University of Twente Student Theses

Login

Including the effect of non-water-retaining objects in the probabilistic modelling of dune safety

Raaben, Thijs (2015) Including the effect of non-water-retaining objects in the probabilistic modelling of dune safety.

[img] PDF
4MB
Abstract:An important function of dunes in the Netherlands is to protect the hinterland against flooding. The land behind the dunes is the most densely populated and economically valuable area of the country. Therefore, dune safety is extremely important for the country. The safety of dunes as flood protection is tested every twelve years. The statutory safety assessment model DUROS+ is based on a sand balance with many simplifying assumptions. One of the assumptions in DUROS+ is that hard elements are absent, which is not valid at many locations along the Dutch coast. When these hard elements have no water retaining function, these elements are called ‘Non Water retraining Objects (Abbreviated as NWO)’. The presence of NWO’s in dunes affect the sediment availability from the dunes. This is because a soft erodible part of the dune is replaced with a hard non-erodible part. Deltares and Arcadis (2014a) developed guidelines (Referred to as: DnA rules) to account for the effect of NWO’s on dune safety. These guidelines are developed for NWO failure and NWO non-failure. However, these guidelines are not implemented in the current dune safety assessment model and parameter/model uncertainties of these rules are neglected. The current statutory safety assessment model is a semi-probabilistic model, which means that the actual safety test is performed with a deterministic model, but input is based on probabilistic calculations. The impact of NWO’s on dune erosion is not taken into account in these probabilistic calculations. Therefore, the impact of NWO’s on dune erosion should be implemented in a probabilistic model to analyse whether the hydraulic input parameters change due to the incorporation of NWO’s in the probabilistic calculations. In addition, another reason to implement the impact of NWO’s on dune erosion in the probabilistic model is to analyse the importance of model and parameter uncertainties of the DnA rules. The probabilistic model is a combination between DUROS+ and a probabilistic method. Because dune erosion has very low failure probabilities, a probabilistic method like Monte Carlo is very inefficient because a very large number of computations is necessary to provide reliable results. Therefore, the First Order Reliability Method is used as probabilistic method, because this method is very efficient for very low failure probabilities. However, this method is not always applicable. When this method is not applicable, Monte Carlo with Importance Sampling is a good alternative. Stochastic distributions for model and parameter uncertainties of the DnA rules are developed. The discrepancy between the predicted model outcomes and reality is called model inadequacy. Important characteristic of model uncertainty is that this kind of uncertainty is present even if there is not a single unknown model parameter in the model. Model uncertainty is hard to determine because no real data about the behaviour of NWO’s during extreme storm surges is available. Therefore, model uncertainty is estimated with the use of other dune erosion models. Parameter uncertainty is dependent on the available information and is case specific. Six academic cases and a field case of the Palace Hotel in Zandvoort are used to analyse the main difference between results of the probabilistic model with the DnA rules included compared an results of the semi-probabilistic model with the DnA rules included. ii The main difference between the semi-probabilistic and the probabilistic dune erosion model is that the semi-probabilistic model overestimates the cross shore location of the 10-5 erosion points in all academic cases and Palace hotel case. Another difference is that the probabilistic dune erosion model provides insight in failure probabilities along the whole dune while the semi-probabilistic model only shows the binary failure/non failure as result. The hydraulic input parameters for the semi-probabilistic dune safety assessment model approximate the combination of input parameters that lead to the location of the 10-5 erosion point. The results of the probabilistic dune erosion model with the incorporation of NWO’s did not show significantly deviating results than values in the HR2006, which is the input for the semi-probabilistic model. Therefore, the semi-probabilistic model could still be used to test dune safety. However, when the dune is ‘just safe’ or ‘just unsafe’ a more accurate calculation is required. The probabilistic model should be used in these cases for a more accurate calculation and to provide insight in failure probabilities along the whole dune. Both parameter and model uncertainties of the DnA rules show significant impact on dune safety. The magnitude of this impact is very case specific. Therefore, parameter and model uncertainties should be considered in the DnA rules. The dune of the Palace Hotel is safe according to the current statutory safety assessment method. However, the dune is not safe when the influence of the NWO on dune safety is considered according to the semi-probabilistic model with the DnA rules. The probabilistic model with the incorporation of the DnA rules shows a safer result than the semi-probabilistic test, but the is also unsafe according to the probabilistic model with the DnA rules.
Item Type:Essay (Master)
Faculty:ET: Engineering Technology
Subject:56 civil engineering
Programme:Civil Engineering and Management MSc (60026)
Link to this item:http://purl.utwente.nl/essays/67023
Export this item as:BibTeX
EndNote
HTML Citation
Reference Manager

 

Repository Staff Only: item control page