Indentation damage identification of carbon fiber composite laminates based on modal acoustic emission and machine learning

Abstract In this paper, a damage signal recognition for carbon fiber composites based on modal acoustic emission is proposed, which can realize the damage classification of acoustic emission signals. First, according to modal analysis results of the acoustic source signal obtained from the pencil lead breakage experiment, digital filters are designed to realize the mode separation of symmetric (S 0 ) and anti‐symmetric (A 0 ) in the acoustic emission signal. Based on the modal characteristics of the damage signals, an algorithm for the recognition of damage signals is established. The accuracy of damage signal recognition in pure matrix cracking and fiber breakage is 97.9% and 96.2%, respectively. Then, the algorithm is used to identify the damage signal during the indentation experiment of the carbon fiber composite laminates and analyze the damage evolution process. Finally, using the machine learning method to locate the damage location during the indentation experiment, the error between the predicted acoustic emission source and the actual results is less than 2.8%. In this study, the classification of damage modes and the location of damaged sound sources during the indentation experiment of carbon fiber composite laminates were realized, which provided a reference for structural health monitoring of composite materials. Highlights AE technique is used to study the indentation damage of CFRP laminates. An algorithm for damage signal recognition based on MAE is proposed. Prediction of damaged location by combining AE and machine learning. Effectively realizes damage signal classification and location prediction.