Application of Impedance-Based Structural Health Monitoring to Piping Systems

Abstract

Pipelines convey natural gas, oil, and some pipelines contain communication and power cables, all of which are very important to maintain functional residential and industrial facilities. The solution proposed in the research for a more reliable, economical monitoring system involves a damage detection process known as structural health monitoring (SHM). To implement an SHM system with a pipeline, an array of sensors deployed at various locations is needed to make observations regarding the damage state of the structure. In this study, piezoelectric Macro-Fiber Composite (MFC) transducers are used for SHM in pipeline systems. Damage sensitive features (i.e. root-mean-square-deviation (RMSD), correlation coefficient (CC)) for impedance signatures were used to decide damages. To apply previous technique which is carried out in experiment study on pipeline system on real structure, sensor array, sensing region, and excitation frequency is important issues. Therefore parameters of impedance-based damage detection techniques (frequency range, sensing region, damage size) are tested through experiment study. It has been also found that loosening bolts damage is more sensitive with continuity of structure members and crack damage is more related to additional structure component between sensors and cracks. Based on these results, efficient location of sensors could be decided. Sensors which are placed on effective location could detect both loose bolts and cracks. Signatures which were from loose bolts and crack have to be classified for effective SHM. Statistical pattern recognition algorithm such as probabilistic neural network (PNN) was employed to classify damage between loose bolt and crack damage. It has been found that the proposed damage estimation based on probabilistic neural network (PNN) is capable for classifying damage between loosening bolts and crack on pipeline systems on the impedance-based SHM technique.