PSI - Issue 17

V.P. Matveenko et al. / Procedia Structural Integrity 17 (2019) 363–370 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

364

2

1. Introduction One of the actual tasks of structural health monitoring (SHM) is to register the appearance of local damages in the structures and assess their development. The importance of solving this problem has increased with the replacement of traditional structural materials with composite materials, in particular, layered polymer composite materials (PCM), which are increasingly used in various fields of production from civil engineering to aerospace products (Staab 1999). The structure of these materials leads to the formation of local damages (defects) often associated with delamination, which represents a significant danger to structures made of composite materials (Adams 2007). One of the basic principles of the registration system of the appearance and development of defects is to ensure non-destructive testing, preserving the integrity of the construction and its structure (Stepinski, Uhl, and Staszewski 2013). The implementation of this task can be achieved on the basis of various approaches. The methods of non destructive testing include various visual methods, the main disadvantage of which is that they do not allow detecting damage located outside the visual inspection zone (Verma, Bhadauria, and Akhtar 2013). Another class of methods, which can be called vibration methods (Fan and Qiao 2011), is based on the interpretation of such mechanical and physical quantities as natural frequencies, vibration modes and parameters of wave processes in a material. An idea of the possibilities of methods using modal parameters (eigenfrequencies and modes of oscillations) is given in the works (Asnaashari and Sinha 2014; Quaranta, Carboni, and Lacarbonara 2016), and methods analyzing wave processes in the work (Raghavan and Cesnik 2007). The advantages and disadvantages of vibration control methods are presented in numerous scientific studies, examples of which are present in the papers (Fan and Qiao 2011; Michaels 2008; Morassi, A. and Vestroni 2008; Zhang et al. 2013). One of the ways to control the appearance and development of defects is associated with the use of information about strains in the material. Currently, one of the most promising sensors for measuring strain is fiber-optic strain sensors (FOSS). In comparison with traditional sensors, they are compact, do not require separate power supply, and they are immune to electromagnetic effects. (Majumder et al. 2008). Numerous papers provide information about FOSS applications at various engineering objects (Ghoshal et al. 2015; Hong et al. 2016; Sierra-Pérez, Torres Arredondo, and Güemes 2016). One of the advantages of FOSS is associated with the real possibility of embedding these sensors in a PCM at the technological stage of manufacturing (Kang et al. 2002; Matveenko et al. 2018; Minakuchi et al. 2013). Among the real applications of FOSS for damage detection, numerous works should be noted, based on the analysis of changes in the spectrum shape of reflected signal during the appearance and development of a defect. Examples of such studies are (Gafsi and El-Sherif 2000; Wagreich et al. 1996). However, it should be noted that this approach to defect analysis can be implemented only when the sensor is located in the defect zone. In this paper, an algorithm for recording the appearance and development of defects based on experimental information about strains in the material and numerical simulation results, which, when using a limited number of strain sensors, make it possible to register changes in strains due to the appearance of a defect, is considered. The algorithm is illustrated when using FOSS on a structurally similar element made of PCM, in which under loading there are defects associated with delamination in the stress concentration zone. 2. Algorithm for damage detection in the material based on strain experimental data Defects registration in a material under quasi-static loading with strain measurements at a finite number of points is based on the use of information about strain changes caused by the defect that has arisen. At the same time, for a reliable conclusion about the defect appearance, two problems must be solved. The first of them is determined by the principle of Saint-Venant (Timoshenko and Goodier 1972), according to which the zone of strain change due to the defect appearance is determined by 3 – 5 of its characteristic dimensions. According to this principle, when there is no additional information on the strains distribution in the material of the monitored object, it is necessary to have a spatial grid of sensors with a cell size of 3 to 5 characteristic sizes of the expected defect. The practical implementation of such a scheme, as a rule, does not make sense. The second task is related to the fact that a change in the strain registered by the sensor can occur not only when a defect occurs, but also when there is a change in external influences on the monitored object. The basis of the proposed method is the following. A variant of the mechanical behavior of the material, which is described by linear physical relations, is