Surface roughness modeling in high speed turning of Ti-6Al-4V – Artificial Neural Network approach

Intelligent machining methods for Ti6Al4V: A review

“…Through these 53 Material: J-C (equation ( 1)) Modified J-C (equation ( 4)) Damage: J-C (equations ( 6) and ( 7)) Friction: C-T (equation ( 12 High-speed machining 2D Li et al 50 Material: J-C (equation ( 1)) Damage: J-C (equations ( 6) and ( 7)) Friction: C-T (equation ( 12 methods, it is possible to create models that relate the entrance variables with the generated response along with the possibility to create machining predictive models. 90…”

“…The adjustment of weight and bias is made to reduce the error between predicted and the training data. 4,90 The neurons in a NN are organized in layers and there are different structures for NN, one of the best known is the feed-forward NN, its main characteristic is that each layer has full interconnection with the next layer but never with neurons from the same layer. Multi-layered perceptron (MLP) and the radial basis function (RBF) neural network are two feed-forward NN topologies, the difference between them is associated with the calculations in the hidden layer.…”

“…Through these methods, it is possible to create models that relate the entrance variables with the generated response along with the possibility to create machining predictive models. 90…”

“…The adjustment of weight and bias is made to reduce the error between predicted and the training data. 4,90…”

“…Multi-layered perceptron (MLP) and the radial basis function (RBF) neural network are two feed-forward NN topologies, the difference between them is associated with the calculations in the hidden layer. 4,90 For MLP a sigmoidal activation function is applied to the product of inputs and weights, this network is typically trained with backpropagation (BP) algorithm. 90 The RBF neural network has only weights between the hidden layer and the output layer.…”