HOME INDEXING CALL FOR PAPERS JOURNAL POLICY MANUSCRIPT CURRENT ARCHIVES ONLINE SUBMISSION EDITORIAL TEAM
   
TITLE : ARTIFICIAL NEURAL NETWORK MODEL TO PREDICT THRUST FORCE IN DRILLING OF HYBRID METAL MATRIX COMPOSITES  
ABSTRACT :

This paper presents, a neural network based on back-propagation (BP) algorithm with hidden layers are used for the modeling of Thrust force in drilling of hybrid metal matrix composites. Materials used for the present investigation are Al 356- aluminum alloy reinforced with silicon carbide of size 25 microns and mica of size 45 microns which are produced through stir casting route. Experiments are conducted on a vertical CNC machining centre using TiN coated carbide drill of 6 mm diameter. The parameters considered for the drilling experiments are spindle speed, feed rate and wt % SiC. The data for training and testing have been taken from experiments conducted as per Design of experiments. An empirical model has been developed for predicting the Thrust force of Al 356/Sicp-mica composites. The result shows that the well trained neural network model can precisely predict the thrust force in drilling of Al 356/SiC-Mica composites. Validation results reveal that the neural network model is suitable for predicting the thrust force in drilling hybrid composites. It was found that the maximum error obtained in training of ANN system when comparing the experimental results is less than 5.0%. The efficiency of the system can be improved by using more number of data point.

Keywords: Hybrid composites; ANN; drilling; thrust force

 
  Download Full Paper
 
Copyrights ©Sathyabama Institute of Science and Technology (Deemed to be University).
Powered By: Infospace Technologies