Real-Time Failure-Detection and Identification Applied to Supervision of Oil Transport in Pipelines

In this paper methods for real-time failure detection and identification are discussed. The methods apply parallel filters of the Kalman type based on plant models that describe the different failure situations as well as normal operation. A failure is determined by identifying the filter having the highest probability of representing the plant. The tests are based on the innovation sequence produced each of the filters. The last section presents simulations from a system designed for supervision of oil transport in pipelines.