Probabilistic Design of Relief Wells as Piping Mitigation Measure

Abstract

The traditional Dutch way to deal with piping for river levees is the implementation of piping berms. The disadvantage of such a measure is the inland space required, especially in urban areas. Relief wells, on the other hand, require less or no inland space and therefore represent an attractive solution as mitigation measure against piping. The aims of this research are first, to show how reliability analysis of relief wells systems can be carried out, and second to examine the costs required to achieve a reliability target for piping failure, as set in the Netherlands. The outcomes of this analysis will help comparing relief wells with piping berms in economic terms. To obtain these results, the statistical parameters of the influencing variables are studied using both, the collected data from existing projects in the Netherlands, and data from relevant literature. A reliability-based design approach is followed to estimate the reliability of relief wells systems. In order to establish the limit state functions, the assessment methods recommended by the Dutch flood defence regulations are used. Applying the probabilistic axioms it is possible to resemble piping failure as a parallel system assessing uplift and heave failure mechanisms. To estimate the hydraulic head in relief wells system, the United States Corps of Engineers method is applied, as well as the latest developments in flood risk analysis, achieved by the VNK project, which are used to determine the reliability target. To estimate the probability of failure and the system reliability, MCS and FORM methods are utilized. A tailor-made comprehensive tool is built in Matlab to compute the hydraulic head in relief wells system and to perform the probabilistic analysis. Subsequently, a LLC analysis is performed with the aim to account for the life cycle of relief wells. A comparison of the net present value of the two alternatives (relief wells and piping berms) is made. Finally, analysis of two case studies with different scenarios are performed to show the possible economic advantages of installing relief wells and sensitivity analysis is used to underpin the robustness of the conclusions. The results show that, using USACE method, the blanket and the aquifer permeability, as well as the hydraulic losses, are the dominant variables (from the 'load' side). The sensitivity factors show high discrepancy between partially and fully penetrated wells. Even when the entrance losses cannot be accurately predicted, a total clogging scenario of the filter can be neglected. One main limitation for the applicability of relief wells system, is that there is a maximum achievable head reduction. This maximum head reduction is limited to the minimum possible well spacing (a>D/4). Results from the case studies show that relief wells are a cost-effective as piping mitigation measure, outperforming piping berms. This advantage can be up to a factor of ten regarding initial investment. This allows accounting for a shorter life cycle for relief wells in order to equate the same LCC as piping berms.