You are on page 1of 6

Defence Science Journal, Vol. 59, No. 2, March 2009, pp.


Numerical Analysis on Protecting Performance of Layered
Arch Structures Subjected to Blast Loading
Dong Yongxiang1, Xia Changjing2,1, Feng Shunshan1, and Shao Zhiyu1
State Key Laboratory of Explosion Science and Technology, Beijing Institute of Technology, Beijing-100 081, China
School of Mechanics & Civil Engineering, China University of Mining & Technology, Beijing-100 083, China

Dynamic responses of layered arch structure composed of different materials subjected to blast loading
are analysed by numerical simulation. The deflection, the particle velocity and the particle acceleration of the
arch inwall and stress curves versus time are obtained comparing properties of blast resistance of different
arch structures with the same amount of charge. The results show that the arch structure composed of foam
concrete-SFRC-steel has good blast resistance. Furthermore, the dynamic performance of the foam concrete-
SFRC-steel composite structures is studied with different amount of charge. Additionally, coupling relationship
of blast resistance and explosion charge is analysed. Comparison of numerical results with experimental results,
show that they are in good agreement. This numerical analysis may provide important guidance for blast-
resistant design and analysis of underground structures.
Keywords: Layered arch structure, underground structures, blast resistance, protective performance, numerical analysis

1. INTRODUCTION arch structures as protective measures are adopted. The
Problems of protection of important facilities and simplified model is shown in Fig. 1. Strip charge is buried
dangerous geological environment always require urgent in the concrete medium which is about 25 cm away from
solutions. Particularly for underground civil and defence the apex of arch. On both sides of the computational model,
facilities, protection problems become difficult because of mass blocks are located for simulating constraints of surrounding
their depth from the earth’s surface. Layered media, especially media of the realistic underground structures. Sizes of
the sandwich material with a soft layer subjected to strong each part in Fig. 1 are shown in Table 1.
impact loading, have attracted much attention since 19701-5. Table 1. Model sizes of arch structure
But for reason of nonlinear coupling effect of the strong
impact loading with surrounding media, it is still a challenge W (cm) W2 (cm) W3 (cm) H (cm) r (cm)
to design reasonable protective structures. In the design
of underground protective facilities, it is not enough to 40 25 40 25.6 25
rely on empirical formulae as these ignore many factors
and can provide very limited information. Nevertheless, 3. NUMERICAL SIMULATION
numerical simulation is an effective tool to achieve reappearance 3.1 Computational Model and Material Parameters
of all the physical processes. 2-D and 3-D simulations of The problem is simplified as a 2-D plane strain problem
the entire process can be realised relying on the large finite to analyse blast resistance of middle cross-section which
element programme through the establishment of a
bears the largest loading. Using numerical simulation technique
comprehensive physical model, inputting the structural
provided by LS-DYNA software, 2-D plane strain computational
parameters and correct boundary conditions. More information
can be obtained from numerical results and can provide model was used in the numerical simulation. Compared
parameters for engineering design. Therefore, protective with the 3-D model, 2-D model can save a lot of computation
performance of arch structure composed of different media time and shorten the period of analysis.
near the explosion charge by numerical simulation has In the calculation, TNT explosive, C30 concrete, C60
been analysed. It may provide important guidance for blast- steel fibre-reinforced concrete (SFRC) and A3 steel plate
resistant design and analysis of underground structures. were used. The JWL equation of state was adopted to
describe the behaviour of explosion product. The Johnson-
2. PROBLEM DESCRIPTION Holmquist constitutive model (JHC) was used for plain
For the realisation of effective protection of the concrete and SFRC 6. The crushable-foam model7 was applied
underground structures close to the explosive region, different for foam materials. The main material parameters8-12are listed
Received 31 March 2008, revised 10 October 2008


N and C are material constants the Same Amount of Charge determined by experiments. m Poisson 3.03 7. 59. m l is the crushing volumetric strain (C). E 0 the internal energy per unit volume.007 0.016 0. s c is the uniaxial compressive strength.0 0. in Tables 2 to 4.4 0. R 1. Table 2.58 0. P CJ the C-J detonation pressure.45 7. modulus. Sketch of analytical model. e fmin is the amount of constant.01 0. K1.60 373. respectively.55 15.4 18.8 kg/m.0 SFRC 2. K 2 and K3 are pressure H1.0 Table 3. the thickness of layer I.80 0.05 0. 20 cm and 0.8 29.61 0.61 0. respectively.02 3. w are material constants of JWL 3.0 0. Six compound modes of layered media are shown in Table 5.2. In linear density of cylindrical TNT charge is 4. layer II.63 6930 20. R 2. f c¢ is the quasi-static uniaxial The arch structure is composed of three parts as shown compressive strength.0 E / MPa 2.8 0.0 24. mc is the crushing volumetric strain corresponding layer I interchanges with the thickness of layer II in group to the pressure P c. MARCH 2009 E x plos iv e ch arg e A rc h s tru c tu re W1 C o nc re te W2 M a ss b lo c k P L AYER I F ou n d atio n W3 L AYER II r L AYER III Figure 1.06 1. H 2 and H3 are 5 cm. D1 and D 2 are damage constants. are on behalf of different materials. II and III which is equal to s / f c¢ (s the actual equivalent stress). Main parameters of the TNT explosive ¦ Ñ(g/cm ) 3 D (m/s) pCJ /GPa A/GPa B/GPa R1 R2 ¦ Ø E0/GPa 1.1 .0 sc / MPa 6.7e+02 0.79 1.72 0.004 SFRC 0. But the G is the shear modulus.75 1. layer I.0 Material efmin PC /GPa mc K1 /GPa K2 /GPa K3 /GPa Pl /GPa m1 T/GPa Plain concrete 0.1 Comparison of Different Arch Structures with equation.10 0.2 m 0. Except that the thickness of density. DEF SCI J.001 85 -171 208 0.013 Table 4.10 0.001 34.18 0.9 0. The constants .15 0. B’.60 -- 132 . Here r and E are density and elastic damping coefficient of materials. and layer III is kept corresponding to the pressure P l. Smax is the normalised maximum strength in Fig.06 1. Material parameters of concrete foam Material properties Foam aluminum Volumetric strain Yield stress/ MPa r/ (g/cm ) 3 0.2 0.35 6.80 0. total thickness of arch structure is kept constant in the Pc = f c¢ / 3 and ml = r grain / r 0 . 2.048 7. NO. D the detonation velocity. Material constants used in calculation for plain concrete and SFRC with JHC model Material r (g/cm3) A’ B’ N C f ’c /GPa Smax G /GPa D1 D2 Plain concrete 2.45 0. A’. In different layered media. The thicknesses of the three layers plastic strain before fracture. D amp the corresponding charge mass is 2 kg. 1. where rgrain is the grain investigation reported here. respectively.6 cm. B.0 24.0 0. In Table 3. The Table 4.0 Damp 0.75 4.0 0.8 3. VOL.60 0.2 Analysis of numerical results ratio and A. T is maximum tensile hydrostatic pressure.009 0.016 0.

0 2.5 2. it stress wave gives rise to the vibration of the whole arch can be observed that the attenuation of group (E) is better structure during a certain time under blast loading.0 2.0 1 . Curves of acceleration of group(A).0 3 .0 t( m s ) t(m s) Figure 3.0 0 .0 1.0 1 . et al. 1).0 1. Figure 5. 50 400 A D 40 B E 300 C F 30 200 20 a (10 m /s ) 2 100 u(m /s) 10 4 0 0 -10 -10 0 -20 -20 0 -30 -40 -30 0 0.5 3.5 2 .5 2.5 1 . The attenuation is the smallest which indicates that concrete foam used ability of group (A) with steel plate as inner layer is better in layer I has a good effect on the wave attenuation. This result shows that though the and particle acceleration curves near the apex of arch P strength of SFRC composed of one whole SFRC layer is point (shown in Fig. 2. Figure 4.0 0. From the deflection of group (B) with group (C). 133 .5 4 . Also. Figure 3 than that of group (F) comparing the deflection of group shows that the extreme particle velocity of group (B) in 5 4 800 3 600 2 400 1 0 200 a(104m /s2) d( m m ) -1 0 -2 -20 0 -3 -4 -40 0 -5 A D -60 0 -6 B E C F -80 0 -7 -8 -10 0 0 0.0 0. Curves of deflection versus time. Curves of acceleration of group(B).5 4.0 0 .5 1 . curve of P point versus time.0 2 .5 4.5 3. Comparing than that of group (F) without the inner steel plate.5 3 .0 3.0 0 .5 3 .0 -12 0 0 0 . Compound modes of calculation model Group Layer I Layer II Layer III A Plain concrete SFRC Steel B Concrete foam SFRC Steel C SFRC Concrete foam Steel D Plain concrete SFRC SFRC E A whole SFRC arch structure F SFRC SFRC SFRC Figures 2 to 5 show the deflection. YONGXIANG.5 1.5 4 . it can be observed Fig. it can also be seen that the deflection near P point that the attenuation of explosive wave of arch structure of each arch structure exists an obvious fluctuation with concrete foam as layer I is better than that of arch phenomenon in a long computation time because local structure with concrete foam as the middle layer. The deflection of group (B) but the stiffness of arch structures is different.0 2 . Figure 2 gives the displacement the same as that of SFRC composed of three layers SFRC.5 2 . Curves of particle velocity versus time. the particle velocity (E) with group (F).5 1.: PROTECTING PERFORMANCE OF LAYERED ARCH STRUCTURES SUBJECTED TO BLAST LOADING Table 5.0 t(m s) t(m s) Figure 2.0 3.0 3 .

the protective performance obtained from numerical results and experiment. These results reveal that the images is better than the structure of group (D) without the inner of inner layer are in good agreement. 7 shows a sharp falling of steel shell and no rebound occurs again in a short time when the -8 0 amount of charge reaches a certain value for the same arch -1 00 structure.4 -6 -30 0 kg/m. 0 It shows that the falling velocity at large amount of charge is far faster than that at small amount of charge. -1 40 the critical amount of charge could be determined. COMPARISON WITH EXPERIMENTS initiation of detonation. respectively. The corresponding charge occur and then cause the fluctuation of wave.0 1. From Figure of group (B) is the best. With the 4.0 kg/m and 31.0 kg/m.0 kg/m.8 kg/m. Therefore. Figure 4 presents the acceleration curves versus time velocity and stress with the amount of charge. the inwall.0 when the linear density of the charge is 8.5 2 . steel plate.0 2 . is shown in Fig.8 cm. 134 .4 kg/ structure. 2.Q 2) and the needs of practical projects.5 3.5 4 .0 0. Velocity with different weights of charge. The t(m s ) numerical results of group (B) at different amount of charge Figure 7. The corresponding charge mass is 0. Fig. Numerical results further show that the average acceleration Figures 10(b) and (c) show the deformation of arch of group (A) is smaller than that of group (D). 59. Figure 9 gives the typical falling acceleration curve of P point 0 . was carried out for each layer under the same conditions As explosive wave propagates and interacts in the arch as experiments.5 1. 20 Figure 7 gives the falling velocity versus time of P point.2 Dynamic Responses of Arc Structure With -20 0 Different Amount of Charge -4 Q3 The blast resistant of group (B) was analysed at three -25 0 linear densities as shown in Table 5. stress peak value decreases significantly.0 0 . the shock wave reached arch inner Based on above analysis. 3. 2. So the structure by simulation and experiment with 31. Q3 -6 0 u(m/s) In addition. Comparing the extreme velocity show that average acceleration of P point increases with value curve of different arch structures.5 4. The linear density of charge is 31. and the (C) shows that the location of concrete foam has an important shoulder of arch structure are all the key positions to bear role on protective performance. Under the conditions 0 2 studied in this research. the optimisation analysis wall at about 100 m s and caused a downward acceleration. the superposition and attenuation of stress wave m and 8.2 kg. The acceleration value keep zero until the explosive wave reaches the apex of arch P point. of group (A). t(m s) Figure 6 presents the deflection curve versus time of Figure 6.3 kg.0 3 . The minimum amount of charge causing the rapid decrease of arch structure can be defined as the critical -1 20 amount of charge.5 3 . DEF SCI J. With the help of the numerical simulation. the conclusion an increase of the linear density of charge which is in is the same as that of the deflection in Fig.0 3. VOL. Figure 8 -2 0 Q1 shows that the equivalent stress of P point of steel plate -4 0 Q2 increases with the amount of charge. it can be seen that the apex. respectively. MARCH 2009 all curves is the smallest. Figure 5 mass is 13. 10(a). NO. The inner thickness of steel plate was shows the different attenuation values of arch structures. -50 whether the protective performance is better or not is determined 0 -10 0 by comprehensive consideration of structural failure criteria d(mm) (Q 3) d(mm) (Q 1.3 kg and 13. The middle-layer thickness of concrete Comparing the average acceleration of each arch structure.2 kg and 3. 8. The linear density of the cylindrical charge is 4.5 2. When stress wave propagates through concrete foam with low-wave impedance. Deflections with different weights of charge. Q2 -2 -15 0 3. the better the protective performance is.5 1 .0 2.4 kg/m linear structure of group (A) as the inner layer with steel plate density of the charge.0 2 kg.0 1 . foam was 6 cm and the thicknesses of two SFRC layers the average acceleration of group (B) is the smallest which were both 10 cm. The composition of arch structure gives the acceleration curve versus time of group (B). Figure 11 gives the equivalent stress of arch structure The above analysis shows that. P point with different amount of charge. agreement with the change trends of the above deflection. Analysis of group (B) and group 11(a). 0.2. It indicates that the deflection increases as the mass of charge increases. the earlier the explosion wave Q1 attenuates.

load distribution 5.0 kg/m linear amount of charge. (a) (b) (c) Figure 10. density of charge. YONGXIANG. the blast loading.5 1 . and damage blast resistance as the upper layer and steel plate as and cracks are easily produced in both sides of arch under an inner layer. Curves of the equivalent stress with different Figure 9.5 4 .5 2 .0 3 . Therefore dimensional deflection (the ratio of deflection of the the corresponding measures in those weak sections could apex to the height of arch) can be used as an index be taken for engineering design.0 0.5 2.5 3 .5 1. Typical stress state of arc structure by simulation and experiment. Comparison deformation of arc structure between simulation and experiment. Separation and sliding phenomena occur (1) The arch structure with concrete foam has a better between two adjacent and bounding layers.0 0 . The deflection curves of the apex. The conclusions are as follows: (3) Coupling effects between load and different layered 135 .0 2. the physical images from the numerical results. non- results are in agreement with experimental results. to characterise the degree of damage of the arch structure. et al.0 1.0 3. (a) (b) (c) Figure 11. Typical acceleration curve with 8.0 0. This provides the reference of the whole process of structure movement are revealed.0 2 . CONCLUSIONS and equivalent stress curve of arch structures and the Through numerical analysis of the deformation and key positions to bear the blast loading are obtained failure characteristics of arch structure.5 3. for design of arch structures. Numerical (2) Comparing numerical simulation and experiments.5 4.0 t(m s ) t(m s) Figure 8.: PROTECTING PERFORMANCE OF LAYERED ARCH STRUCTURES SUBJECTED TO BLAST LOADING 12 00 20 0 10 00 10 0 0 80 0 -10 0 sv-m(MPa) a(10 4m /s 2) -20 0 60 0 -30 0 40 0 -40 0 -50 0 20 0 -60 0 0 -70 0 0 .0 1 . the load in the arch structure [Figs 11(b) and (c)].

LSTC& and the structures. pp. Rev. N. J. 59. ETA. H. D. In the 3. Science and Technology.. Eng. Tedsco. high-strain rate. P. O. of Mechanics. mis-matching principle of wave impedance and the 10. & Perry. 56-60. Struct. Dynamic response of layered structures subject to blast effecfts Contributors of non-nuclear weaponry. Int. 1989. Timothy J Holmquist. SPH algorithm for projectile The authors would like to thank Prof Z... During design of arch structures. impact mat glass fibre laminates to blast loading. This work is sponsored by Open 12. Duan of penetrating into concrete. Impact on laminated composites: Recent Mr Xia Changjing obtained his PhD from the University of Science and Technology of China in 2003. L.. W. MARCH 2009 media have been verified by numerical simulation. Leonard E. 22(3/4). Quebec: American Defense Preparedness Association. Shock Waves. Mechanics of penetration and perforation. J. Beijing Institute of Technology. M. Dearborn. et al. Livermore Software Technology Corporation of Explosion Science and Technology. 27. 2001. Experimental and numerical studies Foundation of State Key Laboratory of Explosion Science of foam-filled sections. 2001.. & Landis. 8. Eng. 25. 591-600. Institute of Technology in 1982. W.. 26. she is working as an associate researcher investigation into the response of chopped strand and her research interests are in explosion dynamics. Int. Schwer. 1998. 24. Nurick. REFERENCES 1. J. Explos. Int. Beijing Institute of 4.. R. At present. Shuncheng. Hallquist John. Sci. Hayes. & Zhou. 625-38. G. R. J. Impact dynamics and explosion protection. Constitutive modelling of optimum zone bearing load among media should be aluminum nitride for large strain. and taken into consideration. G. 283-303. leader in his chosen field of learning at State Key Laboratory California. J. et al. He joined the State Key Laboratory of Explosion Computers & Geotechnology. 2. He is a Professor and the 7. J.X. J. Santosa. & Landis. S. Chinese Academy of Sciences in 2004. DEF SCI J.P.. 2000. 1994. Y. 16(11). surface cap model for geomaterial modelling: A new and flying control. mechanics and explosion dynamics. Comput. D. T. Experimental Technology. China University 6. 136 . and Technology of Beijing Institute of Technology (Grant 509-34. At present.. 2. 2002. Hao. Wilkins. In LS-DYNA Keyword user’s manual. 211-31 ACKNOWLEDGEMENT 11. for 23(1). G. (LSTC). 5.W. Abrate. W. high strain rates and high pressures. Franz. His 9. No. the School of Mechanics and Civil Engineering. J. Ma. Mech. His research interest include the explosive effect and its protection. Song. 26. Institute of Mechanics. (in Chinese) the helpful discussions. Dr (Ms) Dong Yongxiang obtained her PhD from Institute 79-86. He has joined advances.& Cook. T.. flight dynamics. Chinese Academy of Sciences. Serge. pp. H. pp. 2003. Impact Eng. Struct. NO. Michigan. In Seventh International LS- design schemes should be adopted for different media DYNA Users Conference. of Explosion Science and Technology. 16-35. et al. In Fourteenth International Symposium on Ballistics. Different LS-DYNA material type. 2002.. he is working computational constitutive model for concrete subjected as an Associate Professor and his research interest is in rock to large strains. she joined the State Key Laboratory layered systems. Johnson. KFJJ07-5). 2007. 47(11). Appl. & Yvonne D Murray.W. A of Mining and Technology in 2005. W. Int. J. Continuous research interests are in smart munitions. 517-44. 200-201.. Impact Eng. Holmquist. 1987. 1978. Comput. VOL. Tedsco. 639-67. high-pressure applications. 793-807. Mr Feng Shunshan obtained his Masters from the Beijing 1993. M. Modelling of wave Mr Shao Zhiyu graduated from the Beihang University in propagation induced by underground explosion. as a teacher. Wave propagation through same year.