The ubiquity of artificial intelligence in the manufacturing domain draws inspiration for the present article. The implementation of a neural network technique is still a difficult and time-consuming effort for the industry. Prediction of machining variables is a considerable issue that needs to be explored for preventive maintenance of the machine structure and to optimize the surface quality. This work aims at predicting response parameters of the dry turning process for Inconel 825 alloy using deep-cryogenic treated tungsten-carbide insert through artificial neural network technique. Process parameters considered in this work were cutting speed, feed and depth of cut, whereas, surface-roughness, tool-wear, and material-removal-rate were taken as the three response parameters.14 types of training functions were compared based upon their error indices searching for the training function which best suits this work. Artificial Neural Network (ANN) model was developed by taking Bayesian regularization back propagation based training function. The response values predicted by the ANN were in very close approximation to the actual experimental value with the mean square error of only 0.0011?μm2, 39.0882?μm2 and 0.0520?cm6/min2in the prediction of surface-roughness, tool-wear, and material-removal-rate of dry turning process of Inconel 825 using treated carbide tool.
Process modeling for machining Inconel
The ubiquity of artificial intelligence in the manufacturing domain draws inspiration for the present article. The implementation of a neural network technique is still a difficult and time-consuming effort for the industry. Prediction of machining variables is a considerable issue that needs to be explored for preventive maintenance of the machine structure and to optimize the surface quality. This work aims at predicting response parameters of the dry turning process for Inconel 825 alloy using deep-cryogenic treated tungsten-carbide insert through artificial neural network technique. Process parameters considered in this work were cutting speed, feed and depth of cut, whereas, surface-roughness, tool-wear, and material-removal-rate were taken as the three response parameters.14 types of training functions were compared based upon their error indices searching for the training function which best suits this work. Artificial Neural Network (ANN) model was developed by taking Bayesian regularization back propagation based training function. The response values predicted by the ANN were in very close approximation to the actual experimental value with the mean square error of only 0.0011?μm2, 39.0882?μm2 and 0.0520?cm6/min2in the prediction of surface-roughness, tool-wear, and material-removal-rate of dry turning process of Inconel 825 using treated carbide tool.
Log in to your free materialstoday.com profile to access the article.
