Prediction of the product quality of turned parts by real-time acoustic emission indicators

Nowadays, the product quality of turned parts is measured downstream of the actual manufacturing process. This leads to a time-consuming quality control and the risk of a high number of waste and reworking. Even incidents like the fracture of the lathe tool remain undetected until quality issues of the turned parts are measured. Furthermore, certain material defects can't be detected by post-production quality control, which leads to customer complaints because of damages during the use of the parts. This paper presents an in-process approach for evaluating the product quality and tool defects in real-time by using an acoustic emission sensor applied to the tool holder. This paper outlines the identification of feasible quality indicators and explains how the data is recorded and which data sources have to be correlated. This includes for example the recording and correlation of high-frequent acoustic emission signal with further acquired data like machine and computer aided quality (CAQ) data. In dissociation to previous work, this correlation is used directly to develop characteristic factors to predict product quality and to detect tool defects. An overview of several characteristic factors is given. In addition, the test setup is shown and first results are presented, followed by an outlook on further research. The test setup is implemented at a series production without disrupting the daily manufacturing processes. It is shown that solutions in context of Industry 4.0 can be implemented in small and medium-sized companies without a loss of production capacity. The venture is realized within a funded project regarding Industry 4.0 and intelligent quality control systems. Its target is to design smart technologies for manufacturing systems. (c) 2017 The Authors. Published by Elsevier B.V.