Flexible Pressure Sensors Based on CB/CNTS at PDMS for Wind Pressure Field Monitoring of Cable-Stayed Bridge Using Machine Learning

Strong winds can cause severe lateral vibrations in the cables of large-span cable-stayed bridges, leading to damage to the bridge. Real-time and effective acquisition of vibration behavior parameters in service is essential to assess the safety performance of structures. Existing sensing elements for detection include stress gauges, fiber optic gratings, and piezoelectric ceramics, but most of them have the disadvantages of high cost, poor durability, and low sensitivity. In this article, a flexible piezoresistive sensor based on carbon black (CB)/carbon nanotubes (CNTS) at PDMS has been developed with good tensile properties (tensile fracture rate up to 230%) and electrical properties (GF up to 12.1). It adheres closely to the surface of the cable and detects with an accuracy of 97.231%. This article analyzes the force distribution on the sensor surface and static loading by ANSYS, considering the fluid-structure coupling effect, and combined with wind tunnel tests to simulate the vibration response of the sensor when the cable of the cable-stayed bridge is affected by wind. An intelligent monitoring system for cable-stayed bridges was designed using visual studio. The system can upload and display wind pressure field data in real time, enabling the prediction and assessment of the bridge’s health status. Finally, the sensor performs feature extraction and state recognition of the collected data, and a machine learning algorithm based on approximate entropy and BP neural networks was used to achieve classification and recognition of the wind pressure field data (97.231%). This study provides new ideas on the aerodynamic characteristics of cable-stayed bridges and structural health inspection.