A NEW APPROACH OF TOOL WEAR CONDITION MONITORING BASED ON TEXTURE IMAGE RECOGNITION

摘要

Due to the importance of tool wear condition real-time monitoring to the whole of NC machining process, this paper proposes a new approach of tool wear state monitoring. By taking the texture images of machined surfaces and pre-processing them, the characteristic parameters with inner relations to tool wear are extracted. In view of the structure classifier, the LVQ neural network classification model was put forward. Then the relationship between the characteristics data and the degree of tool wear was described with the competition of LVQ neural network. Finally the monitoring and identification diagnosis of tool wear state was indirectly realized. Simulation results show that it is a good way to real-time monitoring tool wear state.