Method for Online Structural Health Monitoring Based on Single-point Response and Neural Network

In the critical field of aerospace maintenance, non-destructive testing technologies play an indispensable role in the comprehensive inspection of aircraft structures. To enhance the efficiency of identifying, localizing, and even quantitatively describing structural damage, this study introduces an innovative online intelligent health monitoring method. The method utilizes single-point vibration response data from cantilever beam structures, which are common in aerospace applications, as the basis for real-time assessment. The non-destructive testing method examined in this research is grounded in the detection of changes in structural modal parameters, which are highly sensitive to damage. By training deep neural networks to extract discriminative features from vibration response data, the method achieves real-time damage identification with high accuracy. Furthermore, the study integrates regression neural networks into the framework, developing a carefully designed online intelligent health monitoring method that is highly suitable for verifying damage in cantilever beams. The results of the experimental validations are promising, indicating that the proposed health monitoring method can assess the condition of specimens online with high precision, with a maximum deviation from the detected damage location of less than 5.04%.