Advanced Method for Detecting Irregularities in Concrete Beams Using IOT Data from Sensors

The increasing reliance on state assessments in civil engineering has sparked extensive research into methods for damage detection based on structural vibrations. Modal parameters, such as natural frequencies and mode shapes, have gained significant attention due to their invariance across structures. These parameters provide a global perspective, meaning their variations can help identify damage without the need for sensor placement directly at the damaged site. This feature is a key advantage in structural health monitoring (SHM) systems. Integrating MEMS sensors into SHM frameworks holds great potential for long-term monitoring, particularly for large-scale infrastructures. This paper introduces an innovative anomaly detection technique that analyzes raw sequential data through a statistical approach to identify damage associated with tendon prestress loss. The technique leverages a distributed monitoring system consisting of six high-performance MEMS sensors. To validate the system, the first mode frequency is initially analyzed, and the method is then tested on acceleration data from a 240 cm beam under three distinct damage scenarios. The results demonstrate high accuracy in damage detection and show that the system can also localize the damage effectively.