A sparse sensor network topologized for cylindrical wave-based identification of damage in pipeline structures

Abstract

A sparse sensor network, based on the concept of semi-decentralized and standardized sensing, is developed, to actively excite and acquire cylindrical waves for damage identification and health monitoring of pipe structures. Differentiating itself from conventional 'ring-style' transducer arrays which attempt to steer longitudinal axisymmetric cylindrical waves via transducer synchronism, this sparse sensor network shows advantages in some aspects, including the use of fewer sensors, simpler manipulation, quicker configuration, less mutual dependence among sensors, and an improved signal-to-noise ratio. The sparse network is expanded topologically, aimed at eliminating the presence of 'blind zones' and the challenges associated with multi-path propagation of cylindrical waves. Theoretical analysis is implemented to comprehend propagation characteristics of waves guided by a cylindrical structure. A probability-based diagnostic imaging algorithm is introduced to visualize damage in pixelated images in an intuitive, prompt, and automatic manner. A self-contained health monitoring system is configured for experimental validation, via which quantitative identification of mono- and multi-damage in a steel cylinder is demonstrated. The results highlight an expanded sensing coverage of the sparse sensor network and its enhanced capacity of acquiring rich information, avoiding the cost of augmenting the number of sensors in a sensor network.