This study attempts to predict the mechanical properties of Al 6061 reinforced with various materials such as AI2O3, SiC, B4C, glass, MoS2, bamboo charcoal, and iron ore. UTS and hardness are considered as output responses, and the stir casting parameters such as stirring speed, stir time, temperature and reinforcement percentage are considered as input variables. Predictive modelling with Artificial Neural Network (ANN) methods in MATLAB is done using 46 data sets from previous studies. The architecture of the ANN model is 4-10-2, which consists of four input neurons, ten hidden nodes and two output nodes. The model parameters are trained to minimise the prediction error. The best validation set is attained with ANN and the value is 82.065 at epoch number 25. Regression analysis-based evaluation is carried to report the performance of the model, which shows good fitting for training, testing and validation datasets and regression values are 0.97932 for training, 0.99227 for testing and 0.97189 for validation, respectively with the overall regression value of 0.97899. The predictions made by the ANN are very close to the actual values for both UTS and BHN, demonstrating the ANN's capability to accurately predict the mechanical properties of AMCs based on stir casting parameters and reinforcement types. This model offers a promising tool for optimizing the production of high-performance Al 6061 hybrid composites.
Gugulothu et al. (Fri,) studied this question.