Non-Destructive Evaluation and Damage Determination of Fiber-Reinforced Composites by Digital Image Correlation

Abstract

AbstractThe trend in recent years shows that fiber-reinforced polymer composites (FRPCs) are steadily replacing traditional engineering materials in the aerospace, automotive, and sports equipment industries, where the safety and reliability of structures are of crucial importance. The increasing popularity of FRPCs is explained by their favorable mechanical properties. However, a serious problem of FRPCs is their catastrophic failure under overload and lack of ductile behavior. This issue makes it necessary to develop a non-destructive examination method that can estimate the structural integrity and predict the remaining properties of a composite structure even after its partial damage. In this paper, we propose a Digital Image Correlation (DIC)-based new method, which is able to meet all the requirements of Industry 4.0, thus this method can be automated, is fast, accurate, and excludes the human factor. To prove the efficiency of the proposed, Non-Destructive (NDT)-DIC method, we created specimens with known impact damage, and tested them firstly by infrared thermography to get a basic reference to compare the subsequent results. Then we performed the NDT-DIC tests, where through the practical example, we presented the steps and the logical deduction of the method. Firstly, the test load limit is determined, proven by AE testing to be in the non-destructive range for the examined composite system. In the case of a simple tensile test setup, the obtained result is the principal strain field from which we calculate our indicator, the X value. The X value can indicate the existence of damage by itself, furthermore, a correlation can be established that defines the X value’s relationship with the reduction of strength, based on preliminary experiments. The use of the correlation enables a higher level of health monitoring than the use of the DIC method without this information because the effect of a random impact damage is predictable for the given composite. In our case, the correlation is already strong with a 0.1% strain/elongation test load, but when the test load is increased to 0.15%, the coefficient of determination (R2) slightly increases from 0.979 to 0.997.