Damage Characteristics And Damage Tolerance Of Fiber-reinforced Composites Under Multiple Low-velocity Impacts

Fiber-reinforced polymer(FRP)composites are widely used in aerospace,transportation,marine,and other fields,due to their excellent mechanical properties.This type of material is typically subjected to complex low-velocity impact environments such as hail impacts,runway debris impacts,and impacts during production processes(such as "tool drops").Due to the weak loadbearing capacity of composites in the transverse direction and their great damage sensitivity to low-velocity impacts,the accumulation of damage caused by multiple low-velocity impacts will inevitably reduce the bearing capacity of the material structure and ultimately lead to failure.However,multiple low-velocity impacts often occur at different positions during actual service.Due to the specific distance between impact positions(DBIP),adjac ent impact damage modes are coupled with each other,resulting in complex and diverse forms and mechanisms of damage and failure,which poses a huge challenge to the life prediction and application of composite structures.Therefore,this paper takes carbon/glass hybrid laminates as the research object,following the principle of simplicity to complexity,first studying the damage behavior of FRP composite s under a single low-velocity impact,then conducting research on the impact resistance and after-impact damage tolerance of multiple low-velocity impacts at different positions,and finally combining machine learning methods to real-time predict the CAI strength of laminates subjected to multiple impacts at different positions.In purpose to improve the rapid and accurate prediction of damage and failure behavior of composites under multiple impact loads,and prevent irreversible disasters caused by accidental impacts.This paper has carried out the following four aspects of research work:The damage mode and failure mechanism of FRP composite s under a single low-velocity impact should be clarified before studying the damage characteristics and damage tolerance of FRP composites under multiple lowvelocity impacts.Therefore,a series of low-velocity impact,ultrasonic C-scan and post-impact compression tests are carried out on carbon/glass composite laminates under different impact energies using a combination of experimental and numerical simulation methods.A single impact "integrated" computational damage model is proposed to deeply explore the comple x failure modes and failure mechanisms of composite laminates.The effects of different impact energies on the impact resistance,impact damage and compression dama ge failure mode of composite laminates are discussed,and the overall mechanical response and damage distribution characteristics of the model are verified by experimental results.In addition,the validated "integrated" finite element model was used to optimize the design of different configurations of carbon/glass composite laminates,providing guidance for engineering applications.An experimental study on multiple low-velocity impact damage and CAI failure behavior of carbon/glass hybrid laminate at different impact positions,that is,the different distances between impact positio ns,is carried out.The lowvelocity impact test,ultrasonic C-scan test,and post-impact compression test of laminates under four impact distances and three impact energies ar e carried out to comprehensively analyze the mechanical response,damage mode,a nd damage evolution and failure mechanism of post-impact compression under multiple low-velocity impacts at different positions.The results show that the mechanical response and damage morphology of different position impacts are induced synergistically by the impact spacing and impact energy.The coincidence degree of low-velocity impact response curves and delamination damage areas are the direct evidence reflecting the interference degree of the two impacts.CAI strength mainly depends on the impact damage mode caused by different DBIP and impact energy.The interference between two impacts is positively correlated with the impact energy when the impact distance is small.As the DBIP increases to a certain range,there is almost no damage interference between the two impacts,which is no longer related to the impact energy.An "integrated" damage calculation model for multiple impacts is established based on the study of singl e-impact damage behavior to simulate the low-velocity impact test and post-impact compression test of carbon/glass hybrid laminates at different positions.The numerical results have a good correlation with the experimental results,which proves the effect iveness and rationality of the proposed numerical model.The parametric study of multiple impact damage behavior is conducted by using the verified prediction model,and the effects of impact spacing and impact angle on the mechanical response,CAI strength and sudden failure position of laminates are evaluated.The results show that the maximum central displacement and absorbed energy of laminates change relatively significantly under different DBIP.The main damage modes of CAI specimens are different,which leads to the final failure behavior of laminates at different impact positions.In addition,the maximum central displacement of laminates is more sensitive under different impact angles.When the ABIP is in the range of 15°–75°,the laminates have fracture failure behavior at both impact positions.This is because the cracks caused by the two impacts spread and confluence in the CAI test compression process,resulting in oblique cracks in the plane.The impact parameters of different combinations are s imulated according to the verified finite element model,and the database of multiple low-velocity impact parameters and CAI strength under different impact positions is established.The artificial neural network model(ANN)and the e Xtreme Gradient Boosting Tree(XGBoost)model are used to predict the damage tolerance of carbon/glass hybrid laminates subjected to multiple impacts at different impact positions,that is,different DBIP,respectively,and the performance of the two models is compared in detai l.The results show that the ANN model with different layers has different prediction effects,and the performance is the best when the model layer number is 6.The XGBoost model is more accurate than the ANN model and can realize high-precision prediction with small data.In addition,the contribution of different characteristic parameters to the output result,that is,the change of CAI strength,is analyzed by Shapley Additional Explanations(SHAP)method.Among them,the impact energy has the greatest influence on the strength of CAI,which is the most critical factor to measure the strength of CAI.The influence of DBIP on the strength of CAI is second only to the impact energy,and the absolute value of SHAP is about 30% of the impact energy,which sho ws that when the number of impacts is greater than 1,the influence of different DBIP on the reliability of laminates must be considered.