Issue 78

A. Sancak et alii, Fracture and Structural Integrity, 78 (2026) 186-206; DOI: 10.3221/IGF-ESIS.78.11

A one-way ANOVA performed across the eleven experimental groups indicated a statistically significant overall difference in absorbed impact energy, F(10,22)=4.509, p=0.0016. This result indicates that the mean absorbed impact energy values of the eleven groups were not all equal. Considering the limited number of specimens per condition and the considerable scatter observed in several groups, the group mean values were interpreted cautiously. To further evaluate the scatter in the Charpy impact results, the coefficient of variation (CV) was calculated for each group using the ratio of the standard deviation to the mean absorbed energy. The CV values confirmed that several groups exhibited substantial scatter. The highest CV values were obtained for the pre-notched seawater specimens tested at 0 °C, 65.68%, the substitute seawater specimens exposed for 10 days and tested at 0 °C, 64.64%, and the seawater specimens exposed for 10 days and tested at 0 °C, 57.94%. The uncorroded reference group tested at 23 °C also showed a high CV of 53.52%, indicating that substantial scatter existed even before corrosion exposure. In contrast, the reference groups tested at − 20 and − 40 °C and the substitute seawater group exposed for 30 days showed relatively low CV values, between approximately 12 and 14%. The high CV values show that several experimental conditions exhibited substantial specimen to-specimen variability. As a result, differences between group means should be interpreted together with the corresponding within-group scatter rather than being attributed solely to the nominal experimental condition. No individual Charpy value was excluded as an outlier. Since each experimental condition consisted of only three specimens, formal outlier rejection was not considered statistically robust. Therefore, all values were retained in the calculation of the mean, standard deviation and coefficient of variation, and were treated as part of the experimental scatter. The reference results demonstrate that the investigated AISI 1008 chain steel is highly sensitive to test temperature under impact loading. Although AISI 1008 is a low-carbon steel, the combination of a sharp V-notch, dynamic loading and reduced temperature resulted in a pronounced decrease in absorbed energy. This behaviour shows that the ability of the ferrite– pearlite steel to accommodate plastic deformation during impact loading decreased markedly as the test temperature was lowered. Another important feature of the reference condition is the large scatter observed at 23 °C. The individual absorbed energies at room temperature were 60.91, 78.59 and 22.19 J. Since these specimens were not exposed to seawater corrosion, the observed within-group scatter cannot be attributed to environmental degradation. It suggests that additional sources of specimen-to-specimen variability, such as extraction position, local ferrite–pearlite distribution or other material heterogeneities, may have contributed to the energy differences. However, the present near-fracture cross-sectional observations do not establish a direct causal relationship between these features and the absorbed impact energy. Effect of seawater environment and immersion duration The effects of seawater and substitute seawater exposure were evaluated by comparing specimens immersed for 10 and 30 days and tested at 0 °C. Seawater was collected from the Üsküdar coastal region of İ stanbul, whereas substitute seawater was prepared according to ASTM D1141. For these groups, the Charpy V-notch was machined after corrosion exposure. After 10 days of immersion, the average absorbed energy was 14.23 ± 8.25 J for seawater and 17.41 ± 11.26 J for substitute seawater. These mean values are slightly higher than the uncorroded 0 °C reference value of 13.37 ± 5.41 J. However, this result should not be interpreted as an improvement caused by corrosion exposure. Both 10-day groups exhibited large scatter. In particular, the substitute seawater group contained one high-energy specimen, 30.41 J, whereas the other two specimens absorbed approximately 11 J. Therefore, at 10 days, specimen-to-specimen variability dominated the average response, and the mean value alone was not sufficient to describe the degradation behaviour. After 30 days of immersion, the seawater group tested at 0 °C exhibited an average absorbed energy of 10.62 ± 5.55 J, which is approximately 20.6% lower than the uncorroded 0 °C reference value. The substitute seawater group tested under the same exposure duration and temperature exhibited an average absorbed energy of 11.02 ± 1.36 J, corresponding to an approximately 17.6% reduction relative to the 0 °C reference condition. Thus, after 30 days of exposure, both seawater and substitute seawater groups exhibited lower mean absorbed energies than the uncorroded 0 °C reference group. The numerical trend associated with immersion time was evaluated by comparing the 10-day and 30-day mean values within each exposure medium. In seawater, the average absorbed energy decreased from 14.23 J after 10 days to 10.62 J after 30 days, corresponding to a reduction of approximately 25.4%. In substitute seawater, the average absorbed energy decreased from 17.41 J to 11.02 J over the same exposure interval, corresponding to a reduction of approximately 36.7%. Accordingly, the mean absorbed impact energy decreased numerically from 10 to 30 days in both exposure media. Seawater produced numerically lower mean absorbed energies than substitute seawater at both exposure durations, although the difference was more pronounced after 10 days and became relatively small after 30 days. However, the relatively large scatter, especially in the 10-day groups, indicates that the mean values alone do not fully represent the impact response.

195