**3.3 Flexural testing**

Three-point bend tests were completed in order to study the flexural properties of the FDM specimens. Following preliminary testing, a three-inch gage length or span between the outermost points was used for all tests (Figure 11). Flexural strengths were found to be greater than tensile strengths for each raster orientation because the specimens are subjected to both tensile and compressive stresses during bending (Riley et al., 2006). In addition, the three-point test configuration results in the measurement of the maximum strength at the outermost fiber of the beam specimens.

Fig. 11. Three-point bend test configuration

The results of the three-point bend flexural tests are displayed in Table 7. The mean ultimate strength value is the highest for the 0° fiber orientation (38.1 MPa), as was the case during tensile testing. The +45°/-45° orientation had the next highest flexural strength values, followed by 45° and then 90° (23.3 MPa) orientations. Consequently, the flexural test results in Table 7 display the same trend as the tensile test results in Table 2. The flexural strength of the 90° raster specimen was only 60.9% of that of the 0° specimen.


Table 7. Flexural test results

Similar to compression testing, however, not all of the specimens ruptured at failure. The 0° raster specimens and the +45°/-45° specimens never fractured, warranting further analysis to be based upon yield rather than ultimate strengths for consistency. A one-way ANOVA was completed and determined that raster orientation had a significant effect on mean flexural yield strengths at the *p* < 0.05 level for the four conditions, with F(3, 16) = 96.44 and p = 0.0001 (Table 8). The coefficient of determination associated with this analysis was R2 = 0.9476.


Table 8. One-way ANOVA results for flexural testing

170 Mechanical Engineering

Three-point bend tests were completed in order to study the flexural properties of the FDM specimens. Following preliminary testing, a three-inch gage length or span between the outermost points was used for all tests (Figure 11). Flexural strengths were found to be greater than tensile strengths for each raster orientation because the specimens are subjected to both tensile and compressive stresses during bending (Riley et al., 2006). In addition, the three-point test configuration results in the measurement of the maximum strength at the

The results of the three-point bend flexural tests are displayed in Table 7. The mean ultimate strength value is the highest for the 0° fiber orientation (38.1 MPa), as was the case during tensile testing. The +45°/-45° orientation had the next highest flexural strength values, followed by 45° and then 90° (23.3 MPa) orientations. Consequently, the flexural test results in Table 7 display the same trend as the tensile test results in Table 2. The flexural strength

(0°) 34.2, 2.6 38.1, 2.3 1549.0, 327.3

45°) 26.5, 0.7 32.2, 0.5 1438.6, 34.7

Similar to compression testing, however, not all of the specimens ruptured at failure. The 0° raster specimens and the +45°/-45° specimens never fractured, warranting further analysis to be based upon yield rather than ultimate strengths for consistency. A one-way ANOVA was completed and determined that raster orientation had a significant effect on mean flexural yield strengths at the *p* < 0.05 level for the four conditions, with F(3, 16) = 96.44 and p = 0.0001 (Table 8). The coefficient of determination associated with this analysis was R2 = 0.9476.

Diagonal (45°) 21.3, 0.2 25.7, 0.6 1250.0, 36.1 Transverse (90°) 20.8, 0.9 23.3, 1.6 1269.7, 149.6

*Mean Ultimate Strength (MPa)* 

Std Dev Std Dev (MPa), Std Dev

*Mean Effective Modulus* 

of the 90° raster specimen was only 60.9% of that of the 0° specimen.

*Mean Yield Strength (MPa),* 

**3.3 Flexural testing** 

outermost fiber of the beam specimens.

Fig. 11. Three-point bend test configuration

*Raster Orientation* 

Longitudinal

Default (+45°/-

Table 7. Flexural test results

Post hoc analysis further indicated a significant difference between all paired mean comparisons other than that of the 45º raster condition (21.3 MPa) in comparison to the 90º raster condition (20.8 MPa). These flexural strength results further confirm that the raster orientation of the FDM specimens contributes to directionally dependent performance. The specimen fracture patterns for the 45º and the 90° specimens were similar to those described for the tensile testing. In contract, the 0º and the +45°/-45° specimens never fractured during three-point bend testing, but retained some degree of permanent deformation.

Examination of the fracture surfaces of those specimens that broke into two pieces, i.e. the 45° raster specimens and the 90° specimens, revealed that failure initiated on the side of the part that was under tension loading. As fracture began, the specimen initially remained together by unbroken fibers on portion of it that was in compression. Crack propagation along load direction was erratic and not uniform. This is apparent in Figure 12, which displays clusters of fibers that have bent and then ruptured individually in a catastrophically brittle manner.

Fig. 12. SEM image of the fracture surface of a 45° raster specimen after flexural loading

Specimens with 0° raster orientation will have fibers that are able to offer more resistance to bending because they are parallel to the bending plane. There is more fiber length over which the load can be distributed. As the raster angle increases to 45° or 90°, the fiber inclination relative to the plane of bending produces rasters with smaller lengths. This results in a net decrease in the ability of the specimen to resist the load. This effect is observed in Figure 13 where the 90° raster specimen shows little evidence of bending. The bottom of the specimen shows a large flat area initially affected by the failure of several

Anisotropic Mechanical Properties of ABS Parts Fabricated by Fused Deposition Modelling 173

Impact tests were completed on 10 specimens with each of the four raster orientations. The mean impact energy results are displayed in Table 9. The absorbed energy was the highest for the longitudinal (0°) fiber orientation (2.989 J/cm) and the lowest for the transverse (90°) orientations (1.599 J/cm). The 45º and +45º/-45º default specimens broke with mean impact

> **Standard Deviation**

**Fracture Type** 

2.27268

2.5143

2.7559

resistances between those of the 0° and 90° specimens.

**Energy (J/cm)** 

 Longitudinal (0º) 2.991 0.103 Hinged Diagonal (45º) 2.339 0.483 Hinged & Complete Transverse (90º) 1.599 0.014 Complete Default (+45º/-45º) 2.514 0.338 Hinged & Complete

The relative impact strengths of the four raster orientations correlated well with the tensile strength results. In addition, the variation of the impact strengths was smallest for the transverse orientation (Figure 15), coinciding with the variation of the tensile test results.

> 95% CI for the Mean **Interval Plot of Impact Energy**

A one-way ANOVA was conducted to compare the effect of raster orientation on mean impact energies in 0º, 45º, 90º, and +45º/-45º conditions. There was a significant effect of raster orientation on impact energies at the *p* < 0.05 level for the four conditions, with F(3, 36) = 37.23, p = 0.0001. Post hoc comparisons using indicated that the mean impact energy for the 45º diagonal condition (2.339 J/cm) did not significantly differ from that of the +45º/- 45º condition (2.514 J/cm), applying a 95% confidence interval (Table 10). All other paired

1.9931

2.3390

2.6848 2.9166

2.9906

**0 45 90 +45/-45**

1.5886 1.6089

1.5987

**Raster Orientation**

**Raster Orientation Mean Impact** 

3.0646

Fig. 15. Interval plot of impact test results

Table 9. Impact test results

**3.25**

**3.00**

**2.75**

**2.50**

**Joules/cm**

**2.25**

**2.00**

**1.75**

**1.50**

Fig. 13. SEM image of fractured 90° flexural specimen magnified (a) 25X and (b) 400X

rasters, from which the crack then splits to the right and left and eventually climbs, as seen in the shear failure of individual rasters in layers. Upon closer microscopic examination, it was observed that individual rasters showed shear failure with a clearly defined exaggerated shear lip on the top of each fiber; something that was not observed in the analysis of 90° raster specimens that failed in tension testing (Figure 7). In both the 45° and the 90° raster specimens, there is little localized plastic deformation before failure initiates and fibers begin to break.
