A Tool Wear Monitoring Method Based on WOA and KNN for Small-Deep Hole Drilling

Abstract

The wear degree of twist drill affects significantly the quality and efficiency of small-deep hole drilling. A monitoring method for tool wear degree based on the signals of sound and current is proposed in this paper, and five kinds of tools with different wear grades are analyzed by using this method. The statistical characteristics of the sound and the current in the time-frequency and psycho-acoustic domains are used to extract the features of drill bits in the proposed method, and the whale optimization algorithm (WOA) is used to optimize features. Finally, the five classifications of twist drill wear degree are realized by K-Nearest Neighbors (KNN). The experimental results show that the combination of the sound and the current can accurately achieve the classification of tool wear, and the recognition accuracy can reach 100%, which can also meet the monitoring requirements of small-deep hole drilling.