The effects of open holes on the fracture behaviors and mechanical properties of glass fiber mat composites

2.1 Materials

A glass fiber chopped strand mat (Nitto Glasstex Co., Ltd.), with gram weight of 450 g/m² was used as reinforcement. The fabric was made of continuous glass fiber bundles, chopped into segments 50 mm in length. The chopped glass bundles were distributed in random directions, as shown in Figure 1.

Figure 1

Glass fiber chopped strand mat.

Unsaturated polyester resin (Showa: RIGORAC 150HRBQNTNW, Showa Denko K.K. Co., Ltd., Tokyo, Japan) was used as the matrix. The resin was mixed with the hardener MEKPO (PERMEK N; NOF Corporation, Osaka, Japan) in a ratio of 100:0.7. Table 1 shows the properties of the UP as measured in lab by tensile test, with a cross-head speed of 1 mm/min.

Composites were created using three, five and seven layers of glass mat through hand lay-up molding. Composite thickness was controlled by spacers, to obtain different V_f in the range of 18%–37%, (determined by burning method according to JIS-K7052). After curing the composite at room temperature for 24 h, the specimens were heated in an oven at 100°C (for 2 h) as a post-cure.

2.2 Specimen preparation

Firstly, smooth specimens with sizes of 20 mm×150 mm were attached to aluminum plate (25 mm in height) and tested to investigate the material constants (such as Young’s modulus).

Furthermore, the notched properties of GMC were also investigated on notched specimens, where a hole was drilled into the center of the specimens. The dimensions and drilling method are illustrated in Figure 2. The wood under the specimens was used to protect the materials from delamination during the drilling process.

Figure 2

Dimension and preparation of specimens.

In order to avoid the effects of micro-damage around the hole caused by the drilling process, the same 10 mm drill was used to drill all holes, with a speed of 2300 rad/min, to ensure that all the notched specimens had holes in similar conditions. Different widths, including 20 mm, 30 mm, and 50 mm, were cut to obtain different width-to-diameter (w/d) ratios, specifically 2, 3, and 5. For the specimens with w/d ratios of 2 or 3, the total length is 150 mm, and the tabs extend 25 mm on both sides, while, for w/d=5 specimens, the length and the tabs are 250 mm and 50 mm, respectively.

2.3 Tensile test

Tensile testing was performed using an Instron universal testing machine (Instron Japan Co., Ltd., Kawasaki, Japan), at a speed of 1 mm/min. Additionally, a strain gauge (Kyowa KFG-10-120-C1-11, Kyowa Electronic Instruments Co., Ltd, Tokyo, Japan) was employed to measure the strain of the smooth specimens during the testing process. For notched specimens, notched strength was calculated using formula (1). The notched strength and smooth strength were compared and analyzed with regard to V_f.

(1)σN=F(W-D)T (1)

Where σ_N is notched strength, the mean strength of the notched specimens; F is maximum load; W is the width of the specimen; D is the diameter of open hole; and T is the thickness of the specimen. The coefficient of variation (C_V) is defined as the ratio of the standard deviation σ to the mean μ. It can be calculated using formula (2):

(2)CV=σμ (2)

C_V is a normalized measure of the dispersion of a set of data. A camera system with a frame rate of 30 FPS was used to capture crack propagation during the tensile test process.

3.1 The tensile test results of smooth specimens

The tensile modulus and strength of smooth specimens are shown in Figure 3A and B, respectively. The trend line shown was drawn according to the least square method. At the same time, the trend line equation and correlation coefficient (R) were calculated. The correlation coefficient was found by equation (3).

Figure 3

Tensile test result of smooth specimens.

(3)R=∑i=1n(xi-x¯)(yi-y¯)[∑i=1n(xi-x¯)2][∑i=1n(yi-y¯)2] (3)

The correlation coefficient gives the degree to which a line generated by the least square method fits the original data. If x and y have a strong linear correlation, R² is close to 1. The results show that both the Young’s modulus and strength were increased with the V_f of glass fiber.

3.2 The tensile test results of notched specimens

Notched strength is shown in Figure 4. Each size of notched specimen is represented with a different shape, based on width. From the trend line described, we can observe that notched strength decreases with the increase of the width-to-diameter ratio. According to the slope of the trend line equation, specimens with lower w/d ratio gain notched strength more easily when V_f is increased.

Figure 4

Tensile test result of notched specimens.

The comparison between smooth strength and notched strength for the specimens with closed V_f but different laminated plies is shown in Figure 5, included 3, 5, and 7-ply specimens with a V_f of 31%, 31%, and 29% respectively. For example, GM3 represents specimens with three layers of glass mat. The decreasing ratio of notched strength from smooth strength of each specimen is shown here. We observed that specimens with a higher w/d ratio had lower notched strength when compared to specimens with the same number of layers. For example, in 3-ply specimens, the decreasing ratios were 10%, 13%, and 21% for specimens with w/d ratios of 2, 3, and 5, respectively. The results also illustrate that notched strength decreased more obviously in 7-ply specimens, which have greater thickness. Consider that the decreasing ratios were 20%, 24%, and 27%, all the largest decreases in percentage. This phenomenon suggests that thickness has a strong effect on the sensitivity of specimens with holes. Our data suggests that GMC with higher thickness was more sensitive to an open hole.

Figure 5

Notched strength decreasing of different specimens in dimension.

3.3 Damage process of notched specimens

A 30-mm wide specimen was selected to explain the damage process of notched specimens shown in Figure 6. The first picture illustrates the specimen during the initial tensile process. The second and third pictures are of specimens carrying maximum load and just after, respectively. The specimen in the fourth picture is totally broken. As the applied load increased, some sound could be heard, expected to be the resin fracturing and the fiber debonding from the matrix. As the test process advanced, the fracture sounds became louder and more frequent, and significant cracks appeared at the edge of the hole, slowly extending outward. Damage near the hole became opaque, as seen in the second and third pictures. After reaching the state shown in the third picture, the specimens would be destroyed suddenly and very fast (within 0.03 s).

Figure 6

Fracture process of notched specimen.

3.4 Fracture characteristics of notched specimens

After the tensile test, the fracture section was not trimmed. The typical fracture zone could be divided into two parts, as shown in Figure 7. This typical fracture formation is closely related to the different stages of the tensile testing process.

Figure 7

Fracture characteristics of notched specimen which were determined as parallel and fan shape areas.

During tensile testing, the cracks occurring at the edge of hole and extending to the outside at the first stage make up the parallel area, due to its relatively stable height. After the parallel area forms, the size of the damaged area increases rapidly in the second stage. After the second stage, the damaged area shows a fan-shaped distribution.

To investigate the parallel length relative to the peak load, the length of the parallel area was measured. Each specimen has two values of parallel area length, one for each side. Some results are displayed in Figure 8 as examples. Length was found to increase with the width-to-diameter ratio.

Figure 8

Measurement of parallel area length.

3.5 SEM observation on the fracture surface of notched specimens

The sketch of the different fractures in the notched specimens and the SEM observation results are shown in Figure 9. In the parallel area, most of the fiber was broken near the surface, caused by concentrated stress at the edge of the hole. As the fracture extended to the outside, the stress was distributed, and the width of the fracture area increased. During this period, the stress was not sufficiently concentrated to break the fibers. The cracks developed in a different direction, changing from perpendicular to the specimen plane to parallel to the specimen plane, which caused pulling out or shear fractures in the glass fiber. To confirm these fracture characteristics, more notched specimens were observed. The typical observations are shown in Figure 10 through different magnifications.

Figure 9

Fracture characteristics of notched specimen in parallel area and fan shape area.

Figure 10

Fracture of notched specimen in parallel area (fiber fractures) and fan shape area (pulling out or shearing fracture of glass fibers).

3.6 Characteristic distance

When notched specimens carried maximum load, there were two points on the line through the center of the hole perpendicular to the loading direction. At these points, the stress was equal to the strength of the unnotched material. The minimum distance from the edge of the hole to this point was defined as characteristic distance, according to point stress criterion theory [22].

In this study, the characteristic distance was calculated by finite element analysis (MSC-Marc), for further understanding of the fracture behaviors of open-hole glass mat composite. In the simulation, a triangle element of three points (element type 6), was used. To obtain more accurate results around the hole, the element near the hole was meshed more closely with the specimen. As the chopped glass fiber of the composite was non-woven and randomly distributed, the interlaminar and intralaminar properties were assumed to be identical. The Young’s modulus and Poisson’s ratio were obtained from smooth specimens for each V_f and used in the simulation. The maximum load of the notched specimen was also applied here. The tensile process for notched specimens was treated as an isotropic linear elastic plane strain problem. The maximum load and stress distribution around the open hole under unit displacement were calculated using the simulation. At that point, the actual stress distribution was calculated according to the ratio of calculated maximum load to the experimental maximum load, according to linear elastic mechanics. The results for all notched specimens are listed in Table 2. From the results, we found that the characteristic distance increased with the width-to-diameter ratio, which supports the experimental results, i.e., the characteristic distance is related to specimen dimensions [24].

3.7 Correlation between parallel area length and characteristic distance

Comparing parallel area length and characteristic distance shows that the parallel area length of most notched specimens was rather close to the characteristic distance, as all of the results illustrate in Figure 11.

Figure 11

Correlation between characteristic distance and parallel area length.

Each notched specimen has two values of parallel area length, with value 1 representing the larger length, and 2 the shorter length. Most of the points are close to the line, especially those specimens with low width-to-diameter ratio. According to the relationship between characteristic distance and notch sensitivity, increasing the parallel area length can enhance the resistance of GMC to the stress of an open hole.