An industrial tool wear monitoring system for interrupted turning

Abstract

An effective wear-monitoring system for machine tool inserts could yield significant cost savings for manufacturers. Over the years, various methods have been proposed to achieve tool condition monitoring (TCM), and recently sensor-based approaches for indirectly estimating tool wear have become highly popular. One difficulty with collecting sensory information from machine tools is that the signal-to-noise ratio of useful information about the tool wear is extremely poor. This problem can be overcome by using advanced signal-processing methods and also by fusing the information obtained from numerous sensors into a single modelling or decision-making scheme such as neural networks (NNs). Neural networks are known for their capacity to solve problems effectively in cases where theoretical/analytical models cannot be established. Furthermore, NNs can handle noisy and incomplete data such as that typically obtained from machining operations. Although numerous authors have proposed the NN approach for TCM, various problems still hamper a practical method of applying the technique for industrial use. This paper proposes a technique which should overcome these difficulties. A cost-effective and reliable tool condition monitoring system (TCMS) was developed, utilising the advantages of NNs for a typical industrial machining operation. The operation considered is interrupted turning (facing and boring) of Aluminium alloy components for the automotive industry. The development and implementation of various hardware and software components for the proposed technique are described in this paper. The main advantages of the technique are its accuracy, reliability and cost-effectiveness.