Forecasting of ice-loads on concrete dams

When evaluating the safety level of existing dams, understanding the loads to which they may be exposed is very important. Concrete dams are exposed to large variations in climate between summer and winter and in Swedish winter conditions the dams are often exposed to ice pressure from the ice cover formed on the reservoir. In order to ensure a high level of dam safety and avoid high costs for stabilisation measures, further research is needed on what controls the origin and magnitude of the ice load.

Current guidelines require dams to be designed to withstand an ice pressure of 5-20 tonnes per metre of dam at the water surface. For low gravity dams that are stabilised by their own weight, it may be difficult to meet the dam stability criteria against such a large horizontal load. Although the ice load is assumed to be large, there are uncertainties in the parameters that control the size of the ice load and with what return time the large load can occur. In several previously completed SVC projects, ice loads have been studied and much of the previous work is referenced together with new knowledge in Hellgren (2022). Through analyses of available data and new measurement data from a larger ice load panel in Rätan (Faxälven) together with numerical modelling, some of the likely parameters that control the size of the ice load have been identified. For example, temperature as well as ice thickness and water level variations seem to be important. However, several uncertainties remain, for example it is unclear how the ice load would vary along a dam. In the present project, further evaluation and analysis of collected measurement data will be carried out to calibrate and further develop developed forecast models for ice load. Among other things, another large-scale ice load panel has been installed in Storrnorrfors (Umeälven), which provides new input data from a plant with different boundary conditions than Rätan. Continuous measurement of ice thickness, now at both plants, is also included in the analysis. The forecast models previously developed will be updated based on new measurements and will then be used to study whether there may have been changes in the ice load over a longer time perspective due to climate change (higher temperatures) and other operating modes in the hydropower plants due to the introduction of more unplanned electricity generation (wind). The latter has resulted in more frequent reservoir water level fluctuations than before. Furthermore, the forecast models are planned to be used to illustrate how the ice load could vary along a river stretch where different facilities have different conditions (regional/facility specific) for how large ice loads can occur. The variation of the ice load along a dam is studied in the final phase of the project to determine which design loads should be used in a future update of the dam safety guidelines. Previous studies have shown that locally measured high ice loads on a short stretch are not necessarily the load prevailing on a longer dam section, which is the interesting part of a stability analysis. Through literature studies and numerical modelling, the phenomenon is planned to be studied and as part of the work, a concept for measuring ice load on a longer distance is also planned to be developed.