Applied Risk Management in Electric Power Plant Decision Making

Large electric power plants incorporate a large number of complex, highly redundant and very reliable systems in order to produce power in safe and efficient manner. Much effort is expended to assure that the systems are maintained to maximize the availability of all normal and safety-related systems. To assure safe operation while performing maintenance or responding to unexpected equipment outages, power plants must continually evaluate the risk of changes in plant configuration. Due to the high redundancy of these power plants, simple, single-point failures are usually not present in the design, so availability and risk calculations require complex calculations. The risks are usually modeled as fault trees and event trees using a Probabilistic Risk Assessment (PRA). To leverage these design-time risk assessment tools for operational purposes, the industry has developed various configuration risk management (CRM) tools to calculate and display the impacts of these configuration changes. Using the current plant configuration, these tools propagate information through the model and quantify risk measures, which are then translated into color-coded status panels, timelines, and lists of relevant and risk-significant activities. Within seconds, plant operators can identify a safety problem and the specific work activities that cause it, and, from this, decide whether the problem is significant enough to warrant special contingency actions. In this way Probabilistic Risk Assessment (PRA) has developed over the last 40 years into a standardized and rigorous activity that is used routinely by utilities that operate power plants. This paper takes a look at current risk monitoring activities in the United States, using modern tools such as the EPRI Phoenix Risk Monitor. It discusses how these tools are implemented to communicate in the language of its users, minimize operator impacts, provide quality risk metrics, incorporate both in-plant and external hazards such as spatially-related and scenarios-based impacts from as fires, flooding, seismic, high winds, and tsunamis, and then produce results that support practical and effective decision making for justification of continued operation and design modifications.