Swarup Bag

Accurate prediction of fatigue life in additively manufactured components remains challenging due to the complex interplay of mechanical properties, surface states, and microstructural factors. Correlation analysis enables the identification of key influencing parameters, while deep learning models offer the capability to capture complex temporal patterns associated with fatigue behaviour. This study presents a correlation guided temporal deep learning approach for predicting the fatigue life of AlSi10Mg alloy fabricated using laser powder bed fusion. The data are collected from independent literature from the analysis of V-notched specimens. The mechanical, microstructural, and surface characteristics including yield strength, elongation, hardness, residual stress, and surface roughness data are collected and corelated. The prediction shows elongation as the most influential parameter, whereas traditional strength indicators and surface hardness are statistically insignificant under cyclic loading. Guided by these findings, deep learning models are developed, among which the temporal convolutional network (TCN) achieved the highest predictive accuracy with an R ² of 96.6%. Its ability to model long term dependencies without recurrence proved highly effective for capturing fatigue behaviour. The proposed framework demonstrates the potential of combining statistical insights with temporal deep learning to improve fatigue life prediction in additively manufactured components.

Making thin strips directly from the melt by twin roll casting requires precise control over process parameters to obtain defect-free components. Strategically controlling the kissing point location and the solidification front can mitigate the probability of crack formation. A two-dimensional low Reynolds number k – ε turbulence model is used to identify the location of kissing point for a vertical twin roll casting process. The temperature-dependent viscosity and apparent heat capacity method is employed to imitate the flow of melt and the solidification behaviour. The effect of different roll rotational speeds (20–28 rpm) and inlet melt temperatures (1733–1813 K) on solidified shell thickness and kissing point location is explored. The formation of a laminar sub-layer near the roll surface ensures uniform temperature distribution and high Nusselt number indicates dominant convective heat transfer from the upper roll surface. The shifting of the kissing point location is obvious by increasing roll rotational speed and inlet melt temperature accompanied by reduced solidified shell thickness. A polynomial relation between these parameters offers a practical solution for optimising heat transfer efficiency with direct applicability to industrial twin roll casting operations. This insightful study provides variation in kissing point location with process variables of twin roll casting confronting various dimensionless numbers.

Wire arc additive manufacturing (WAAM) is amongst the emerging technologies of the layer-by-layer deposition process to manufacture the metallic parts. Multi-layer bead deposition using the WAAM process leads to the fabrication of complex products with practical utility. The bead profile of each layer controls the geometry of the final product. However, the melting efficiency and the bead geometry depend on the various process parameters. The primary process parameters affecting the melting efficiency and the bead geometry are the wire feed rate, travel speed, diameter of the wire, and power. Owing to the various complexities during metal deposition, predicting the dimensions at each deposition attribute is not always feasible. Hence, the current work is focused on the utilization of different machine learning (ML) algorithms to understand the relationship between the process parameters, melting efficiency, and bead geometry. The different ML models used for the current work are linear regression (LR), decision tree regressor (DTR), random forest (RF), support vector regression (SVR) and, extra tree regressor (ETR). The ETR is found to predict the melting efficiency with the highest prediction rate of 97.4%, whereas, the SVR and LR predict the bead width and height with the highest accuracy rate of 97.4% and 98.7%, respectively.