Using a systems approach to analyze the operational safety of dams
Abstract
Dam systems are arrangements of interacting components that store and convey water for beneficial purposes. Dam failures are associated with extreme consequences to human life, the environment and the economy. Existing techniques for dam safety analysis tend to focus on verifying system performance at the edge of the design envelope. In analyzing the events which occur within the design envelope, linear chain-of-events models are often used to analyze the potential outcomes for the system. These chain-of-events models require that combinations of conditions are identified at the outset of the analysis, which can be very cumbersome given the number of physically possible combinations. Additional complications arising from feedback behaviour and time are not easily overcome using existing tools. Recent work in the industry has begun to focus on systems approaches to the problem, especially stochastic simulation. Given current computational abilities, stochastic simulation may not be capable of analyzing combinations of events that have a low combined probability but potentially extreme consequences. This research focuses on developing and implementing a methodology that dynamically characterizes combinations of component operating states and their potential impacts on dam safety. Automated generation of scenarios is achieved through the use of a component operating states database that defines all possible combinations of component states (scenarios) using combinatorics. A Deterministic Monte Carlo simulation framework systematically characterizes each scenario through a number of iterations that vary adverse operating state timing, impacts and inflows. Component interactions and feedbacks are represented within the system dynamics simulation model. Simulation outcomes provide useful indicators for dam operators including conditional failure rates, times to failure, failure inflow thresholds, and reservoir level exceedance frequencies. Dynamic system response can be assessed directly from the simulation outcomes. The scenario results may be useful to dam owners in emergency decision-making to inform response timelines and to justify the allocation of resources. Results may also help inform the development of improved operating strategies or upgrade alternatives that can reduce the impacts of these extreme events. This work offers a significant improvement in the ability to systematically characterize the potential combinations of events and their consequences.