Prediction of surface roughness and diameter deviation in drilling ti-6al-4v alloy using artificial neural networks and Taguchi's design of experiments
Abstract
[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT REQUEST OF AUTHOR.] Titanium and titanium alloys are excellent materials for aerospace applications owing to their high strength to weight ratio and excellent corrosion resistance. Drilling is an essential process in the structural frames of an aircraft and contributes 40 to 60% of the total material removal operations. Two main attributes of drilled-hole quality are surface roughness and diameter deviation. Major factors that affect the job quality are spindle speed, feed rate and thrust force. This research is focused on the effects of these parameters in drilling Titanium alloy (Ti-6Al-4V) with carbide tool and predicting the quality characteristics using ANN. Experiments were conducted with five different levels of speed and feed using Taguchi's design of experiments. Instead of having to test all possible combinations (such as the factorial design), the Taguchi method tests pairs of combinations resulting in collection of the necessary data to determine which factors most affect quality with a minimum amount of experimentation. The process outputs: surface roughness, holes' accuracy, force and torque were then fitted to 3-D surfaces with the output on the Zaxis and the speed and feed on X and Y axis respectively. Process contour maps were then derived for each output parameter giving an outline of the effects of feeds and speeds on each output and the optimum process conditions were found for each parameter. Signal-to-noise ratio analysis was conducted and results are plotted. Artificial Neural Network (ANN) toolbox from MATLAB is implemented for prediction of drilled-hole quality characteristics (surface roughness and diameter deviation). The experimental results are used to train, validate and test the artificial neural network. Performance of the neural network designed predicted the quality characteristics with reasonable degree of accuracy compared with experimental measurement.
Degree
M.S.
Thesis Department
Rights
Access to files is limited to the University of Missouri--Columbia.