*2.2. Material Structure Analysis*

Metallographic analysis was performed on the non-tested spare samples for all thermoplastic types. The analysed samples contain only minor porosities (<20 μm in the radius and <200 μm in the flat parts). The void size and quantity were not sufficient to visualize a negative effect on interlaminar strength in cases where the specimens were exposed to a quasi-static loading. No significant deviations between the sets were observed. A typical cross section before testing of PAEK thermoplastic is shown in Figure 4.

**Figure 4.** Example of the cross section before testing: PAEK.

#### *2.3. Test Method*

The objective of the curved beam strength test is to determine the strength characteristics of the composite material in the out-of-plane (z) direction. Radial tensile stress in this direction of the composite (through the thickness of the material) is induced in the curved region of the test specimen when bending is applied. The bending load is applied using a four-point bending fixture with two pairs of cylindrical supports with different span lengths (lt and lb).

Before the tests started, a comparison of the two most used test methods was performed (ASTM D6415 [2,3]—the standard test method for measuring the curved-beam strength of fibre-reinforced polymer-matrix composites—and AITM 1-0069 [24]— determination of curved-beam failure loads). This method of loading induces a constant bending moment in the curved region of the specimen. The main motivation was to compare the effect of the spans. Three configurations were prepared: the first was per the ASTM standard, where fix values of span were used (lt = 75 mm and lb = 100 mm); the second configuration used a modified ASTM (ASTM mod) span (lt = 45 mm and lb = 75 mm); and the third was per the AITM standard. In this method, the spans are calculated based on the sample geometry in Equations (1)–(4). Based on these calculations, lt = 26.4 mm and lb = 40.6 mm were set.

$$\mathbf{l}\_{\mathbf{t}} > 2 \cdot \left( \left( (\mathbf{R}\_{\mathbf{i}} + \mathbf{t} + \frac{\mathbf{D}}{2}) \cdot \sin(\varphi) + \left( \frac{\mathbf{t}}{4} + 1 \right) \cdot \cos(\varphi) \right) \right. \\ \left. \pm 0.5 \tag{1}$$

$$\mathbf{d}\_{\mathbf{t}} > 2 \cdot \left( \left( (\mathbf{R}\_{\mathbf{i}} + \mathbf{t} + \frac{\mathbf{D}}{2}) \cdot \sin(\varphi) + \left( \frac{\mathbf{t}}{4} + 1 \right) \cdot \cos(\varphi) \right) \right. \\ \left. \pm 0.5 \tag{2}$$

$$\mathbf{l\_b > l\_t + t + 10 \pm 0.5} \tag{3}$$

$$\mathbf{l\_b} > \mathbf{l\_t} + \mathbf{t} + \mathbf{20} \neq \mathbf{0.5} \tag{4}$$

In these equations, lt denotes the span of the top fixture, lb is the span of the bottom fixture, Ri is the inner radius, t is the thickness of the sample, D is the roller diameter and *ϕ* is the angle from the horizontal of the sample legs.

Tests were performed on a static load machine Instron 55R1185 (Norwood, MA 02062- 2643, 825 University Ave, USA) with an installed load cell with a capacity of ±10 kN and with control system Instron K5178. Recording of the force, displacement and extensometer data was ensured by the software Bluehill 3. The test setup is shown in Figure 5. A test specimen was placed on the bottom cylindrical bars. Then, the extensometer Instron 2620- 604 with a base of 50 mm was installed. The extensometer recorded the axial displacement between the upper and lower parts of the fixture. The specimen was loaded by a constant crosshead speed of 2 mm/min and the test ended when the loading rapidly decreased (approximately a 30-percent drop).

**Figure 5.** Curved-beam strength test.
