Damage detection of structures using signal processing and artificial neural networks

Over the last two decades, extensive research has been conducted on structural health monitoring and damage detection in order to reduce the life-cycle cost of structures and improve their reliability and safety. These methods are divided into modal-based and signal-based approaches. Recent advances in the field of sensor technologies have facilitated the use of signal-based methods as practical solution to detect damages in structures. After a sever earthquake, usually there is no possibility for visiting the individual structures. Therefore, application of methods that can detect damage of structures only by using signals recorded at the time of the earthquake is noteworthy. Many existing methods, especially methods based on signal processing are not able to determine the damage severity. This article presents a signal-based seismic structural health monitoring technique for damage detection and evaluating damage severity of a multi-story frame subjected to an earthquake event. As a case study, this article is focused on IASC–ASCE benchmark problem to provide possibility for side-by-side comparison. First three signal processing techniques including EMD, HVD and LMD, which are categorized as instantaneous time-frequency methods, have been compared to find a method with the best resolution in extracting frequency responses. Based on the results EMD has proved to outperform than the others. Second, EMD is used to extract the acceleration response of the sensors. Results show that by taking advantage of signal processing and artificial intelligence techniques in this research, damage detection of structures was carried out for three levels including damage occurrence, damage severity and location of the damage.