Evaluation of seismic induced liquefaction and related effects on dynamic behaviour of anchored quay walls

Evaluation of seismic induced liquefaction and related effects on dynamic behaviour of anchored quay walls

van Elsäcker, W.A.

Jonkman, S.N. (mentor)
de Gijt, J.G. (mentor)
Brinkgreve, R.B.J. (mentor)
Besseling, F. (mentor)

Civil Engineering and Geosciences

Hydraulic Engineering

2016-07-26

Liquefaction may have large influence on performance of anchored quay walls and is important to include in seismic design. Design codes provide evaluation methods that provide hardly any insight in the actual development of excess pore pressures and the don’t consider soil structure interaction. There is a need for tools that include these aspects, since performance based design principles are more often adopted in earth- quake engineering. In this research effects of excess pore pressures on dynamic earth pressures and dynamic behaviour of an anchored are investigated. Different evaluation methods are analysed on their performance. Initially prescribed pseudo-static methods in design codes are applied, followed by a dynamic analysis using finite element model PLAXIS. After a broad literature overview of available pseudo-static methods and sensitivity analysis. A case study in Akita Port is adopted where two similar anchored quay walls were both hit by the Nihonkai Chubu Earthquake in 1983. One quay wall survived the seismic event, the other suffered severe damage. Damage was related to occurrence of liquefaction in the backfill. Pseudo-static- and dynamic analyses are applied to assess both structures. Performance of both analyses is evaluated and effects of liquefaction are investigated. Pseudo-static analysis of the quay wall without liquefaction are, after reduction of the seismic coefficient, rea- sonably in correspondence with observations. Active earth pressures are well predicted, while passive earth pressures are overestimated. The relatively simple method is however suitable to give a first indication of failure/non-failure of the structure. On the other hand the pseudo-static method is not capable to realisti- cally include effects of excess pore pressures. Large stiffness differences between layers lead to exaggerated bending moments and displacements, again passive resistance is overestimated. Since Mononobe-Okabe is a limit equilibrium method, it is not well able to capture progressive failure. By performing a dynamic analysis, using the finite element model PLAXIS, insight is obtained in the actual soil structure interaction. The Hardening Soil small strain (HSsmall) constitutive material model is adopted to model the static and dynamic behaviour of soil not vulnerable to liquefaction. A calibrated dynamic cal- culation lead to reasonable accurate for the case without liquefaction. Bending moment distribution and displacements are in line with observations. Different earthquake motions lead to relatively large spread in displacement of the structure, less differences are found for bending moment distribution. To include effects of liquefaction the user-defined UBC3D-PLM constitutive material model is assigned to potential liquefiable layers. Liquefaction resistance depends on the type of soil and on the stress state of the soil. By calibrating the UBC3D-PLM model a reasonably accurately prediction of the onset of liquefaction can be obtained for the considered stress state. Performance of the model deteriorates for variation of the stress state, especially for initial static shear stresses. Therefore a calibration method is developed to get around this limitation. Initially zones with a specific stress state are identified around the structure. Subsequently the model is calibrated to accurately predict the onset of liquefaction for each considered stress state, leading to a calibrated model parameter set for each zone. After adopting all parametersets for each zone a calibrated dynamic calculation is performed of the system, leading to reasonably accurate results of the development of excess pore pressures around the structure. Calculated displacements and bending moments are in line with observations in the field and the failure mode is well predicted. Insight is gained in the development of excess pore pressures in time at different locations. Leading to the conclusion that the UBC3D-PLM model is capable to reasonable accurately evaluate liquefaction and corresponding effects on an anchored quay wall. Finally link is made between the followed procedure and applicability of this procedure for performing a dy- namic analysis including liquefaction for one of the anchored quay walls in the Eemshaven in Groningen. Important aspects that are typical for Groningen are the location of the bedrock and corresponding transla- tion of the earthquake motion and characteristics of the motion.

Liquefaction
Dynamic analysis
Plaxis
UBC3D-PLM
HSsmall
Anchored quay walls

http://resolver.tudelft.nl/uuid:57172707-229a-4f4d-8f9a-70c96c61dfa2

Student theses

master thesis

(c) 2016 van Elsäcker, W.A.