Effect of Incident Direction of Earthquake Motion on Seismic Response of Buried Pipeline

Hyungjoo Kwak, Duhee Park, Jangguen Lee, Jae‐Mo KangJournal of the Korean Geoenvironmental Society 2015 · DOI

Abstract

In this paper, a 3D shell-spring model that can perform time history analysis of buried pipelines is used to evaluate the effect of the incident direction of the earthquake motion. When applying harmonic motions, it is shown that the period of vibration has pronounced influence on the response of buried pipelines. With decrease in the period, the curvature of the pipeline and corresponding response are shown to increase. To evaluate the effect of the incident angle, the motions are applied in the direction of the pipeline, horizontal, and vertical planes. Under motions subjected in horizontal and vertical planes at an angle of 45° from the longitudinal axis of the buried pipeline, the axial deformation is shown to contribute greatly to the response of the pipelines. When imposing two-components simultaneously, the calculated response is similar to the case where only single-component is imposed. Therefore, it is concluded that use of a single-component is sufficient for estimation of the longitudinal response of buried pipelines.

Introduction: The Seismic Vulnerability of Buried Lifelines

Buried pipelines for gas and water are critical lifelines, but their seismic behavior differs fundamentally from above-ground structures. While buildings are driven by inertial forces, buried pipes are "slaves" to the surrounding soil; their deformation is governed by ground strain and soil-structure interaction. Existing design codes often rely on simplified models, leaving a gap in understanding how the incident direction of seismic waves and multi-directional loading affect the longitudinal strain. This study utilizes a 3D shell-spring model to determine if simpler 1D analyses are sufficient for accurate risk assessment.

Methodology: 3D Shell-Spring Numerical Modeling

The authors developed a 3D shell-spring numerical model using the frame analysis program Zeus NL. Unlike traditional beam models, the pipeline is modeled as a cylindrical shell to capture local deformations. Two types of steel pipes were tested: API 5L X65 and X80.

Soil-Structure Interaction

The soil-structure interaction is represented by non-linear elasto-plastic springs in three dimensions—axial, transverse-horizontal, and transverse-vertical—following the American Lifelines Alliance (ALA) guidelines.

Non-linear soil spring models for axial, vertical, and transverse directions.

Loading Scenarios and Incident Angles

The model was subjected to both idealized harmonic motions and seven pairs of real-world earthquake records. Three primary incident cases were evaluated:

  1. LM: Waves traveling parallel to the pipe axis.
  2. HM: Waves in the horizontal plane at a 45° angle to the pipe.
  3. VM: Waves in the vertical plane at a 45° angle to the pipe.
Schematic of a buried pipeline subjected to seismic waves with incident angles ϕ and θ.

Key Insights and Results

The Period Trap

Shorter vibration periods (high frequency) are more dangerous. As the period decreases, the ground's curvature increases, forcing the pipeline into sharper bends and higher strains.

Axial Dominance

When waves hit at a 45° angle (Cases HM and VM), the resulting axial strain is the primary driver of the total response. In contrast, waves traveling parallel to the pipe (Case LM) only induce bending strain, resulting in significantly lower overall stress. Loading at a 45° angle produced normalized strains roughly 10 times higher than loading parallel to the pipe axis.

Sufficiency of Single-Component Analysis

A practical finding is that 2-component loading (applying SH and SV waves simultaneously) results in a response very similar to 1-component loading, with differences generally less than 20%. This is because the secondary component typically only adds minor bending strain, which is "drowned out" by the dominant axial strain of the primary component.

Conclusion

The study concludes that a single-component analysis that induces axial displacement is sufficient to accurately and conservatively predict the longitudinal seismic response of buried pipelines. Future work could explore localized buckling or "wrinkling" at specific points like elbows or T-junctions, as well as the behavior of segmented pipelines.

Summary updated

Abstract

이 연구에서는 사실적인 매설관의 종방향 지진 응답특성을 규명하기 위하여 3차원 시간 이력해석이 가능한 쉘-스프링 수치모델을 구축하였으며 지진파 입사방향, 2방향 재하의 영향을 평가하였다. 조화운동을 적용한 결과, 이의 주기는 큰 영향을 주는 것으로 나타났다. 이는 주기가 감소할수록 매설관의 곡률이 증가하며 응답이 증가하기 때문인 것으로 판단된다. 입사방향의 영향을 평가하기 위하여 매설관 축 방향, 수평평면, 그리고 수직평면으로 운동을 재하 하였다. 매설관 축 방향으로 하중을 재하할 경우 휨 변형만을 유발하므로 응답이 매우 작게 계산되었다. 수평 및 수직평면에서 매설관 축 방향과 <TEX>$45^{\circ}$</TEX>로 재하 하는 경우 축방향 변위가 응답에 지배적인 영향을 주는 것으로 나타났다. 또한 2방향 재하 시, 휨 변형만을 유발하는 요소의 영향이 극히 작으므로 1방향으로만 재하 하는 경우와 결과가 유사하게 계산되었다. 계측기록을 재하 하는 경우에도 응답은 유사하게 계산되었고 1방향 및 2방향 해석의 결과는 크게 다르지 않은 것으로 나타났다. 결과적으로 매설관의 종방향 지진 응답은 축 방향 변위를 유발하는 1방향 재하만으로도 충분히 정확하게 예측할 수 있을 것으로 판단된다. In this paper, a 3D shell-spring model that can perform time history analysis of buried pipelines is used to evaluate the effect of the incident direction of the earthquake motion. When applying harmonic motions, it is shown that the period of vibration has pronounced influence on the response of buried pipelines. With decrease in the period, the curvature of the pipeline and corresponding response are shown to increase. To evaluate the effect of the incident angle, the motions are applied in the direction of the pipleline, horizontal, and vertical planes. When the motion is applied parallel to the direction of the pipeline, it only induces bending strains and therefore, the response is the lowest. Under motions subjected in horizontal and vertical planes at an angle of <TEX>$45^{\circ}$</TEX> from the longitudinal axis of the buried pipeline, the axial deformation is shown to contribute greatly to the response of the pipelines. When imposing two-components simultaneously, the calculated response is similar to the case where only single-component is imposed. It is because one component only induces bending strain, resulting in very small increase in the response. The trend of the response is shown to be quite similar for recorded motions. Therefore, it is concluded that use of a single-component is sufficient for estimation of the longitudinal response of buried pipelines.
BibTeX
@article{kwak2015effect,
  author = {Hyungjoo Kwak and Duhee Park and Jangguen Lee and Jae‐Mo Kang},
  title = {{Effect of Incident Direction of Earthquake Motion on Seismic Response of Buried Pipeline}},
  journal = {Journal of the Korean Geoenvironmental Society},
  year = {2015},
  doi = {10.14481/jkges.2015.16.9.43},
}
Figure