Abstract
An investigation for seabed liquefaction induced by progressive water waves is vital for the protection of marine structures from damage to the structure foundations. The residual liquefaction in sedimentary seabed has been found to be of progressive nature and experiments have also demonstrated that the liquefied soil behaves as a visco-elastic-plastic material. Building on the previous research, this work develops various numerical models to re-examine the key factors which influence the progressive liquefaction processes.To investigate the effect of randomness of wave height on seabed liquefaction, ensemble modelling approach is adopted in a two-layer inviscid fluid flow model, whereby the liquefied soil is treated as inviscid heavy fluid. Probabilistic study of soil liquefaction processes indicates that the random wave-induced liquefaction depth is much larger than that corresponding to regular waves with Equivalent Wave Height. The larger liquefaction depth in random waves is shown to be due to the fact that the highest waves rather than average waves in the wave series tend to dominate the liquefaction extent. It is also shown that the time needed for liquefaction to reach the bottom of investigated domain can vary considerably in the case of random wave time series. The longer period of low waves between the large waves will delay the time for the maximum liquefaction depth to be reached within the simulation time considered. The current design practice, which is entirely based on the regular wave models, can under-estimate the liquefaction depth and lead to unsafe design. It is recommended that the evaluation of liquefaction potential due to random waves should be based on the appropriate extreme values in the wave height distribution rather than average values such as significant wave height or root-mean-square wave height.
Secondly, a two-layer viscous fluid model representing a visco-elastic-plastic liquefied soil is constructed. The upper seawater and liquefied soil were treated as viscous fluid and described by the linearized Navier-Stokes and continuity equations. Simulation results confirmed that shear stress solved from infinite seabed solution can lead to significant errors and underestimate the liquefaction depth. The viscosity of liquefied soil computed by the present model reveals a clear state change, i.e., from viso-elastic stage to visco-plastic stage, due to the increasing deformation rate of liquefied soil layer. The strain rate dependent viscosity can influence the liquefaction process relative to constant viscosity although not very strongly. Deeper liquefaction is more likely to take place in shallower water under the same wave loading. Smaller soil permeability prevents residual pore pressure dissipation and consequently enhances the liquefaction.
Finally, the two-layer viscous model is extended to a multi-layer model in order to investigate the effect of stratification of liquefied soil layer. It is found that the liquefaction depth estimated using the N-layer model is sensitive to water content, which is contrary to that predicted by the two-layer model. The continuously increasing liquefied soil density is found to overcome the numerical difficulty in achieving a convergent viscosity. The predicted liquefied soil viscosity, liquefaction depth and interface wave amplitude are all different from that predicted by constant water content model. The sensitivity of liquefaction to both wave and soil parameters are enhanced by the stratification of liquefied soil viscosity and density. The thickness of seabed is also found to affect liquefaction but the trend is not monotonic. There seems to be a critical seabed thickness, at which the effect of seabed thickness on liquefaction reverses. Below the critical thickness, the liquefaction depth is smaller due to the relatively short drainage distance in thinner seabed but beyond the critical thickness, increasing seabed thickness damps the wave energy and consequently prevents the liquefaction. Seabed liquefaction is very sensitive to the soil plastic model parameters contained in the residual pore pressure build-up equation. Therefore, a reliable procedure for quantifying these parameters is extremely important.
Date of Award | 2012 |
---|---|
Original language | English |
Sponsors | China Scholarship Council |
Supervisor | Ping Dong (Supervisor) |