Rapid design of tool-wear condition monitoring systems for turning processes using novelty detection

Abstract

Condition monitoring systems of manufacturing processes have been recognised in recent years as one of the key technologies that provide the competitive advantage in many manufacturing environments. It is capable of providing essential means to reduce cost, increase productivity, improve quality and prevent damage to the machine or work-piece. Turning operations are considered one of the most common manufacturing processes in industry. It is used to manufacture different round objects such as shafts, spindles and pins. Despite recent development and intensive engineering research, the development of tool wear monitoring systems in turning is still on-going challenge. In this paper, force signals are used for monitoring tool wear in a feature fusion model. A novel approach for the design of condition monitoring systems for turning operations using novelty detection algorithm is presented. The results prove that the developed system can be used for rapid design of condition monitoring systems for turning operations to predict tool-wear.

The Application of Force and Acoustic Emission Sensors for Detecting Tool Damage in Turning Processes

“…However, automated signal processing and analysis is needed to extract the important information from the signals (i.e. Sensory Characteristic Features (SCFs)) [1,6,7]. The test from the first tool is used to investigate the process characteristic and to select the sensitive Sensory Characteristic Features…”

“…The most critical aspect of the design of a condition monitoring system is how to select a suitable sensor or signal processing system to design the experiments. The authors in previous publications have suggested an Automated Signal and Signal Processing Selection System (ASPS) [6,7]. This is preformed by processing the available sensory signals with a selection of signal processing algorithms.…”

“…In his paper this theory has been tested using acoustic emission and force signals using different features within the time and frequency domains. For decision making and classification, novelty detection algorithm [7] has been implemented to detect the normal conditions and investigate any new conditions within the selected features. For more details, see [6,7].…”

“…For decision making and classification, novelty detection algorithm [7] has been implemented to detect the normal conditions and investigate any new conditions within the selected features. For more details, see [6,7].…”

“…The response of the Gaussian kernels φ j , as shown in equation 1, is defined by a covariance matrix (a spherical matrix in this case) and a centre (i.e. the centroid of the input clusters), see [7] for further details. A single variance parameter for each Gaussian component is calculated using 4 centres in the mixture which has been found to be a suitable structure that gives relatively quick learning process and consistent results.…”