Real time implementation of on-line tool condition monitoring in turning

Statistical and Regression Analysis of Vibration of Carbon Steel Cutting Tool for Turning of EN24 Steel Using Design of Experiments

“…Many sensors were reported for tool condition monitoring system, namely; power sensors, vibration sensors, vision sensors, temperature sensors, touch sensors, force sensors, acoustic emission sensors, flow sensors and so on [4,2,5].The role of sensors and their signal interpretation capability is critical in any automation process. The processing and investigation of signals is significant because it will improve production capacity, reliability, reduced downtime and enhanced machining quality [6]. Byrne et al [7] described that 46% of the sensors monitoring systems were fully functional, 16% had limited functionality, 25% were non-functional owing to technical limitations and 13% were switched over to alternate systems.…”

On-Line and Indirect Tool Wear Monitoring in Turning With Artificial Neural Networks: A Review of More Than a Decade of Research

“…Indirect methods, on the other hand, employ parameters which are easier to measure, but the computational effort is usually very high (see, e.g. [46,73,75,130]). Table 1 (representation adapted from [36] and extended) lists different direct and indirect monitoring methods.…”

“…Some authors (see, e.g. [19,76,125,139,147,162]) describe the use of filters (such as band-or low-pass filters) in order to reduce the influence of disturbances like noise or other non-measurable contributions (see Section 1). However, the use of these techniques is not undisputed [184]: 'While relatively easy to implement, these techniques have not proven to be particularly effective at reducing the variations and tend to remove vital information.…”

“…Signal increments (see, for example, [76,108,125,176]) and force ratios (see, for example, [108,144,176,178,205]) are used to eliminate or to reduce the influence of variable cutting conditions. This objective is mentioned in [209], where increments of the axial and the radial force are determined.…”

Tool wear estimation using an analytic fuzzy classifier and support vector machines

“…Resonance Theory-ART2 (Hong et al 2001), Kohonen Self-organizing Maps (Silva et al 1998), Time-Delay NN (Sick 1998), and Recurrent Neural Networks (Venkatesh et al 1997;Ghasempoor et al 1999;Kamarthi et al 2000) are also occasionally proposed. Recently, Wang et al (2008) proposed a TCM model developed using Fully Forward Connected NN which is a generalized version of MLP NN trained with the Extended Kalman filter algorithm, and Silva (2009) utilized a self-organized Spiking Neural Network.…”