretically and numerically investigate the confinement effect of inert materials on the detonation of insensitive high explosives. An improved shock polar theory based on the Zeldovich-von Neumann-D öring model of explosive detonation is established and can fully categorize the confinement interactions between insensitive high explosive and inert materials into six types for the inert materials with smaller sonic velocities than the Chapman-Jouguet velocity of explosive detonation. To confirm the theoretical categorization and obtain the flow details, a second-order, cell-centered Lagrangian hydrodynamic method based on the characteristic theory of the two-dimensional first-order hyperbolic partial differential equations with Ignition-Growth chemistry reaction law is proposed and can exactly numerically simulate the confinement interactions. The numerical result confirms the theoretical categorization and can further merge six types of interaction styles into five types for the inert materials with smaller sonic velocity, moreover, the numerical method can give a new type of interaction style existing a precursor wave in the confining inert material with a larger sonic velocity than the Chapman-Jouguet velocity of explosive detonation, in which a shock polar theory is invalid. The numerical method can also give the effect of inert materials on the edge angles of detonation wave front.
Insensitive high explosives (IHEs) are gaining popularity in weapon engineering due to their safety. Because of two main characteristics of IHEs: slower detonation velocity and wider chemical reaction zone, the detonation of IHEs is more easily influenced than the detonation of sensitive explosives by confinement materials. The confinement interaction of different inert materials can change the front shape and propagation speed of detonation shock wave of IHEs, and leads to the IHEs detonation transform to the non-ideal state when its front shape turns into a curve or its propagation speed decreases under the Chapman-Jouguet (CJ) velocity of the explosive detonation.
Earlier studies on the confinement effects focused on the driving capability of IHEs and the accepted work of the driven materials. With the development of numerical simulations and experimental techniques, it is possible to obtain more information on the detonation flowfield to better understand the mechanism of the confinement effect. Aslam and Bdzil [
The present paper studies the physical mechanism for the confinement effect of different inert materials on PBX-9502, a widely-applied IHE. Part 2 presents the theoretical study, which includes the establishment of an improved shock polar theory and the categorization of the confinement effect for inert materials with a sonic velocity smaller than the CJ velocity of explosive detonation. Part 3 presents the numerical simulation study, which comprises: 1) the full derivation of a second-order two-dimensional cell-centered Lagrangian hydrodynamic method with a phenomenological detonation reaction model; and 2) a detailed analysis of the confinement interaction for typical inert materials with sonic velocities smaller and larger than the CJ velocity of explosive detonation, and a representative comparison and discussion of the detonation edge angles between the improved shock polar theory and the Lagrangian numerical simulation. Part 4 presents the main conclusions.
A relatively simple shock polar theory can provide a good leading-order prediction of the confinement effect. The shock polar analysis generally utilizes a reference frame that moves along with the intersection point between the detonation shock wave and the material interface, and considers the matching relation of pressure p behind shock wave to the deflection angle θ of the streamline across the shock wave fronts of the IHEs and the inert materials.
Traditionally, detonation under the shock polar theory is described by the CJ model, which neglects the chemical reaction zone and regards the detonation as a strong discontinuity front with zero thickness [
Because the chemical reaction zone of IHEs is relatively wide, the Zeldovich-von Neumann-Döring (ZND) model is more precise for the analysis of the confinement effect as it considers the chemical reaction structure. Under the ZND detonation model, a leading shock wave and a closely following chemical reaction zone move at CJ velocity, which is the intrinsic velocity of a explosive, such that the leading shock wave and the chemical reaction zone form an indivisible whole structure called the detonation wave, and the detonation wave can self-sustainingly propagate by means of the releasing energy of the chemical reaction. When a detonation wave interacts with a confining material, its leading shock wave first refracts into the confinement material. As a result, it is more reasonable to establish a shock polar at the leading shock wave of the unreacted explosives.
makes the leading shock wave not generate any reflection wave. Therefore, the confinement interaction is determined by the polar curve of the leading shock wave within the unreacted explosive and the polar curve of the refraction shock wave within the confining material. Evidently, the main difference between the traditional shock polar theory and the improved shock polar theory lies in the difference between neglecting and considering the chemical reaction zone while adopting different detonation models. When the propagation speed of detonation shock and the equations of state (EOS) of the explosives and the confining inert materials are known, the expression of the improved shock polar curve can be analytically obtained.
For the unreacted explosives, a Jones-Wilkins-Lee (JWL)-type of EOS is generally adopted, of which the form of the shock polar curve of the leading shock wave can be expressed as follows:
{ p = b ( 1 ) a ( v ¯ ) − b ( v ¯ ) a ( 1 ) b ( 1 ) − b ( 1 ) ( 1 − v ¯ ) b ( v ¯ ) / ( 2 ρ 0 ) tg θ = ρ 0 D C J 2 p ( 1 − v ¯ ) − p 2 ρ 0 D C J 2 − p (1)
where v ¯ = v v 0 ; a ( v ¯ ) = A ( 1 − ω R 1 v ¯ ) exp ( − R 1 v ¯ ) + B ( 1 − ω R 2 v ¯ ) exp ( − R 2 v ¯ ) ;
b ( v ¯ ) = ω ρ 0 v ¯ ; ρ 0 is the standard density and v is the specific volume, and D C J
is CJ velocity of the explosives. In addition, A , B , R 1 , R 2 and ω are the constants related to the unreacted explosives [
When the detonation product expands, under the JWL-type of EOS, the form of the polar curve of the rarefaction wave can be expressed as follows:
{ d e d v ¯ = − v 0 a ( v ¯ ) − v 0 b ( v ¯ ) e d q 2 d v ¯ = ± 2 c 2 v ¯ d θ d v ¯ = ∓ c 2 ( q 2 − c 2 ) v ¯ q 2 p = A exp ( − R 1 v ¯ ) + B exp ( − R 2 v ¯ ) + C v ¯ − ( 1 + ω ) (2)
where e is the specific internal energy, q is the local velocity of detonation, and c is the sonic velocity of detonation, c 2 = v 0 v ¯ 2 A R 1 exp ( − R 1 v ¯ ) + v 0 v ¯ 2 B R 2 exp ( − R 2 v ¯ ) + v 0 C ( 1 + ω ) v ¯ − ω . The initial conditions of the above ordinary differential equations are the values of the explosives at sonic state.
For the inert materials, a p ( ρ , T ) − e ( ρ , T ) -type of EOS is generally adopted, wherein the form of the shock polar curve of the refraction shock wave can be expressed as follows:
{ e − e 0 = 1 2 ( p + p 0 ) ( v 0 − v ) tg θ = ± ( ρ 0 D ) 2 ( p − p 0 ) ( v 0 − v ) − ( p − p 0 ) 2 ρ 0 D 2 − ( p − p 0 ) p = p ( ρ , T ) = p c ( ρ ) + p n ( ρ , T ) + p e ( ρ , T ) e = e ( ρ , T ) = e c ( ρ ) + e n ( ρ , T ) + e e ( ρ , T ) (3)
where v 0 , ρ 0 , p 0 and D are the initial specific volume, initial density, initial pressure, and shock wave speed, respectively, and the three terms on the right hand side of the equality sign for p ( ρ , T ) and e ( ρ , T ) denote the cold pressure and energy, the thermal pressure and energy of ions, and the thermal pressure and energy of electrons, respectively [
Under the shock polar theory, the edge angle of detonation (the angle between the leading shock wave and the interface of the undisturbed confinement
material at their intersection point) is α = arcsin ( 1 D p − p 0 ρ 0 ( 1 − ρ 0 / ρ ) ) , which is
an important variable to characterize non-ideal state.
Shock polar theory in this paper | Traditional shock polar theory [ | Experiment [ | |
---|---|---|---|
Angle (degree) | 78.8 | 82.9 | 78.5 |
If the sonic velocity of a confinement material is smaller than the CJ velocity of the explosive detonation, the intersection point “O” in
streamline, wherein the solid line represents the polar curve of the leading shock wave of detonation and the dashed line represents the polar curve of the refraction wave for the confinement material, and the points “Ce” and “Cm” represent the sonic state of polar curve. The six types of inert materials are steel (Fe), phenolic resin, PMMA, silicone rubber foam, lithium hydride (LiH), and lithium deuteride (LiD). The six presented types of matches correspond to the six types of confinement effects.
The shock polar theory only presents the flow state near the intersection point between the leading shock wave of detonation and the interface of the confinement material, however, numerical simulations can provide more details on the entirety of the flowfield for both the detonated explosives and the confinement materials. Additionally, the shock polar theory can only analyze the confinement effect of inert materials that have smaller sonic velocities than the CJ velocity of the explosives. For the inert materials having larger sonic velocity than the CJ velocity of the explosives, a precursor wave in the confinement material may appear ahead of the detonation wave, thereby the flow near the interaction point become unsteady and the shock polar theory become invalid. Therefore, numerical simulations are essential. The confinement interaction of inert materials on IHEs can be regarded as a compressible multi-material flow problem and be suitable to adopt a second-order, cell-centered Lagrangian hydrodynamic method. It is well known that the key point of the cell-centered Lagrangian method is its determination of the velocity at the cell vertex, and the presented study obtains the velocity at the cell vertex by means of constructing a two-dimensional Riemann solver at the cell vertex based on the characteristic theory of multidimensional reactive flow equations.
The governing equations of the chemical reaction flows in the Lagrangian formulation are presented as follows:
d d t ∫ Ω ( t ) d Ω = ∫ ∂ Ω ( t ) u ⋅ d l (4.1)
d d t ∫ Ω ( t ) ρ d Ω = 0 (4.2)
d d t ∫ Ω ( t ) ρ u d Ω = − ∫ ∂ Ω ( t ) p d l (4.3)
d d t ∫ Ω ( t ) ρ E d Ω = − ∫ ∂ Ω ( t ) p u ⋅ d l (4.4)
d d t ∫ Ω ( t ) ρ λ d Ω = ∫ Ω ( t ) ρ r d Ω (4.5)
where ρ is the density, u is the velocity, p is the pressure, E is the specific total energy, E = e + u 2 / 2 , e is the specific internal energy, λ is the chemical mass fraction, r is the chemical reaction rate about the three-term ignition-growth reaction law [
Equations (4.1)-(4.5) are the conservation laws for geometry, mass, momentum, total energy, and chemical mass fraction, respectively. In particular, the geometry conservation law is essential for the Lagrangian finite volume method as it ensures the compatibility between the grid motion and the area variation of a control volume. The EOS of the JWL-type for explosives and the p ( ρ , T ) − e ( ρ , T ) -type for inert materials were adopted.
For a discretized control volume Ω c with a mass m c = ∫ Ω c ρ d Ω and an area A c = ∫ Ω c d Ω , the average value of any physical variable f can be defined as
f ¯ c = 1 m c ∫ Ω c ρ f d Ω . Therefore, Equation (4.2) becomes an algebraic equation
ρ ¯ c A c = m c , and Equations (4.1), (4.3), (4.4), and (4.5) can be expressed as follows:
d U ¯ c d t = − 1 m c ∫ ∂ Ω c H ⋅ n d l + r ¯ c (5)
where U ¯ c = ( − τ ¯ c , u ¯ c , v ¯ c , E ¯ c , λ ¯ c ) T , r ¯ c = ( 0 , 0 , 0 , 0 , r ¯ c ) T , τ ¯ c = A c / m c , n is the outward unit vector normal to the boundary surface of the control volume, and H = ( u , p , 0 , p u , 0 ) T i + ( v , 0 , p , p v , 0 ) T j is the tensor of the flux.
For any non-overlapping structured quadrilateral mesh with sides denoted by I k (k =1, 2, 3, 4), the semi-discrete finite volume discretization of Equation (5) can be written as
d U ¯ c d t = − 1 m c ∑ k = 1 4 ∫ I k H ⋅ n d l + r ¯ c (6)
Based on the trapezium rule to approximate the integration ∫ I k H ⋅ n d l about the numerical interface flux and the mathematical meaning of the semi-discrete model to describe the instantaneous behavior of a dynamics system at some initial time, a temporal-spatial second-order and implicit-explicit Runge-Kutta scheme can be used to solve Equation (6) as follows (the stiffness of chemical reaction leads to the implicit scheme to solve the chemical mass fraction equation):
U ¯ c ( ∗ ) = U ¯ c ( n ) − Δ t 2 m c ∑ r = 1 4 [ H r ( E 0 R c U ¯ c ( n ) ) + H r + 1 ( E 0 R c U ¯ c ( n ) ) ] ⋅ n r , r + 1 L r , r + 1 + Δ t 2 r ¯ c ( ∗ ) U ¯ c ( n + 1 ) = U ¯ c ( n ) + U ¯ c ( ∗ ) 2 − Δ t 4 m c ∑ r = 1 4 [ H r ( E 0 R c U ¯ c ( ∗ ) ) + H r + 1 ( E 0 R c U ¯ c ( ∗ ) ) ] ⋅ n r , r + 1 L r , r + 1 + Δ t 2 r ¯ c ( n + 1 ) (7)
where E0 is the multidimensional Riemann solver to compute the instantaneous evolution solution of a mesh vertex at time t n + = t n + 0 , namely U ¯ ( t n + ) = E 0 U ¯ ( t n ) , Rc is a reconstruction operator to transform the average value of a mesh cell to a spatially linear distribution within the mesh cell [
In the present notation, t n + denotes the infinitely small time interval in terms of the initial time tn, expressed as
t n + = t n + 0 = lim τ → 0 ( t n + τ )
where τ is a small time interval. Here, E0 may be also named as the vertex solver to generate the solution of the mesh vertex, of which the following subsection will present its expression.
The interface flux of a mesh cell is only correlated to three physical variables, namely u, v, and p from Equation (5). Therefore, the solutions of the three presented variables, specifically u, v, and p can be obtained for the vertex solver. The central idea of the vertex solver is to compute the theoretical solution along every bi-characteristic direction for a small time interval from the given initial conditions by means of the characteristic theory about the hyperbolic partial differential equations, of which certain approximation operations and limit operations can be derived to obtain the analytical solution at the infinitely small time interval [
To obtain the theoretical solution about the nonlinear hyperbolic system, a suitable linearized treatment in terms of the primitive variables is necessary to reduce the bi-characteristics to straight lines. For this purpose, it is convenient to start with the following flow equations about the primitive variables:
d q d t + A 1 ( q ) ∂ q ∂ x + A 2 ( q ) ∂ q ∂ y = r q (8)
where q = [ ρ u v p λ ] , A 1 ( q ) = [ 0 ρ 0 0 0 0 0 0 1 ρ 0 0 0 0 0 0 0 ρ c 2 0 0 0 0 0 0 0 0 ] , A 2 ( q ) = [ 0 0 ρ 0 0 0 0 0 0 0 0 0 0 1 ρ 0 0 0 ρ c 2 0 0 0 0 0 0 0 ] , and r q = [ 0 0 0 − r ∂ p ∂ λ | ρ , e r ] , wherein c is the sonic velocity under the chemical reaction.
The flow equations are linearized by freezing the Jacobian matrices about the reference state q ˜ = ( ρ , u , v , p , λ ) t = t n T at the initial point P ˜ = ( x , y , t ) t = t n . The linearized system with frozen Jacobian matrices can be written as follows:
d q d t + A 1 ( q ˜ ) ∂ q ∂ x + A 2 ( q ˜ ) ∂ q ∂ y = r q (9)
The presented special characteristic surface, namely the so-called characteristic cone or Mach cone, generated by all the bi-characteristic lines passing an evolution point P = ( x , y , t ) t = t n + τ (τ is the evolution time) is considered. This solution at the evolution point P = ( x , y , t ) t = t n + τ for Equation (9) can be given in terms of the conditions at the initial time t n .
Consideration of any unit vector denoted by n ( θ ) = ( cos θ , sin θ ) T , θ ∈ [ 0 , 2 π ] , presents a matrix pencil A ( q ˜ , θ ) = cos θ A 1 ( q ˜ , θ ) + sin θ A 2 ( q ˜ , θ ) , which has five real eigenvalues, namely λ 1 = c ˜ cos θ , λ 2 , 3 , 4 = 0 , and λ 5 = − c ˜ cos θ , and five corresponding linearly independent right eigenvectors. The eigenmatrix R is constructed by the five right eigenvectors, and the characteristic variables can be defined as w = R − 1 q .
Multiplying the system in Equation (9) by R − 1 from the left generates an eigen-system as follows:
d w d t + B 1 ( q ˜ ) ∂ w ∂ x + B 2 ( q ˜ ) ∂ w ∂ y = r w (10)
where B k ( q ˜ ) = R − 1 A k and r w = R − 1 r q , where k = 1, 2.
Equation (10) can then be transformed into the following quasi-diagonalized system:
d w d t + Λ 1 ∂ w ∂ x + Λ 2 ∂ w ∂ y = s + r w (11)
where Λ k = ( λ k , 1 , λ k , 2 , ⋯ , λ k , 5 ) (k = 1, 2) is a diagonal matrix.
Given the initial condition at time t n , the solution at the evolution point P = ( x , y , t ) t = t n + τ of the component w l (l = 1, 2, 3, 4, 5) of the characteristic variables w for Equation (11) can be obtained from the characteristic theory of the two-dimensional linear hyperbolic partial differential equations as follows:
w l ( x , y , t n + τ , θ ) = w l ( x − λ 1 , l τ , y − λ 2 , l τ , t n ) + s ^ l + r ^ w , l (12)
where s ^ l = ∫ t n t n + τ s l [ x − λ 1 , l ( t n + τ − ξ ) , y − λ 2 , l ( t n + τ − ξ ) , ξ ] d ξ and r ^ w , l = ∫ t n t n + τ r w , l [ x − λ 1 , l ( t n + τ − ξ ) , y − λ 2 , l ( t n + τ − ξ ) , ξ ] d ξ .
Therefore, the solution of Equation (9) may be obtained by multiplying Equation (12) with the right eigenmatrix R from the left, q = R w , and then integrating about θ from 0 to 2π [
u ( P ) = 1 2 u ( P ˜ ) + 1 2π ∫ 0 2π [ − p ( Q ) ρ ˜ c ˜ cos θ + u ( Q ) cos 2 θ + v ( Q ) sin θ cos θ ] d θ + 1 2π ∫ 0 2π ∫ t n t n + τ S [ z + c ˜ ( t n + τ − ξ ) n ( θ ) , ξ , θ ] cos θ d ξ d θ − 1 2 ρ ˜ ∫ t n t n + τ ∂ p ( z , ξ ) ∂ x d ξ + 1 ρ ˜ c ˜ ∫ t n t n + τ r ∂ p ( z , ξ ) ∂ λ d ξ (13)
v ( P ) = 1 2 v ( P ˜ ) + 1 2 π ∫ 0 2π [ − p ( Q ) ρ ˜ c ˜ sin θ + u ( Q ) cos θ sin θ + v ( Q ) sin 2 θ ] d θ + 1 2π ∫ 0 2π ∫ t n t n + τ S [ z + c ˜ ( t n + τ − ξ ) n ( θ ) , ξ , θ ] sin θ d ξ d θ − 1 2 ρ ˜ ∫ t n t n + τ ∂ p ( z , ξ ) ∂ y d ξ + 1 ρ ˜ c ˜ ∫ t n t n + τ r ∂ p ( z , ξ ) ∂ λ d ξ (14)
p ( P ) = 1 2π ∫ 0 2π [ p ( Q ) − ρ ˜ c ˜ u ( Q ) cos θ − ρ ˜ c ˜ v ( Q ) sin θ ] d θ − 1 2π ρ ˜ c ˜ ∫ 0 2 π ∫ t n t n + τ S [ z + c ˜ ( t n + τ − ξ ) n ( θ ) , ξ , θ ] d ξ d θ + ∫ t n t n + τ r ∂ p ( z , ξ ) ∂ λ d ξ (15)
where S ( z , t , θ ) = c ˜ [ ∂ u ( z , t , θ ) ∂ x sin 2 θ + ∂ v ( z , t , θ ) ∂ y cos 2 θ ] − c ˜ 2 [ ∂ u ( z , t , θ ) ∂ y + ∂ v ( z , t , θ ) ∂ x ] sin 2 θ , z = ( x , y ) ,
z + c ˜ ( t n + τ − ξ ) n ( θ ) = [ x + c ˜ ( t n + τ − ξ ) cos θ , y + c ˜ ( t n + τ − ξ ) sin θ ] , andQ denotes the initial position ( x + c ˜ τ cos θ , y + c ˜ τ sin θ , t ) t = t n .
For discretized grids, θ i b and θ i e are denoted as the starting and ending angles of a mesh with common vertices to update coordinates. Some approximation operations with similar procedures [
u ( t n + ) = 1 π ∑ i = 1 4 [ − p i ρ ˜ c ˜ ( sin θ i e − sin θ i b ) + u i ( θ i e − θ i b 2 + sin 2 θ i e − sin 2 θ i b 4 ) − v i cos 2 θ i e − cos 2 θ i b 4 ] (16)
v ( t n + ) = 1 π ∑ i = 1 4 [ p i ρ ˜ c ˜ ( cos θ i e − cos θ i b ) − u i cos 2 θ i e − cos 2 θ i b 4 + v i ( θ i e − θ i b 2 − sin 2 θ i e − sin 2 θ i b 4 ) ] (17)
p ( t n + ) = 1 2π ∑ i = 1 4 [ p i ( θ i e − θ i b ) − ρ ˜ c ˜ u i ( sin θ i e − sin θ i b ) + ρ ˜ c ˜ v i ( cos θ i e − cos θ i b ) ] (18)
where i is the counterclockwise numbering of the mesh cells with common vertices.
1) The steady structure of a one-dimensional planar detonation wave
The Von Neumann spike values of the PBX-9502 explosive are (in cm-g-μs units) pN = 0.375 and uN = 0.253; and the values for the CJ state are pCJ = 0.285, uCJ = 0.192, and DCJ = 0.7655. The analyzed explosive is 5.0 cm long and is initiated by the CJ condition on its left-hand side.
2) The front shape of the detonation shock wave in an explosive confined by steel
In an explosive-confiner experiment [
utilized to record the front shape of the detonation shock wave in the insensitive explosive and the front shape of the refraction shock wave in steel.
The confinement materials were characterized into two groups based on their respective compressibility, of which one group exhibited a smaller sonic velocity than the CJ velocity of the detonated explosive, which exhibited a steady flow near the intersection point between the leading shock wave of detonation and the interface of the IHE/inert material. In comparison, the other group exhibited a larger sonic velocity than the CJ velocity of the detonated explosive, which exhibited an unsteady flow near the intersection point.
To compare the above confinement materials with the aforementioned theoretical analysis, confinement materials with smaller sonic velocities were also selected, specifically steel, phenolic resin, PMMA, silicone rubber foam, LiH, and LiD. The simulation results presented in
relatively thick reaction zone width and a subsonic region in the confinement material. The subsonic flow resulted in a higher refracted shock pressure in the confinement material.
The above numerical result of the “one strong and one weak confinement” case exhibited almost identical flow structures with the “weak confinement” case. Therefore, the two types of confinement interactions may be regarded as the same type, such that the six types of confinement interactions categorized by the improved shock polar theory can be merged into five types. The flow state of the strong branch in the “one strong and one weak confinement” case is a type of unphysical solution that cannot be conducted in reality due to its higher entropy production.
The confinement material with a larger sonic velocity is selected as Beryllium (Be), whose sonic velocity is about 0.799 cm/μs at standard state.
The edge angle of detonation is a very important parameter based on the detonation shock dynamics and is generally applied to IHE engineering calculations [
A notable discrepancy between the improved shock polar theory and the numerical simulation was observed in the “one supersonic weak solution” case for LiH confinement. In this case, the match of the shock polar was observed through a Prandtl-Meyer rarefaction fan in LiH into the explosive, wherein the refraction shock wave in LiH “pulled” the leading shock wave of detonation. The leading shock front exhibited a backward curved shape for the most part and a locally forward curved shape only within a very small region near the explosive interface, in which the edge angle is obtuse. The flow structure in this case is presented in
The presented paper demonstrated the significant confinement effect of the inert materials on the flow behavior of the system for IHEs and inert materials. The following conclusions were drawn:
1) An improved shock polar theory was established on the basis of the ZND model for IHEs detonation. The theory yielded seven categories of confinement effects. Six categories were for the inert materials that exhibit a smaller sonic velocity than the CJ velocity of the explosive detonation, which have steady flow structures and can be analyzed by the improved shock polar theory, and the remaining category was for the inert material that has a larger sonic velocity than
Type of confinement | Confinement material | Edge angle of denotation (degrees) | ||
---|---|---|---|---|
Shock polar theory | Numerical | Experimental [ | ||
Strong | 2024 Aluminium | 85.3 | 84.6 | 85.1 |
HR-2 Steel | 80.9 | 80.3 | 80.6 | |
Brass | 78.8 | 78.3 | 78.6 | |
W-Mo alloy | 83.8 | 83.4 | 83.5 | |
U-Ni alloy | 82.2 | 81.7 | 81.9 | |
W-Fe-Ni alloy | 84.9 | 84.3 | 84.6 | |
Weak | Phenolic resin | 44.4 | 45.2 | 45.0 |
PMMA | 44.4 | 45.2 | 45.0 | |
“Three-weak solutions” | Silicone rubber foam | 44.4 | 45.4 | 45.1 |
“One supersonic weak solution” | LiH | 75.2 | 85.5 |
the CJ velocity of the explosive detonation, which has unsteady flow structure and cannot be analyzed by the improved shock polar theory.
2) A second-order, cell-centered Lagrangian hydrodynamic method is proposed on the basis of the characteristic theory of the two-dimensional first-order hyperbolic partial differential equations with chemical reaction law. The numerical method confirmed the theoretical categorization and presented the detailed flowfield structures.
a) For the inert materials that had a smaller sonic velocity than the CJ velocity of the explosive detonation, the six types of confinement effects divided by the improved shock polar theory might be merged into five types, and the edge angles of detonation could be cheaply and conveniently obtained from the improved shock polar theory with enough accuracy for the strong and weak confinement cases that are most frequently applied in engineering.
b) For the inert material that had a larger sonic velocity than the CJ velocity of the explosives detonation, the precursor wave was produced in the confinement material and generally exhibited a double-wave structure, including an elastic-plastic wave and a following shock wave. In addition, the refraction of the precursor wave could precompress the unreacted explosive.
This research is supported by Natural Science Foundation of China with Grant No. 11772066 and 11272064, Science Challenge Project with Grant No. TZZT2017-A1-HT001-F, and Foundation of Chinese Key Laboratory of Computational Physics with Grant No. 9140C 690101150C 69300.
Yu, M. (2018) The Confinement Effect of Inert Materials on the Detonation of Insensitive High Explosives. World Journal of Mechanics, 8, 161-181. https://doi.org/10.4236/wjm.2018.85012