Geotechnical Reliability of Flood Defences

Geotechnical reliability calculations explicitly take into account how soil uncertainties influence the performance of flood defences. Soil uncertainties are usually relatively large and traditionally not explicitly and properly taken into account in the design of flood defences (where loads used to be the main driver for design). The recent flood disaster in New Orleans and local dike failure in Wilnis emphasize the importance of taking the reliability of flood defences into account. Soil uncertainties can be split into inherent uncertainties and modelling uncertainties. Inherent uncertainties stem from natural variability in soils (i.e. continues uncertainties and potential anomalies). Modelling uncertainties are the result of describing the real world with models and could be due to measurement uncertainty, transformation uncertainty, statistical uncertainty and physical model uncertainty. Reliability analysis can be used to assess how the different uncertainties affect he flood defences performance. Main advantages of the reliability analysis of flood defences are e.g. cost-optimal designs, insight in which uncertainties contribute most to the failure probability, effectiveness of measurements, and cost-effective solutions.

Breached floodwall in New Orleans due to hurricane Katrina (by IPET, 2006)

Recent projects

1. The weakest link – Spatial variability in the piping failure mechanism of dikes
Ph.D. thesis by Wim Kanning, defended on December 14, 2012

Piping is an important failure mechanism of flood defense structures. A dike fails due to piping when a head difference causes first the uplift of an inland blanket layer, and subsequently soil erosion due to a ground water flow. Spatial variability of subsoil parameters causes the probability of piping failure to increase, often to unacceptable levels. The general research question is: How can we incorporate spatial variability in a flood defense system design dealing with the piping failure mechanism? The question in solved in three steps: first by quantifying the spatial variability in subsoil parameters, second by assessing the influence of this spatial variability on the piping mechanism and third by analyzing optimal decisions to deal with unacceptable situations. There are two new models presented in this thesis. The first model, is a simple design model that uses historical failures to assess the piping safety. The second model describes the formation of piping erosion paths in spatial variable soils. Especially the second model might potentially lead to improvements in piping modeling and potentially in cost reductions. The main conclusions from this thesis is that the piping mechanism and influence of spatial variability in the subsoil are a very significant threat to flood defenses. However, performing local soil measurements in combination with local dike improvements can be a cost-effective method to deal with unacceptable piping failure probabilities.

Schematic formation of a piping channel below a dike in heterogeneous soil

2. ...
Ph.D. thesis by Timo Schweckendiek, defended around the summer of 2013

3. Piping, reality or calculation error  

This recent ENW report (in Dutch) assesses whether the calculated high probabilities of failure are indeed worrying or merely a calculation error. The main outcome is that the probabilistic calculation give a realistic representation of the safety against piping, which is not sufficient in the Netherlands.

Future projects (MSc thesis subjects)

There are several highly interesting in the field of geotechnical reliability of flood defences, which can serve as a basis for MSc thesis.