ABSTRACT
The Tennessee Valley Authority (TVA) is interested in estimating the frequency distribution of dam failure risk to inform the decision making process for identifying, prioritizing and justifying structural and non-structural dam safety risk-reduction measures. Flood risk is comprised of the combination of the frequency of flood events, the potential failure modes that could occur during flood events, and the associated consequences of failures. Riverside is working with TVA, CADSWES and others to develop a framework to evaluate flood risk. One aspect of this includes development of flood frequency (hazard) curves for the reservoir system. This involves (1) generating stochastic precipitation events thousands of times, (2) sampling initial soil moisture and reservoir levels, (3) generating hydrologic inflows to the system, (4) executing a rule-based RiverWare model of the system, and (5) assembling simulation statistics (e.g., peak headwater or duration above critical threshold) from all events to develop the flood hazard curves. The rule-based RiverWare model builds on previous development work and will reflect current operating policy. In future work, we plan to incorporate stochastic inputs in the RiverWare model to simulate failure of gates, debris blockage, and potentially dam breaches as a part of the stochastic simulation process. Once fully built, the stochastic flood simulation framework will provide a means for TVA to assess the impact of policy changes and structural improvements on the flood hazards and risks throughout their system. This presentation will provide an overview of the simulation framework along with unique aspects of the RiverWare modeling.

Click HERE for a PDF version of the presentation slides.