Robust Fault Diagnosis for Pipelines

TL;DR

This paper develops a fault diagnosis system for pipelines using nonlinear observers and redundant relations, capable of detecting, isolating, and reconstructing various faults based on a fluid dynamics model.

Contribution

It introduces a novel staged diagnosis approach combining nonlinear modeling and observers for comprehensive pipeline fault detection and reconstruction.

Findings

Successfully detects and isolates faults in a hydraulic pipeline

Reconstructs fault behavior using nonlinear observation algorithms

Validated through experiments on a 200-meter pilot pipeline

Abstract

This paper presents the design of a diagnosis system for the detection, identification and reconstruction of faults in pipelines. The design of such diagnosis system is based on redundant relations and nonlinear observers, taking into account faults in sensors, damages in pumps, and unknown extractions. The proposed algorithm is developed based on a model described by nonlinear equations of the fluid behavior in a pipeline, considering the principles of conservation of mass and momentum. In order to distinguish among different types of faults and to reconstruct their behavior, the diagnosis system operates in stages. The first stage called detection & fault isolation aims to isolate a fault symptom with a set of redundant relations deduced from the analysis of the model in nominal conditions and assuming measurements of the standard variables at the extremes of the pipeline. In the second stage named fault reconstruction, nonlinear observation algorithms estimate the temporal evolution of the isolated fault. The complete diagnosis system is validated through a series of experiments in a hydraulic pilot pipeline of 200 [m].