Intelligent prediction using AI-based modeling and optimization of surface roughness in Al7049 end milling with coconut oil under minimum quantity lubrication.

Metal cutting researchers attempt to develop technology for metal cutting and use of cutting fluids have been continuous due to increasing demands for high productivity. This work reports on the effect of coconut oil under minimum quantity lubrication (MQL) during end milling of AL7049. Dry, wet and MQL approaches are all used to conduct the experimental studies. In contrast to dry and wet machining, it is apparent that MQL generates surfaces with less roughness. A multilayer perceptron (MLP) prediction model was developed considering the real-time dataset with three inputs, and an output that led to prediction of surface roughness. It was analyzed to determine the appropriate activation function and observed that ReLU activation function outpaces sigmoid and tanh and adapted to the proposed solution. K fold cross-validation was done for the developed MLP model with backpropagation to substantiate high accuracy in predictions than KNN and linear regression. The MQL will be good alternative to wet machining and also environmentally friendly machining solution. [ABSTRACT FROM AUTHOR]