the finite element analysis for concrete filled steel tubular columns under blast load

7/31/2019 The Finite Element Analysis for Concrete Filled Steel Tubular Columns Under Blast Load

1/6

The Finite Element Analysis for Concrete Filled Steel TubularColumns under Blast Load

J.H. ZHAO, X.Y. WEI AND S.F. MA

School of Civil Engineering, Changan University, Xian 710061, China

1. Introduction

The responses of blast load are always taken into consideration for the significant building

and protective construction. Presently, concrete filled steel tubular (CFST) is widely used in

construction because it has the beneficial qualities of both concrete and steel. In order to

study the mechanical behavior of the CFST column under blast load, the dynamic responses

of a square CFST column under surface explosion were simulated by the nonlinear finite

element program ANSYS/LS-DYNA. The JHC model was used for concrete material and

the MAT_PLASTIC_KINEMATIC m odel which a ccounted for the strain r ate used for steel.

The failure behavior of the CFST column at scale distance equal to 1.0 was analyzed. The

results indicate that the inner concrete was seriously damaged, however, the deformation of

concrete was restricted by the steel tube. It shows that CFST column has excellent ductilityand blast resistance. The time-history curve of displacement of key nodes at different scale

distance are compared, which indicates that the deformation of column obviously decreases

with the increase of scale distance.

2. Numerical Simulation

2.1. Numerical model

As shown in Fig. 1, the responses of CFST column under surface blast occurred at various

stand-off distances are investigated. The clear height of the CFST column is H= 3 m. Assum-

ing the column has square cross section and the width, the depth and the thickness of the

steel tube is 500 mm, 500 mm and 10 mm, respectively. The to p a nd botto m o f the column isconsidered as fully fixed. A 3-D numerical mo del of concrete filled steel tubu lar column was

set up. Solid elements are used to model both the concrete and the steel. There are a total of

40460 elements in t he numerical model. Con vergence test is conducted and it was found that

further refinements in mesh density did not significantly improve global response.

2.2. Material model

The Johnson-Holmquist (J-H) material model is used for concrete. This model can be used for

concrete subjected to large strains, high strain rates, and high pressures. The equivalent stress

is expressed as a function of pressure, strain rate and damage. A more detailed description

can be found in LS-DYNA theoretical manual.8 The pa rameters of concrete used in this study

are shown in Table 1.The MAT_PLASTIC_KINEMATIC material model is used for steel. Isotropic, kinematic,

or a combination of isotrop ic and k inematic hardening may be obtained by varying a par ame-

ter between 0 and 1. For = 0 and = 1, respectively, kinematic and isotropic hardening

Corresponding author. E-mail: [email protected]

Analysis of Discontinuous Deformation: New Developments and Applications.

Edited byGuowei MA and Yingxin ZHOU. Published by Research Publishing Services.

Copyright c 2009 by Society for Rock Mechanics & Engineering Geology (Singapore).

ISBN: 978-981-08-4455-4

doi:10.3850/9789810844554-0070 669


2/6

Analysis of Discontinuous Deformation: New Developments and Applications

Figure 1. Concrete filled steel tubular column under blast.

Table 1. Concrete material parameters (g-mm-ms).

Parameter M ID RO G A B C N

Value 1 2.25E-3 1.38E4 0.75 1.65 0.007 0.76

Parameter FC T EPS0 EFM IN SFM AX PC UC

Value 40 3.92 0.001 0.01 7 13.33 7.3E4

Parameter PL UL D1 D2 K1 K2 K3

Value 800 0.1 0.038 1 1.74E4 3 .8 8 E4 2 .9 8 E4

Table 2. Steel tube material parameters (g-mm-ms).

Parameter M ID RO E PR SIGY ETAN BETA SRC SRP FS VP

Value 2 7.85E-3 2.1E5 0.3 345 1180 0 40.4 5 0.3 0

are obtained. Strain rates effect is accounted for using the Cowper-Symonds model which

scales the yield stress by a strain rate dependent factor. The parameters of steel used in this

study are shown in Table 2.

2.3. Blast loading model

The explosive process is not included in this study. The blast pressures are generated using

procedures outlined in TM5-1300 and the loading functions corresponding to these blast

pressures are then applied to the numerical model. TM5-1300 is widely used by blast engi-

neers for preliminary design purpo se. It adop ts the cube-root scaled distance for considering

various stand-off distances and charge weight. The scaled distance is defined a s

Z = R/W1/3 (1)

in which R is the distance from the source and W is the weight of explosives.

Figure 2 shows a free-field typical pressure-time history. At any point away from the burst,

the pressure disturbance has the shape shown in Fig. 2. The shock front arrives at a given

location at time tA and after the rise to the peak value, Ps0 the incident pressure decays to

the ambient value P0 in time to which is the positive phase duration. This is followed by

a negative phase with a duration t0

. The negative pressure has a maximum value of Ps0

.

670


3/6


Figure 2. Free-field pressure-time variation.

Usually the negative phase is less important in the design than is the positive phase. Hence,

only the positive phase of blast pressure is considered in the n umerical simulation.

The empirical pressure-time history in Ref. 2 is used herein:

P(t) = Ps0 (1 t/t0) exp ( bt/t0) (2)

in which b is the parameter of the shock wave.

The shock waves propagate with supersonic velocity and finally it hit the building. They

reflect from the building with amplified overpressures and it can be determined from TM5-

1300. Assuming the stand-off distance is 5 m, three blast scenarios are considered, i.e., the

scaled distance Z = 0.7, 1.0 and 1.3. The blast pressure is uniformly loaded on the column

surface.

3. Numerical Results

Nu merical simulations are carr ied o ut t o estimate the blast response and d amage of the CFST

column subjected to explosive blast loading based on the transient dynamic finite element

program LS-DYNA.

3.1. Results of scaled distance = 1

Figure 35 shows the deflection in X direction and maximum principal stress of concrete of

time t= 2 ms, 5 ms, 9 ms, respectively. It is observed that the maximum deflection o ccurs atthe middle of the column. It is expected because the column has symmetrical supports and

it is under uniform load. The deflection increases with time and reach its maximum value

of 117 mm when t= 9 ms. From the stress contour of the column, it can be found that the

tensile damage occur first at the top and bottom of the concrete. The maximum principal

stress reaches the tensile strength of concrete. When time increases to 9 ms, the concrete at

the middle of the column is also damaged and erosion occurs. However, the ratio between

the deflection and the height of column is 3.9% . H ence, it can be concluded that the steel

tube effectively restricted t he lateral deflection of the column and thus can impro ve the b last

resistances.

671


4/6


(a) Column deflection in x direction (b) Maximum principal stress of concrete

Figure 3. Deflection and stress oft= 2 ms.

(a) column displacement in x direction (b) maximum principal stress of concrete


3.2. Comparison of Displacement

Figures 6(a) and (b) shows the d eflection in x d irection of t he column for scaled d istance z =

0.7m/k g1/3 an d z = 1.3 m/k g1/3, respectively. It can be seen that the maximum deflections

of the column decrease significantly with increase of the scaled distances.

672


5/6


(a) column deflection in x direction (b) maximum principal stress of concrete


(a) scaled distance z=0.7m/kg

1/3(b) scaled distance z=1.3m/kg

1/3

Figure 6. Deflection in x direction.

4. Conclusion

The following conclusions are deduced from the n umerical r esults:

The Johnson-Holmquist (J-H) material model can be applied to simulate reasonably both

the compr essive crush zone a nd tensile damage.

When scaled distance is 1.0 m/kg1/3, th e ratio b etween th e deflection and the height of col-

umn is 3.9% . It can be concluded t hat the steel tube effectively restricted the latera l deflection

of the column and thus can improve the blast resistances.

The maximum deflections o f th e column decrease significantly with increase of t he scaled

distances.

673


6/6


Acknowledgements

The supports of the Fund for the Doctoral Program of Higher Education of China

(20040710001) and Shaan Xi Province Natural Scicence Foundation (SJ08E204) are grate-

fully acknowledged.

References

1. Georgios Giakoumelis, Dennis Lam. Axial capacity of circular concrete-filled-tube columns. Journal

of constructional Steel Research, 60, 2004, pp. 10491068.

2. Ben Youn g, Ehab Ellobody. Experimental investigation o f concrete- filled cold-formed high strength

stainless steel tube columns. Journal of Con structional Steel R esearch, 62, 2006, pp. 484492.

3. Zh ang, F.G. and Li, E.Z. A computat ional model for concrete subjected to large strains, high strain

rates, and high pressures. Explosion and Shock Waves. 2002, 22(3), pp. 198 202.

4. LSTC. LS-DYN A keywords manual, Version 970, Livermore Software Technology Corporation,

Livermore, C A, 2003 .

5. Wei, X.Y. Dynamic response of concrete and masonry structure und er explosive and impact loads.Reports of post PhD, 2007.

6. TM 5-130 0. Structu res to resist the effects of accidental explosions. US Army, USA, 1990 .

7. LSTC. LS-DYNA theoretical man ual, Livermore Softwar e Technology Corp orat ion, Livermor e, CA,

1998.

674

the finite element analysis for concrete filled steel tubular columns under blast load

Documents