PSI - Issue 68

Hamidreza Rohani Raftar et al. / Procedia Structural Integrity 68 (2025) 1066–1073 Hamidreza Rohani Raftar et al./ Structural Integrity Procedia 00 (2025) 000–000

1071

6

and environmental factors on crack growth rates, emphasizing the need for deeper exploration and validation across varied material systems and operating conditions. The model demonstrated a satisfactory level of accuracy, highlighting the effectiveness of machine learning in predicting hydrogen-assisted fatigue behaviour in pipeline steels.

Fig. 2. Overview of Data Analysis and Predictive Model Results (a) predicted da/dN using bagged trees, (b) accuracy assessment of the training procedure through regression analysis, (c) Accuracy assessment of the test procedure via regression analysis, (d) feature importance scores derived from the predictive model. 4. Conclusions The machine learning approach was used based on experimental results, demonstrating its capability to predict hydrogen embrittlement in pipeline steels. However, achieving a deeper comprehension of fatigue phenomena in hydrogenous environments necessitates a thorough and detailed study of material characterization using machine learning techniques. The applied methodology emphasizes the importance of exploring various factors, such as alloy elements, to enhance understanding and enhance the fatigue strength of materials in hydrogenous settings. This study supports incorporating additional parameters as inputs to advance comprehension of hydrogen environment embrittlement. Future investigations could concentrate on examining the interaction among multiple parameters that influence hydrogen-assisted fatigue in pipeline steels and leveraging machine learning for material characterization. This approach holds promise for guiding strategies to mitigate hydrogen-assisted fatigue in pipeline materials. Below is the finding information related to this research: