Resumo

TORRES, Lizeth; VERDE, Cristina; CARRERA, Rolando  e  CAYETANO, Raúl. Diagnostic Algorithms to Detect Faults in Pipelines. Tecnol. cienc. agua [online]. 2014, vol.5, n.4, pp.57-78. ISSN 2007-2422.

This paper presents the design of an online diagnostic system (functioning while the transportation system remains in operation) to detect, identify and rebuild pipeline faults based on redundant relations and state observers. The faults include those occurring in measurement instruments, pumps, and unknown extractions from the pipeline. The algorithms that compose the diagnostic system were developed from nonlinear partial differential equations that characterize the behavior of the fluid according to the principle of conservation of mass and momentum. These equations were approximated in space using the Finite Differences Method. In order to distinguish between the different types of faults and reconstruct their behaviors, the diagnostic system operates in stages. The first one -which is called Fault Detection and Isolation- aims to isolate each fault symptom with the help of a set of redundant relations that are deduced from the nominal model of the pipeline, i.e., in normal conditions. This first stage simplifies the second, called Fault Reconstruction, which is composed of observation algorithms that estimate the temporal evolution of the isolated faults. The overall diagnostic system is validated through a series of experiments carried out in a pilot hydraulic pipeline at the Engineering Institute, UNAM. This pipeline is approximately 200 m long and was implemented and built specifically to carry out tests to monitor pipelines.

Palavras-chave : Diagnostic systems; fault detection; hydraulic pipelines; pipeline monitoring.