soil-pipe interaction of fault crossing segmented …

SOIL-PIPE INTERACTION OF FAULT CROSSING

SEGMENTED BURIED DUCTILE IRON

PIPELINES SUBJECTED TO DIP FAULTINGS

Mohammad Hossein ERAMI1, Masakatsu MIYAJIMA2 and Shougo KANEKO3

1 PhD Candidate, Earthquake Engineering Dept., Kanazawa University (Kakumamachi, Kanazawa, Ishikawa 920-1192, Japan)

E-mail: [email protected] 2 Member of JSCE, Professor, Earthquake Engineering Dept., Kanazawa University

(Kakumamachi, Kanazawa, Ishikawa 920-1192, Japan) E-mail: [email protected]

3 Research Engineer, Ductile Iron Pipe R&D Dept., KUBOTA Corporation (26, Ohamacho 2-chome, Amagasaki, Hyogo 660-0095, Japan)

E-mail: [email protected]

This study investigates the necessity of considering different soil resistance against pipeline relative movement in upward and downward directions. In this way, results of FEM analyses are verified by experimental tests on a segmented ductile iron pipeline with 93mm diameter and 15m length installed at a 60cm depth from the ground surface in the moderate dense sand backfill condition. Fault movement, totally 35cm, has three same steps occurring in reverse way and intersection angle of 60 degrees with the pipe. This study demonstrates how assuming same resistance for soil against both upward and downward relative movements of pipeline, as suggested in JGA guideline, eventuates in imprecise FEM models.

Key words: Soil-Pipeline Interaction, Reverse Fault, Experimental Test, FEM Analysis, Segmented

Buried Pipeline, Ductile Iron Pipe

1. INTRODUCTION

As human life undeniable dependence on pipeline systems, access to these facilities, even through high seismic risk areas, is inevitable. One of these seismic risks which caused significant damages on pipeline systems in previous seismic events is movement of faults crossed by them. Uninterrupted serviceability of pipeline systems which have important function in post-earthquake recovery of human societies needs comprehensive knowledge about these structures and consequently, appropriate methods for their analysis and design.

Studies on structural behavior of fault crossing continuous buried pipelines, especially those cross strike slip faults, have considerable progress and recently have led to some reasonable analytical equations which are confirmed by comprehensive numerical models, experimental tests or comparison with previous real seismic events1) 2)3).

But behavior of segmented buried pipelines, especially subjected to dip type of faulting, is not well-studied. And sufficient understanding about

this type of pipelines needs more researches4). One of the most important uncertainties, in this regard, is the function of the connection joints in integrated structural behavior of segmented pipeline5)6) which is ignored in most commonly used manuals such as “Guidelines for the seismic design of oil and gas

pipeline systems” issued by ASCE. This guideline recommends the same soil-pipe interaction relations for continuous and segmented pipelines7). Moreover, some other guidelines, such as “Manual for seismic design of gas distribution pipelines” of Japan Gas Association (JGA), disregard different resistance of soil against upward and downward relative movement of pipe and suggest the same soil-pipe interaction relations for both cases8).

Hence, in this study, behavior of a fault crossing segmented buried ductile iron pipeline subjected to reverse faulting is studied to demonstrate how ignoring different resistance of soil against pipeline relative movement in upward and downward directions, as suggested in JGA guideline, leads to incorrect analysis results.

Journal of Japan Society of Civil Engineers, Ser.A1 (Structural Engineering & Earthquake Engineering (SE/EE)), Vol. 68, No. 4, I_781-I_789, 2012.

I_781

Fig. 1 Geometry of FE model for entire studied pipeline (a), Geometry of FE model in vicinity of pipe-fault intersection point (b)

In this way, a simple equation for vertical

transverse interaction of pipeline and surrounding soil in FEM analysis is introduced and results of computer-aided analyses based on this equation are compared to actual behavior of pipeline, observed in full-scale experiments, conducted by Kubota Corporation, Japan.

2. MODEL DESCRIPTION

In this study, a segmented ductile iron pipeline,

nominally Φ75, with external diameter of 93mm and

7.5mm wall thickness in total length of 15m composed by nine 1m length segments between two 3m length ones, at both ends, is considered. Pipeline is buried at the depth of 60cm in the moderate sandy soil with unit weight of γ=17.7kN/m

3, sub-grade reaction of k=40800kN/m3. Table 1 indicates the components of movements in horizontal and vertical directions for considered faulting which has three same steps occurring in reverse way and intersection angle of 60 degrees with pipeline as shown in Fig 1. Table 1. Components of fault movement in studied test cases

Step No. Test Case Fault movement in horizontal direction

(cm)

Fault movement in vertical direction

(cm) First Step D-1 5.75 10

Second Step D-2 11.5 20 Third Step D-3 17.25 30

Modeling details of this problem in both computer aided simulation and experimental tests are as follows. (1) FEM modeling

Using recently developed progressive methods for pipeline modeling such as Discrete Element Method (DEM) or procedures using shell or combined shell-beam elements as pipe bar have undeniable advantages. Though, such methods are

useful for fully detailed observation of pipe internal attempts in critical status like during or after buckling and performance of joints through failure or post failure stages4)9).

As mentioned previously, in this study, we focus on reasonably accurate numerical evaluation of soil transverse resistance against segmented buried pipeline relative vertical movement at two sides of a dip fault. Hence, for analysis of models, we use FEM software, namely DYNA2E10), which basically is developed for analysis of framed structures but can be used in this study by introducing pipe as beam elements connected to springs as surrounding soil. a) Modeling of Pipe

Fig. 1 shows the geometry adopted for the proposed finite element model. A 15m long straight pipeline is considered for the analysis. The fault is assumed to cross the pipeline at the center of its length.

The pipeline is modeled by using linear beam elements. The entire 15m length of the model is divided into five regions. Element size is kept uniform within each region. Region 3, including the fault crossing point has a total length of 1m (0.5m on either side of the fault crossing point). The smallest element size of 1cm is adopted in this region. That is, according to “Seismic Guidelines for Water Pipelines”

11) published by American Lifelines Alliance and in order to get adequately accurate and converged results, the length of elements is decided less than one-tenth of pipe diameter in vicinity of intersection point of pipe and fault.. The element size is increased to 2cm in regions 2 and 4, which begin at the ends of region 3 on both sides and extend up to 5m. Regions 1 and 5 represent the rest of the length of the model on both sides and they have elements of size of 5cm (half of the pipe diameter). b) Modeling of soil-pipe interaction

As shown in Fig. 2, the soil surrounding the pipeline is modeled by pairs of springs having axial stiffness only with one end attached to the pipe body


I_782

Fig. 2 Soil equivalent springs in FE model

and the other end fixed. The first spring is perpendicular to the pipe and represents the transverse interaction between pipe and soil; the other is tangential to the pipe and represents the sliding interaction of pipe with surrounding soil. Input displacements are applied at the fixed ends of springs to simulate fault displacements. Each spring is active in both directions of relative displacement between pipe and soil. The direction of all soil springs are modified at each step of solution procedure to preserve the angle they made with the pipe in the initial configuration. Properties of soil-equivalent springs used in finite element analysis of this study are shown in Fig. 3 and 4.

Parameter Fixed Side Subsiding Side

K1 334.3 kN/m 334.3 kN/m K2 0.3343 kN/m 0.3343 kN/m δ1 0.002 m 0.002 m

Fig. 3 Stiffness diagram for soil frictional springs

Parameter Fixed Side Subsiding Side

Kt1 3794.4 kN/m 3794.4 / (1~5) kN/m Kt2 37.944 kN/m 37.944 / (1~5) kN/m δt1 0.005 m 0.005 m

Fig. 4 Stiffness diagram for soil transverse springs

Basis of stiffness for tangential spring is frictional force in soil and pipe interface and evaluated according to equation recommended in guideline of Japan Ductile Iron Association (JDPA)12) as follows:

)1( / L × H × × D ×× =K 1axial

Where, “Kaxial” is axial stiffness of soil equivalent

spring (kN/m), “μ” is soil-pipe surface interaction coefficient, “D” is external diameter of pipe (m), “γ”

is soil unit weight (kN/m3), “H” is pipe burial depth

from ground surface to center of pipe (m), “L” is

pipe element length (m) and “δ1” is achieved by

experiment and equals to 0.002 (m)13). In terms of transverse springs, as resistance of

soil against upward and downward relative movement of segmented buried pipeline is different14) so different stiffness equation is needed for each direction. Hence, herein, to evaluate the resistance of soil against downward relative movement of segmented buried pipeline a simple and logical equation is introduced. That is, soil stiffness is considered equal to sub-grade reaction times to interfaced area of soil and pipe projection in horizontal plane, so we have:

(2) × D × =K ltrasnversa LK

Where, “Ktransverse” is transverse stiffness of

soil-equivalent spring (kN/m), “K” is value of

sub-grade reaction (kN/m3), “D” is external

diameter of pipe (m) and “L” is length of pipe

element (m). Furthermore, for soil resistance against upward

relative movement of pipeline, applicability of using the reduced soil resistance against downward relative movement of pipeline as its resistance in upward direction is investigated. This method is previously used by Ivanov et al. (2004) for PVC pipes in which, for FE modeling, downward stiffness of soil-equivalent springs is multiplied to a reduction factor and used as upward resistance of simulated soil against pipeline relative movement15). c) Modeling of joints

A joint is introduced in the pipe model by specifying two elements instead of one at nodes where joints are presented. The forces and moments arising from the relative displacements and rotations of two elements are computed according to the constitutive behavior as shown in Fig. 5 and obtained by laboratory tests13).


I_783

Parameter Both Sides of Fault

Ka 980 kN/m Kb 9.80 kN/m Kc 980 kN/m δa 0.001 m δb 0.05 m


Kra 12.7 kN.m/rad Krb 108.7 kN.m/rad θa 0.0855 rad


Ks 196000 kN/m

Fig. 5 Diagrams of Constitutive Behavior of Joint

Fig. 6 Layout of experimental test Facility (a), Strain gauges arrangement (b)

3. EXPERIMENTAL TESTS

Fig. 6(a) shows an elevation view of conducted experimental test. As shown in this figure, the segmented pipeline is buried in a box filled with soil which is introduced in model description. Both ends of the pipe are firmly fixed to a steel frame attached to the box, restricting both translations and rotations.

The right-hand side of the box subsides 35cm in fault trace direction (in intersection angle of 60 degrees with the pipe) at three same steps as indicated in Table 1. Strain gauges are arranged along the pipe segments at locations depicted in Fig.

6(b). The used connection joints are earthquake resistant joints of NS-Type with constitutive behavior as shown in Fig. 5 13).


I_784

Fig. 7 Results of FEM analysis and experiment for test case of (D-1)



4. VERIFICATION AND DISCUSSION

In order to find the appropriate reduction factor to

reduce soil downward resistance to use as its upward stiffness in computer-aided simulations, comprehensive numerical models are made assuming this factor varies from one to five and results of these analyses are compared to findings of full-scale experimental tests.

With the purpose of using measured parameters

during experiments, with no need to convert them to terms of pressure or force, comparisons are done on transverse displacement of pipeline in vertical plane and angular deflection at connection joints. The vertical displacement of segments are directly obtained through measurement of strain gauges and the angular deflection at a connection joint is calculated based on relative rotation of two adjacent segments connected to each other by the joint. Fig.

7(a) and 7(b), respectively, show the results of FEM


I_785

Table 2. Deviation of results in FEM and physical models for test case of (D-1)

Reduction Factor

Deviation in Vertical Displacement of Joints (cm)

Deviation in Angular Deflection (deg)

J3 J4 J5 J6 J7 J8 J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 3 0.00 -0.11 0.24 -1.49 0.00 0.1 -0.03 0.11 0.04 -0.49 -0.35 1.80 -1.25 0.00 0.07 -0.05 4 0.00 -0.13 0.31 -0.37 -0.03 0.07 -0.03 0.11 0.02 -0.43 0.18 0.57 -0.60 0.01 0.05 -0.04 5 0.00 -0.14 0.34 0.66 0.04 0.03 -0.03 0.11 0.01 -0.41 0.71 -0.52 -0.13 0.10 0.01 -0.04


Reduction Factor



J3 J4 J5 J6 J7 J8 J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 3 -0.01 -0.05 0.70 -1.16 -0.47 0.24 -0.02 0.16 0.75 -0.38 -0.06 1.04 -0.44 -0.16 0.16 -0.12 4 0.00 -0.11 0.89 0.18 -0.25 0.15 -0.03 0.16 0.70 -0.22 0.36 -0.24 0.03 0.06 0.10 -0.11 5 0.01 -0.15 1.00 1.00 0.01 0.06 -0.03 0.17 0.67 -0.13 0.62 -0.95 0.13 0.31 0.04 -0.11


Reduction Factor



J3 J4 J5 J6 J7 J8 J1 J2 J3 J4 J5 J6 J7 J8 J9 J10 3 -0.08 0.13 1.14 -1.11 -2.12 -0.18 -0.13 0.21 0.59 0.37 -0.51 1.10 0.94 -0.76 0.86 -0.17 4 -0.06 0.00 1.61 1.21 -0.91 -0.40 -0.13 0.22 0.49 0.75 0.05 -0.55 0.79 0.16 0.74 -0.18 5 -0.04 -0.08 1.82 2.74 0.08 -0.44 -0.13 0.24 0.44 0.93 0.53 -1.60 0.58 0.74 0.76 -0.21

analyses and experimental test for first step of adopted fault movement (indicated in Table 1). While, similar information for second and third steps of fault movement is shown in Fig. 8 and 9, respectively. Note that, in these figures, when reduction factor is one, stiffness of soil in upward and downward directions are assumed equal and as this factor goes up to five the stiffness of soil in upward direction is divided to this factor and reaches to 20 percent of its initial value.

These figures reveal that for higher reduction factor, lower resistance of surrounding soil against pipeline upward relative movement, longer portion of pipeline contributes in its accommodation to ground deformation induced by fault movement. In other words, with increase in reduction factor, pipeline integrated response overspreads to more number of pipeline elements located at subsiding side of fault.

That is, for all three steps of fault movement, increase in reduction factor results in more upward movement of pipe6 and pipe7 (shown in Fig. 6(a)) and larger angular deflection at J5 but smaller one at J6 (shown in Fig. 1(a)).

In addition, for more detailed comparison between results of computer-aided analyses and experimental tests, deviation of results in physical and FEM modeling for all three test cases (steps) and reduction factors of 3, 4 and 5 are indicated in Table 2 to 4. Comparing these resulted deviations reveal that minimum variant and closest correspondence in terms of both pipeline vertical displacement and angular deflection at joints relates to the case in which soil stiffness in upward direction is considered one-fourth of downward one.

Hence, hereafter we focus on comparison of pipeline behavior in cases with reduction factor equal to one and four to investigate how increase in reduction factor changes the magnitude of internal attempts in pipeline elements.

Fig. 10 to 12, respectively, depict the results of FEM analysis and experimental test for first, second and third step of adopted fault movement. According to these figures, for larger reduction factor, upward force (resistance) applied on pipeline in fixed side of fault will be compensated by low rated downward soil force (resistance) but in longer portion of pipeline in subsiding side of fault. This fact in numerical result diagrams is obvious as all internal force and moment diagrams are shifted to right by decreasing the stiffness of soil in upward direction to one-fourth of its initial value and their overspread in positive direction of abscissa axis at subsiding side of fault.

Moreover, it can be realized, for case with equal soil stiffness in upward and downward directions, pipe acts as beam with both ends fixed and has almost symmetric reactions in terms of produced transverse force and bending moment in whole pipeline and connection joints. While, for the case with reduced stiffness of soil in upward direction, behavior of pipe changes to fixed-pined beam and its reaction is no longer symmetric and the fixed side of fault works as fixed end of beam with larger restriction against rotation. So, pipeline elements including the segments and join connections have to carry larger moments and transverse force in its vicinity.

As shown in f Fig. 10 to 12, ratio of produced internal attempts regarding bending moment in fixed


I_786

Fig. 10 Internal attempts in pipeline for reduction factors of 1 and 4 in test case of (D-1)


side of fault are one and a half time of corresponding values in subsiding side and this ratio is fixed for all three steps of fault movement.

Results of this study suggested that ignoring reduction of soil stiffness in upward direction for

shallow depths, often used for burial of pipelines, in addition to incorrect knowledge on interaction of pipe and surrounding soil, causes overestimated analyses and uneconomical designs for one of most massively developing structures.


I_787


5. CONCLUSIONS

This study investigates the necessity of considering soil different stiffness in upward and downward directions. In this way, a simple equation is introduced to estimate the transverse vertical resisting force of soil on segmented buried pipelines crossing dip faults by considering different behavior of soil in two sides of subsidence trace. Numerical evaluation of this resisting force is based on simple concept of equality of soil stiffness in downward direction to sub-grade reaction times to interfaced area of soil and pipe projection in horizontal plane and multiplied to an appropriate reduction factor as its upward stiffness.

To find appropriate value as reduction factor, comprehensive numerical modeling has been done with varying reduction factor from one to five and results of computer-aided analyses are compared to findings of experimental tests.

Furthermore, it is realized that reduction in upward stiffness of soil may reduce it to fixed-pinned beam with relative larger moments and transverse force in pipeline elements in vicinity of stiffer soil wall side of fault. While for the case with equal soil stiffness in upward and downward directions, pipe acts as beam with both ends fixed and has almost symmetric reactions in terms of induced transverse force and bending moment in whole pipeline and connection joints. In terms of pipeline behavior, this study suggests that ignoring

reduction of soil stiffness in upward direction, as suggested in “Manual for seismic design of gas

distribution pipelines” by Japan Gas Association, causes incorrect knowledge on soil-pipe interaction and overestimated design of pipelines.

However, since the equation used herein for numerical evaluation of soil stiffness depends on sub-grade reaction, to confirm the consistency of specified numbers as results of this study such as soil reduction factor, it is needed to repeat the analyses and experimental tests for other soils with different values of sub-grade reaction. ACKNOWLEDGMENT: The cooperation of Ductile Iron Pipe R&D Department of Kubota Corporation, Japan in providing experimental data for Ductile Iron pipes and joints is highly appreciated. This research would not have been possible without it.

REFERENCES 1) Trifonov, O.V. and Cherniy, V.P.: A semi-analytical

approach to a nonlinear stress-strain analysis of buried steel pipelines crossing active faults, Soil Dynamic Earthquake

Engineering, pp. 1-11. 2010. 2) Vazouras, P., Karamanos, S.A. and Dakoulas, P.: Finite

element analysis of buried steel pipelines under strike-slip fault displacements, Soil Dynamic Earthquake Engineering,

30(2010), pp. 1361-1376, 2010.


I_788

3) Vougioukas, E.A. and Carydis, P.G.: Performance of buried pipelines, subjected to fault movement, Tenth World

Conference on Earthquake Engineering, pp. 5449-5457, 1992.

4) Erami, M.H., Miyajima, M. and Fallahi, A.: A State–of–the–art Review on Fault Crossing Pipeline Analysis, Paper No. 10-2228, Sixth International

Conference on Seismology and Earthquake Engineering,

Iran, 2011. 5) Kuwata, Y., Takada, S. and Ivanov, R.: DEM response

analysis of buried pipelines crossing faults and proposal for a simplified method to estimate allowable fault displacements, JSEE, Vol. 8, No. 4, , pp. 195-202, 2007.

6) Ivanov, R. and Takada, S.: Assessment of the vulnerability of jointed ductile iron pipeline crossing active faults, Japan

Society of Civil Engineering, Journal of Earthquake

Engineering, , pp. 1-6, 2003. 7) Guidelines for the seismic design of oil and gas pipeline

systems, Committee on Gas and Liquid Fuel Lifelines, American Society of Civil Engineers, 1984.

8) Manual for seismic design of gas distribution pipelines,

Japan Gas Association, 1982. 9) Liu, A., Takada, S. and Hu, Y.: A shell model with an

equivalent boundary for buried pipelines under the fault movement, Paper No. 613, Thirteenth World Conference on

Earthquake Engineering, pp. 1-8, 2004. 10) DYNA2E User’s Manual, Version 7.2.9, CRC Solutions,

2007. 11) Seismic Guidelines for Water Pipelines, American Lifelines

Alliance, p. 76, 2005. 12) Design manual for joints of ductile iron pipe type NS, S II

and S, Japan Ductile Iron Pipe Association, pp. 17-42, 2010 (In Japanese).

13) Experimental study on mechanical behavior of earthquake resistant joint of NS-Type, Kubota Co. ltd, 2011.

14) O’Rourke, M.J. and Liu, X.: Response of Buried Pipeline Subject to Earthquake Effects, New York, MCEER, 1999.

15) Ivanov, R., Takada, S. and Morita, N.: Analytical assessment of the vulnerability of underground jointed PVC pipelines to fault displacements, Paper No. 613, Thirteenth

World Conference on Earthquake Engineering, 2004.

(Received Dec. 16, 2011

Revised Mar. 3, 2012

Accepted Mar. 6, 2012)


I_789

soil-pipe interaction of fault crossing segmented …

Documents