On improving data and models on corrective control failures for use in probabilistic reliability management

One of the most pressing concerns in the investigation of new probabilistic reliability criteria pertains to the data required as input to the evolving probabilistic models. This paper discusses an area of failure data collection that has been overlooked in power system reliability studies: corrective control actions. Background information is provided on the need for evolution of data collection systems in this context. Further, steps that can be taken to build a database of parameters necessary for modelling corrective actions are provided, to be useful in assessing system behaviour from a probabilistic reliability management perspective. Modelling of corrective control actions using event trees is illustrated. Throughout, the various challenges foreseen in the building of corresponding databases and models are outlined.