Dimensional analysis and ANN simulation of chip-tool interface temperature during turning SS304

OBRABOTKAMETALLOV Vol. 23 No. 3 2021 MATERIAL SCIENCE EQUIPMENT. INSTRUMENTS 4 namely input, output, and hidden layers as shown in Fig. 6. The input layer has 3 neurons, the output layer has 1 neuron, and the hidden layer has 8 neurons. A feed-forward neural network displays a data set of numeric inputs with a set of numeric targets. The Neural Fitting app of MATLAB Toolbox will help in the selection of data and for the creation and training a network and evaluate its performance using mean square error and regression analysis. A two-layer feed-forward network with sigmoid hidden neurons and linear output neurons is selected in the present study that fi t multi-dimensional problems arbitrarily well, given consistent data and enough neurons in its hidden layer. The network has been trained with the Levenberg- Marquardt backpropagation algorithm. Fig. 6 . ANN architecture to predict chip-tool interface temperature In a neural network, three kinds of samples are used for the training and validation of test data. In the present work, around 70 % of the data (experimental results of cutting tool temperature) is used for training the neural network. The network is adjusted according to its error. Around 15 % of the data is used for validation of the results predicted by the trained neural network. These validation data sets are used to measure network generalization, and to halt training when generalization stops improving. And around 15 % of data is used for testing the results predicted by the neural network. These data sets do not in fl uence on training and so provide an independent measure of network performance during and after training. The next important step is to determine network architecture, i.e., to set the number of neurons in the fi tting network hidden layer. The neurons in the hidden layer are selected by checking the accuracy of the network. The number of neurons on the hidden layer can be changed if the network does not perform well after training. In the present study, the neural network is modeled considering a different number of hidden neurons to obtain better accuracy of the predicted results. In the present study, a better-predicted accuracy of 0.995 has been observed with 8 neurons at the hidden layer. Further, the network is to be trained using either the Levenberg-Marquardt algorithm or Bayesian Regularization, or Scaled Conjugate Gradient algorithm. The Bayesian Regularization algorithm is preferred for small and noisy data sets. This algorithm results in good generalization but requires more time. Scaled Conjugate Gradient algorithm requires less memory and stops automatically when generalization stops improving. However, the researchers have mostly used the Levenberg-Marquardt algorithm for training the neural network. This algorithm is comparatively faster than other algorithms. However, this algorithm requires more memory and training automatically stops when generalization stops improving, as indicated by an increase in the mean square error of the validation samples. Neural network training performance is measured in terms of mean squared error which is the average squared difference between outputs and targets. Lower values are more preferable and in the present work, the best validation performance of 417.9654 was observed at epoch 7. Regression ( R ) values measure the correlation between outputs (predicted values) and targets (inputs). Neural network regression graphs with regression coef fi cients obtained while training the model, during validation, testing, and for the entire data set are shown in Fig. 7, a–d respectively. The values of regression coef fi cients close to 1 for training, validation, testing, and for the entire data set shows that the developed neural network model could be reliably used for predicting chip-tool interface temperature during turning of SS304 for the given tool-workpiece pair. The results predicted by the neural network are shown in Table 6.