علي علي اكبرلو

Friction stir welding is an advanced welding process that has been created in order to weld light weight and high strength materials such as aluminum, magnesium and titanium, which are difficult to weld with conventional welding methods, in order to reduce energy consumption, the effect of greenhouse gases and Reduce carbon dioxide footprint in various industries. Due to the fact that friction stir welding is a complex process and it seems difficult to predict the final properties of a set of welding parameters such as hardness profile, which affect the quality of welding, In this research, after collecting 1405 data from the hardness measurement process of friction stir welding process, to investigate the effective input parameters of friction stir welding on predicting the hardness profile with the help of machine learning models such as Linear Regression, Support Vector Machine, The K Nearest Neighbor, Decision Tree, Random Forest, Gradient Boosting and Multi-layer Perceptron Neural Network have been discussed, then finally after eva‎luating the models, the Random Forest model can perform the best with the help of the effective input parameters selected by the mutual information statistical method

جوشكاري اصطكاكي اغتشاشي , پروفايل سختي , يادگيري ماشين

friction stir welding , hardness profile , machine learning

Ali Aliakbarli

Dr. Mohammad Reza Mohammad Aliha