With the increasing application of steel materials, the metallographic analysis of steel has gained importance. At present, grain size analysis remains the task of experts who must manually evaluate photos of the structure. Given the software currently available for this task, it is impossible to effectively determine the grain size because of the limitations of traditional algorithms. Artificial intelligence is now being applied in many fields. This paper uses the concept of deep learning to propose a fast image classifier (FIC) to classify grain size. We establish a classification model based on the grain size of steel in metallography. This model boasts high performance, fast operation, and low computational costs. In addition, we use a real metallographic dataset to compare FIC with other deep learning network architectures. The experimental results show that the proposed method yields a classification accuracy of 99.7%, which is higher than existing methods, and boasts computational demands, which are far lower than with other network architectures. We propose a novel system for automatic grain size determination as an application for metallographic analysis.

Quench–polish–quench (QPQ) nitro-carburizing of AISI 4140 steel in a salt bath was performed in this study. Nitro-carburizing in a salt bath enhanced the formation of Fe-nitride on the outer surface layer. Moreover, the oxidizing treatment formed a thin oxide layer decorated on the outermost part of the QPQ-treated sample. The dense compound layer formed after nitro-carburizing in a salt bath consisted of refined granular Fe₃N and transformed to Fe₂N after post-oxidation treatment. Micro-shot peening (MSP) was adopted before QPQ treatment to increase the treated steel’s fatigue performance. The results indicated that MSP slightly increased the thickness of the compound layer and harden depth, but it had little effect on improving the fatigue strength/life of the QPQ-treated sample (SP-QPQ) compared to the non-peened one (NP-QPQ). A deep compressive residual stress (CRS) field (about 200 μm) and a hard nitrided layer showed a noticeable improvement in the fatigue performance of the QPQ-treated ones relative to the 4140 substrates tempered at 570 °C. The ease of slipping or deforming on the substrate surface was responsible for its poor resistance to fatigue failure. The cracking and spalling of the brittle surface layer were the causes for the fatigue crack initiation and growth of all of the QPQ-treated samples fatigue-loaded at/above 875 MPa. It was noticed that fatigue crack initiation at the subsurface inclusions was more likely to occur in the SP-QPQ sample fatigue-loading at 850 MPa or slightly above the fatigue limit.

Gas nitriding was implemented in the current work at a constant nitrogen potential (K_N) of 2.0 for 8 h to enhance the fatigue properties of SCM 440 steel, and the results were compared with those of the substrate tempered at the nitriding temperature (475 °C). Fine particle peening (FPP) prior to nitriding imposed a refined structure and induced compressive residual stress (CRS) in the near-surface peened zone. The fine-grained structure provided numerous paths to enhance nitrogen diffusion inwards during nitriding. The compound layer formed on the nitrided SCM 440 steel primarily comprised a mixture of Fe₃N and Fe₄N; however, the pre-peened and nitrided (SPN) specimens exhibited a higher proportion of Fe₃N and a thicker compound layer than the non-peened and nitrided (NPN) counterparts. In addition, FPP prior to nitriding increased both the case depth and the magnitude of the CRS field compared with nitriding alone. The fatigue limits of the substrate (SB), NPN, and SPN samples were approximately 750, 1050, and 1400 MPa, respectively. Gas-nitriding at 475 °C significantly improved the fatigue performance of SCM 440 steel. Moreover, pre-peening prior to nitriding further enhanced fatigue strength and life of the treated SCM 440 steel by introducing a deeper case depth and higher CRS field. Multiple cracks initiation at the outer surface of the SB sample accounted for its lowest fatigue limit among the tested samples. Surface microcracks and pits on the surface of the NPN specimen would be crack initiation sites and harmful to its fatigue resistance. These surface dents were considered to be responsible for fatigue crack initiation in the SPN specimens. Therefore, polishing after nitriding to reduce surface roughness and/or microcracks was expected to further increase the fatigue resistance and the reliability of nitrided SCM 440 steel.