PSI - Issue 18

Valerii Matveenko et al. / Procedia Structural Integrity 18 (2019) 12–19 Author name / Structural Integrity Procedia 00 (2019) 000–000

14

3

the expected defect. It is usually impossible to satisfy these conditions. In addition, it is necessary to take into account that a change in the single sensor measurement with varying loads in time does not necessarily indicate the defect appearance. The concept of the proposed method for recording the appearance and development of defects is as follows. Let one of the options of external loads take place for the object under consideration.         1 1 2 2 , , , , , , , , , . j j n n i i i i i i i i p x t k p x t k p x t k p x t k       (1) Here 1, 2, , j n   − external load option, i x − the coordinates components of points on the surface of the body, t − time, j i p − the vector components of external forces in the corresponding coordinate system   1, 2,3 i  , j k − constants. It should be noted that in practice a significant number of objects are subject to a limited number of load options. The processes of elastic linear behavior of the material were considered. In this case for each option of external loads a set of ratios of the corresponding components of the strain tensor at a given number of points can be defined. Here , j j p q   – the strain tensor components at points p and q for j – load option, k – number of controlled points, j pq a – strain ratios. If the integrity of the controlled volume of the material is preserved, then the set of ratios for each of the load options will remain unchanged for different values j k . The defect appearance, the zone of which captures one of the points s , will cause at this point a local change in the strain and a change in the ratios j j s q   , which will indicate the appearance of a defect in the vicinity of the point s . For the practical application of this technique, it is necessary to limit the amount of sensors for a reasonable number. For this purpose, it is proposed to perform a preliminary numerical simulation of the stress-strain state of the object under study. Based on the simulation results, the scheme of the sensors location is determined with their mandatory placement in the stress concentration zones and moderate stress zones. Further, in the monitoring process for each load option, the initial strain ratio j pq a is recorded based on the data of different sensors. If over time or with changing loads, some of the relationships change, then these changes, tied to a specific sensor, will indicate the appearance and j p j pq j q a    , 1, 2, , , , p s k    , 1, 2, , , q p p s k      (2)

development of defects in the area corresponding to that sensor. 3. Practical implementation of damage detection algorithm

The considered object of study was a rectangular glass fiber reinforced plastic (GFRP) plate marked as VPS-48. Sample sizes were 250  50  3 mm. Optical fiber has a diameter 0.124 mm and protective polyimide coating with thickness 0.012 mm. To measure the plate strain, the optical fiber was glued to the plate surface. The mechanical characteristics of GFRP, optical fiber, protective coating and glue are shown in Table 1 and 2. The samples were tested for uniaxial tension on a universal testing machine Shimadzu AG-X Plus. Uniaxial tension was performed in an elastic zone with a load of up to 20 kN with minute holding at loads of 5 kN, 10 kN, 15 kN, 20 kN and subsequent unloading. Loading rate was 2 mm/min.

Table 1: Mechanical properties of VPS-48. X E , GPa Y E , GPa Z E , GPa XY 

XY G , GPa

YZ G , GPa

XZ G , GPa

YZ 

XZ 

22 3 The deformation control from the machine was driven by an optical extensometer. Three optical fiber lines with fiber Bragg gratings (s1, s2, s3) were glued to the surface of the sample to register transverse strains and one line with 22 8.7 0.14 0.14 0.14 3 3