Experimental investigation and prediction model on residual mechanical properties of GFRP strips under wind-blown sand erosion

This study examines the impact of wind-blown sand erosion on the durability of glass fiber reinforced polymer (GFRP) strips through experiments and modeling. Erosion tests were conducted considering the angle, velocity, and time of exposure, followed by evaluation of the residual mechanical properties. From the test results, there were extensive erosion pits on the specimens’ surfaces. In some areas, the epoxy adhesive had been peeled off, resulting in the formation of through-holes. Scanning electron microscopy (SEM) further indicated severe epoxy deformation and spalling, with fiber exposure and fractures in highly damaged areas, leading to material degradation. Test results showed that as the erosion angle increased to 90°, there was a 16.4 % decrease in the ultimate tensile stress of the specimens, but the elastic modulus of the GFRP strips was not sensitive to the erosion angle. When the erosion velocity increased to 31 m/s, the maximum reduction in ultimate tensile stress reached 25.0 %, and the reduction of elastic modulus was approximately 8.7 %. Furthermore, after erosion for 50 min, the reductions in tensile stress and elastic modulus increased to 36.1 % and 7.3 %, respectively. The strain in the erosion center area was significantly higher than that in the edge area, and the material strain continued to increase with increased erosion angle, velocity and time. The wind-blown sand erosion could cause the fiber and adhesive damages of GFRP strips and the erosion angle, velocity, and time are critical factors in the mechanical property degradation of GFRP strips. Additionally, the proposed PC-SE-ResNet model effectively predicted residual mechanical properties, offering reliable support for GFRP application in wind-blown sand environments. ● Effects of wind-blown sand erosion on the durability of GFRP strips. ● Wind-blown sand erosion causes the degradation of the mechanical properties. ● Erosion angle, velocity and time could influence the residual mechanical properties. ● Prediction model is established to predict the residual mechanical properties.