Most contemporary risk assessment techniques, such as failure modes and effect analysis (FMEA), fault tree analysis (FTA), and probabilistic risk analysis (PRA) rely on a chain-of-event paradigm of accident causation. Event-based techniques have some limitations for the study of modern engineering systems; specifically hydropower dams. They are not suited to handle complex computer-intensive systems, complex human-machine interactions, and systems-of-systems with distributed decision-making that cut across both physical and organizational boundaries. The emerging paradigm today, however, is not to analyze dam systems separately by breaking the major disciplines into stand-alone vertical analyses; but to explore the possibilities inherent in taking a systems approach to modeling the reliability of flow-control functions within the entire system.

This dissertation reports on the development and application of systems reliability models to operational aspects of a hydropower cascade in Northern Ontario: The Lower Mattagami River (LMR) Project operated by Ontario Power Generation (OPG). The reliable performance of a spillway system depends on the many environmental and operational demand functions placed upon it by basin hydrology, the hydraulic conditions at reservoirs and dams, operating rules for the cascade of reservoirs in the basin, and the vagaries of human and natural factors such as operator interventions or natural disturbances such as ice and floating debris (Regan, 2010). These systems interact to control floods, condition flows, and filter high frequencies in the river discharge. Their function is to retain water volumes and to pass flows in a controlled way. The reliability of flow-control systems is a broad topic that covers structural, mechanical, electrical, control systems and subsystems reliability, as well as human interactions, organization issues, policies and procedures. All 3 of these occur in a broad spectrum of environmental conditions. A systems simulation approach is presented for grappling with these varied influences on flow-control systems in hydropower installations.

The Mattagami River cascade operated by Ontario Power Generation is a series of four power stations along the Mattagami River and the Adams Creek bypass channel from Little Long GS at the top to the cascade to the Mattagami River below Kipling GS at the bottom. The number of riparians in the river flood plain is few and there is no commercial riverine navigation, so potential loss of life is small or negligible and operational safety dominates. Upstream of Little Long dam is a seasonally-varying inflow and a reservoir. The remaining three dams downstream (Smokey Falls, Harmon, and Kipling) have little storage capacity. Each dam has two vertical lift gates and all four structures have approximately the same spillway capacity. Far downstream, the river discharges into Hudson's Bay. Hydrological and climate frequency data are available for a period of 50 years. The problem facing the project was to conceptualize a systems engineering model for the operation of the dams, spillways, and other components; then to employ the model through stochastic simulation to investigate protocols for the safe operation of the spillway and flow control system. Details of the modeling, analysis, and results for safe operation of the cascade are presented.