Сompilation of a Set of Informative Features for Neural Network-Based Determination of Minimum Vibration of Cutting Tools

The article is devoted to the development of neural network models of digital counterparts of metal cutting processes on metal-cutting machines of the turning group. The article considers an example of a study of vibration signals removed from a cutting tool. The results of a series of experiments conducted on a 1K625 lathe made it possible to form a sample of eleven informative features, five of which turned out to be enough to form a list of input signals of the projected neural network. The designed neural network, after training using the error back propagation algorithm, allows us to estimate the proximity of the current processing mode to a certain local optimum in terms of cutting speed.