PSI - Issue 50

12 R.S. Akhmetkhanov et al. / Procedia Structural Integrity 50 (2023) 11–16 R.S.Akhmetkhanov / Structural Integrity Procedia 00 (2019) 000 – 000 (average radius of the equivalent disk). Figure 1 shows the resulting images (Ganiev S.R., at etc., 2014 ) . From these distributions, we can clearly see how the filler was milled and mixed with the matrix. 3. Methodology The image research method is a computer analysis of the structure of the material based on the selection of the main structural characteristics, such as the distribution of the grain composition of the filler by their size, the total perimeter of the grains, the total mass fraction of the filler, determining the fractal dimension, the position of the static moment of inertia of the image, etc.

Fig. 1. Visual distribution of the filler grains at different times of the wave mixing process: a - 1 minute: b - 120 minutes (filler is marked in black). 4. Results Table 1 shows statistical data on grain size (area) and statistical deviations depending on the mixing process. The mixing process lasted 120 minutes. The data are given in pixels (pk). Filler grains are highlighted by segmentation using the threshold method. A threshold is a feature (property) that helps to divide the searched signal into classes. The thresholding operation consists in comparing the brightness value of each pixel of the image with the given threshold value. The degree of comminution according to the average values of the areas of the filler grains by area size is 3.2 (300.3/93.7) (120 minutes of mixing). Both milling and aggregation of small grains into larger grains occurs during mixing. Figure 1b shows a large number of grains with sizes ranging from 0.3 mm to 1.2 mm, whereas in the first minutes of the process there was a fairly large number of large grains along with small grains.

Table 1. Results of structural analysis of filler by grain size and statistical data.

Time t , minutes

Grain area S z , pk

Median (median value), pk

Standard deviation, pk

Interquartile range ( IQR ), pk

1

300.3 120.2 118.5

28.0 36.0 41.0 40.0 43.0

1356.4

49.8 92.0 76.0 64.0 78.0

15 60 90

239.8 240.2 132.6 145.0

89.2 93.7

120

IQR is a measure of variability based on dividing the data set into quartiles.

Figure 2 shows the distribution of filler grains by the average radius of the equivalent disk, the size is given in millimeters (the material image was conventionally assumed to be 40*50 mm to obtain the grain sizes in mm). The main number of grains is between 0.5 mm and 2.0 mm (fig.2c). The difference is that after grinding the grain size became less than 4 mm, while before grinding it was 12 mm (fig. 2a).