Development of an intelligent system for tool wear monitoring applying neural networks

Purpose: The objective of the researches presented in the paper is to investigate, in laboratory conditions, the application possibilities of the proposed system for tool wear monitoring in hard turning, using modern tools and artificial intelligence (AI) methods. Design/methodology/approach: On the basic theoretical principles and the use of computing methods of simulation and neural network training, as well as the conducted experiments, have been directed to investigate the adequacy of the setting. Findings: The paper presents tool wear monitoring for hard turning for certain types of neural network configurations where there are preconditions for up building with dynamic neural networks. Research limitations/implications: Future researches should include the integration of the proposed system into CNC machine, instead of the current separate system, which would provide synchronisation between the system and the machine, i.e. the appropriate reaction by the machine after determining excessive tool wear. Practical implications: Practical application of the conducted research is possible with certain restrictions and supplement of adequate number of experimental researches which would be directed towards certain combinations of machining materials and tools for which neural networks are trained. Originality/value: The contribution of the conducted research is observed in one possible view of the tool monitoring system model and it’s designing on modular principle, and principle building neural network.