José A. Villanueva and María C. Verde
Fluid model in pipelines, Fault diagnosis, Structural analysis, Transportations systems, Intelligent signal processing
This work copes with the fault diagnosis problem for branched pipelines. Since the models of networks are usually sparse, a fluid model transformation is proposed that simplifies the residual generator design for fault detection. The transformation is achieved by the application of structural analysis together with properties of equivalence classes of the fluid model equations. This transformation allows to obtain a set of subsystems in which each of them is robust to a fault set by considering the structural model redundancy. As a case study, the monitoring of a pipeline with two branches is presented where the outflow rate deviation at each branch must be detected. Experimental data of a hydraulic pilot pipeline shows the simplicity of the algorithm.
Important Links:
Go Back
