Study on Dynamic Splitting Tensile Damage Characteristics and Damage Stage Identification of Steel Fiber–Reinforced Fly Ash Concrete Based on Acoustic Emission

This paper investigates the dynamic splitting tensile damage characteristics of steel fiber–reinforced fly ash concrete (SFRFAC) under various steel fiber contents, fly ash replacements, and loading rates. The Brazilian disc splitting tensile test was combined with the acoustic emission (AE) detection system for the analysis. Then, the crack scale of SFRFAC under different working conditions was analyzed using the wavelet analysis method (wavelet energy spectrum coefficients). In addition, the damage stages of SFRFAC under dynamic splitting tensile load were identified using the convolutional neural network (CNN) model. The test results indicate that the appropriate steel fiber content and fly ash replacement can improve the ductility and energy absorption of concrete, as evidenced by AE characteristic parameters. The scale of cracks can be identified through the analysis of AE signals after wavelet decomposition. Significantly, the CNN performs well in identifying the damage stages of SFRFAC, with recognition rates over 96%, which is more suitable for concrete conditions with low loading rates and high fiber content.