Interaction between flood defences and pipelines subject to induced earthquake loads in Groningen

Interaction between flood defences and pipelines subject to induced earthquake loads in Groningen

De Greef, J.

Jonkman, S.N. (mentor)
Bardoel, J.W.S. (mentor)
Kruse, H.M.G. (mentor)
Vrouwenvelder, A.C.W.M. (mentor)
Vardon, P.J. (mentor)

Civil Engineering and Geosciences

Hydraulic Engineering

2015-07-03

In the province of Groningen in recent years strong induced earthquakes are observed, during which a lot of energy is released in the form of seismic waves. These ultimately lead to horizontal ground accelerations at the ground surface, which affect amongst other things the stability of embankments. The main question to be answered in this report is whether the presence of a pipeline imposes an additional risk to the embankment stability, as a pipeline may fail before the embankment does. For this purpose use is made of various empirical and semi-analytical methods and the outcomes of these methods are compared with the outcomes found in the finite element software PLAXIS 2D. In this report a method is proposed in which first the stability of the embankment is checked before the earthquake begins. Ground accelerations may cause loosely packed sandy soils to rearrange to a compacter state. This is only possible if the water in between the soil particles can dissipate. As the shaking of the subsoil during an earthquake is so rapid, this is not the case and the pressure of the water increases during the earthquake. This causes the effective stresses between the soil particles to decrease, thus reducing the soil’s shear strength. When the shear strength reductions are highest at the end of an earthquake, the reduced embankment stability is checked again. The gradual reduction of the embankment stability during the earthquake is determined by coupling this to the intensity of accelerations during the earthquake. By using the accelerations higher than the acceleration at which stability is no longer guaranteed, the sliding plane displacement of the embankment is calculated. By applying this method to a case study soil profile consisting of sandy soil layers it is found that the calculated displacement increases exponentially with increasing values of the peak ground acceleration. When a soil profile of clayey soil layers is considered, the calculated deformations increase linearly with increasing values of the peak ground accelerations. The response of a continuous pipeline is determined by imposing the sliding plane displacement on a spring-supported elastic beam that represents the pipeline. The springs represent the interface stiffness of a pipeline that moves relative to the soil in four-directions: axial, lateral, upward and downward. Plasticity of the pipeline is accounted for by looking at the pipeline cross-section at the intersection between the sliding plane and the pipeline and by using an iterative procedure between this cross-section and the spring-supported beam analysis. It is found that a continuous pipeline can indeed fail before the embankment does, but for failure to occur the displacements have to be rather higher, the soil surrounding the pipeline rather strong and the limit state to indicate failure rather strict. The response of a segmented pipeline is determined by stating that the sliding plane displacement is accommodated by a maximum of two pipeline segments, of which the displacements depend on the relative stiffness of the upward and downward soil-pipeline interfaces. It is found that the segmented pipelines can fail on the criteria of axial pull-out and joint rotation before the embankment fails. As the axial pull-out magnitude does not depend on the soil stiffness, it is expected that the probability of failure is higher than for the continuous pipeline at the same sliding plane displacement magnitudes. The attempt to model the pipeline in PLAXIS 2D did not succeed as there are yet limitations to how the structural element can be applied. The ground deformations for the clayey soil profile are found to be in good agreement with the hand calculations. The ground deformations of the sandy soil profile are not. Depicting what causes the differences is very difficult, but it appears that the model used to determine the soil behaviour is very sensitive to sloped ground conditions. Insight can be given in mechanisms that are neglected in the hand calculations, but properly interpreting the results requires a lot of experience.

earthquakes
macro-instability
pipelines

http://resolver.tudelft.nl/uuid:b925ef5c-6273-4a7a-a438-55b41771342d

Student theses

master thesis

(c) 2015 De Greef, J.