Dam Failure Analysis of the Parteboda Hydropower Plant : A deterministic approach

Abstract: Despite all the benefits a dam can provide, hundreds of years of experience have shown the destructive forces that can be unleashed by dam failure. To carry out risk management and execute mitigative measures for a dam facility, it is first necessary to assess the incremental consequences of a potential dam failure. This in turn was the initial subject of this thesis. While conventionally a dam failure analysis is carried out applying a deterministic approach, this thesis aimed to further develop the ap- proach and provide a probabilistic method. Due to the results and the time frame of the thesis, however, it was decided to pursue the probabilistic approach only theoret- ically.  In contrast to the deterministic approach, the probabilistic approach does not only con- sider single scenarios but instead a multitude of scenarios over a whole range of pos- sible parameter combinations. This offers the possibility to create probabilistic flood maps, which have the advantage of indicating the probability of flooding for an area. Accordingly, mitigation measures can be dimensioned and planned in a cost-benefit efficient way.  For the implementation of the deterministic approach at the Parteboda hydropower plant, the combination of a one-dimensional and a two-dimensional numerical hydrau- lic model was chosen. For the one-dimensional modeling, which was used to predict the dam breach and connected outflow discharge, the software package MIKE Hydro River (1D) was chosen. The generated breach hydrographs were then used as input for the two-dimensional modeling, for which HEC-RAS (2D) was chosen.  Modeling of the dam breach, the erosion-based approach was first selected. However, the results were extended by creating additional user-defined breach processes (time- based approach). The manual breach processes were based on empirical formulas. Subsequently, the hydrographs of the addressed scenarios were extracted and served as boundary condition for the two-dimensional modeling.  Already for the breach hydrographs, differences between the dams and scenarios were relatively small. Even a variation of the parameters breach width and formation time showed little impact. As already indicated by the results in the form of the hydro- graphs, there were no significant differences with respect to the flood areas. Differ- ences in incremental consequences were therefore relatively insignificant. As a reason, the small volume of the turbine channel could be identified, which is separated from the main reservoir and acts as the first discharged volume. In addition, the from the reservoir inflowing water is limited by a constriction of the upper turbine channel. Consequently, as soon as the breach exceeds a certain width, the outflow is no longer determined by the breach itself but instead by the narrowing cross section upstream.  Due to the small differences between the dams and scenarios, events with a wide range of likelihood can lead to similar incremental consequences. This in turn makes it difficult to hierarchize individual dams or measures.