Implementation of signal processing methods in a structural health monitoring (SHM) system based on ultrasonic guided waves for defect detection in different materials and structures

The local defect inspection in longitudinal structures such as plates or pipelines implies high economical costs and it is time consuming mainly in underground infrastructures, energy or water, and aerospace sectors. Moreover, if these structures are non-accessible, their local inspection is not possible. Ultrasonic (US) inspection technique based on guided waves is one of the potential alternatives to address this issue. The US inspection based on these type of waves could be applied in many scenarios to monitor the damage state of structures; i.e., in water underground pipelines to identify the wall thickness losses or impact damage detection on Carbon Fiber Reinforced Composites (CFRC). A SHM system based on guided waves requires a special signal processing in order to identify possible damage in the structure. The signal emitted and received is a combination of different propagation modes which are difficult to identify and analyse. However, if the signals are compared to each other (signal related to non-damaged components compared to damaged signal) it is possible to measure their difference as a distance that can be used to estimate the damage level. In this work, signals corresponding to non-damaged samples have been captured and then different types of damage have been applied for different cases. After the data acquisition phase, the comparison between signals has been carried out by applying different mathematical methods and distance metrics (SDC, DTW, Euclidean, Manhattan and Chebyshev), with the aim of detecting defects in different structures and materials. For this purpose, two cases have been analysed: 1) In CFRC plates subjected to impact damage and deformations and 2) In a pipe coated by cement-mortar in order to quantify the wall thickness losses. In both cases ultrasonic PZT sensors, an ultrasonic multichannel pulser/receiver and a software developed ad-hoc have been used. Although the SHM system components were similar, it must be noted that the type of ultrasonic guided waves used were different; in the case of CFRC plates, Lamb waves were excited whereas in the case of the pipeline, Love waves have been used. A comparison between the above mentioned methods is provided. The results show the validity of the approach for damage characterization.