Study On Damage Identification Of Pipeline Based On Evidence Theory

Pipelines are widely used in many areas of national life and industrial production.Changes in load, deterioration with age and environmental influences may cause local or global damage to pipelines. So it becomes increasingly important to conduct continuous damage identification on pipeline to prevent sudden structure failures. Significant progress has been achieved in the area of damage identification based on modal parameters.However, single damage identification method may easily lead to wrong conclusion of damage diagnosis in practice due to the measurement noise and limitations of testing conditions.This paper combines D-S evidence theory with pipeline damage identification. Firstly,the noise sensitivity and applicability of six commonly used damage identification methods is studied. Simulation results prove that single damage identification method shows low ability in resisting noise and limitations in applicability.Secondly,a two stage pipeline damage identification method based on evidence theory and simplex genetic algorithm is proposed to improve the accuracy of damage identification with measurement noise.Two kinds of information sources are integrated by evidence theory to determine the damage area,and then simplex genetic algorithm is used to precisely detect damage location and extent.Satisfactory results can be achieved under small noise.Finally, in order to solve practical engineering problems, an improved evidence theory based on multiple information sources is proposed for pipeline damage identification. Experiment results show that the proposed method improve the identification results and it is superior to the traditional evidence theory.