Many flood control dams are founded on alluvial deposits, which usually grade from fine-grained materials near the ground surface to coarse sediments with high permeability in the lower part of the strata. If dykes or levees do not have a cut-off wall fully penetrating the aquifer, underseepage occurs beneath the dam structure during high river levels. Such seepage leads to an excessive water pressure in the pervious foundation (aquifer) landside of the embankment dam. Most dangerous situations occur when high hydrostatic pressure acts on the upper soil layers and ground failure develops in the form of uncontrolled boils or uplift of the blanket. In these cases, the installation of pressure relief elements at the landside embankment toe prevents erosion and hydraulic rupture of the stratum. ^Relief wells and relief drainages in the form of geotextile wrapped stone columns or trenches are highly permeable drainage elements that ensure a defined hydraulic connection with the aquifer. They relieve the uplift pressure below the impervious blanket while water discharges freely to the surface. Therefore, the safety factor against hydraulic failure increases significantly. However, the hitherto existing design criteria for relief drainages were insufficient. Most approaches determine the pressure relief and the discharges only based on assumptions and experience from former projects. Geometrical and hydraulic parameters of relief drainage elements (e.g. diameter, embedment length, centre-to-centre distance, permeability, etc.) are usually not considered. These parametric influences on the relief behaviour and corresponding within general requirements for installation were not sufficiently described until now. In this dissertation, mathematical approaches are presented for an underseepage analysis, referring to embankment dams (dykes, levees) founded on pervious and semi-pervious soil layers. Analytical solutions are derived for the pressure distribution in a flow field of finite and infinite length with open and blocked boundary conditions. This simplified underseepage analysis for steady state conditions provides a basis for the judgment, if pressure relief measures are necessary. In addition to these mathematical analyses, different pressure relief systems are described. Besides a proper installation technique, the design of the geotextile filter of relief drainages is significant for the long-term behaviour. Therefore, the most important design criteria and recommendation are discussed. Furthermore, this dissertation includes also the general design for the installation of relief wells. The detailed study of pressure relief behaviour and the estimation of discharge from relief drainages were based on a combined modelling method by using experimental and numerical techniques. Firstly, a small-scale (1:10) dam model on a two-layer strata was used to analyse the influence of main parameters (diameter, embedment length, permeability and number of drainage relief columns) on the relief behaviour of stone columns and trenches. The results of these tests built the basis for calibration of the numerical model and for the large-scale modelling (scale of 1:1) as well. Furthermore, this gradual approach allowed parametric studies exceeding the geometric and hydraulic limits of the small-scale model. Although the homogenous permeability of the pervious soil layer could not be entirely kept in the large-scale model, it was possible to analyse the relief behaviour of drainage stone columns under laboratory conditions on a scale of 1:1. In this way, the results from small-scale modelling were verified by means of several tests in the large-scale facility. The high risk of hydraulic ground failure due to underseepage was also confirmed. Based on the combined modelling results, the relief behaviour of drainage stone columns can be described with high accuracy regarding the practical application. Finally, recommendations for design and estimation of discharge are derived.