Compact and even portable devices integrated with Structural Health Monitoring systems are a promising solution to overcome issues regarding to practical implementation in real structures. This paper compares the computational capabilities of two Linux based development platforms for implementing structural monitoring algorithms. The Beagle Bone Black and Odroid-U3 single-board computers are studied. Attractive features as open-source software availability, easy configuration, low power requirement and hardware flexibility design, facilitate the use of these hardware platforms as an embedded system to contribute in recent advances obtained in the field of SHM and damage detection. From these devices, in this paper, storage capabilities, memory requirements, computational complexity and time processing consuming are analyzed and compared to evaluate benefits of embedding structural damage detection algorithms. In this work, a Piezo - diagnostics approach based on guided waves methods is embedded. The main components of the proposed damage identification system consist of a piezo actuator active system, a computational core (BBB or Odroid-U3 hardware), and a modelling block based on statistical features estimated by means of principal component analysis. The embedded algorithm was validated by using experimental data gathered from a steel carbon pipe section. Measurements from piezoelectric devices attached to the surface structure are used to distinguish damaged and undamaged conditions. Fifteen damage classes are seeded in the specimen by adding masses at different locations on the surface, where 200 experiments per damage are conducted. The feasibility to implement an automated real time diagnostic system is demonstrated.
doi: 10.12783/SHM2015/147