Prediction of chatter phenomenon in the internal turning process using the results of experimental tests and artificial neural network

Machining is a complex process that many variables determine the result. Among them, tool vibration is a critical phenomenon because it causes unacceptable dimensional errors on the work pieces and a sharp reduction in tools life. Whereas the dynamic of the tool and the work piece depends on many parameters, it's very difficult to prediction tool vibration by using relationship and formulas. In this research, internal tuning experiments done with various parameters. Output of experiments shows occurrence or non- occurrence of chatter phenomenon by finishing surface shape, Fast Furrier transform and Spectral density of acceleration signal. Then, by writing a program with Matlab software, using different artificial neural networks, various processes are performed to make the program with highest efficiency between them and using it to predict chatter phenomenon this specific turning machine. Finally, for the validation of results, tests have been designed and planned. The results show that the predication of chatter phenomenon with an acceptable percentage is done correctly.