A4 Conference proceedings

A Study of Sensor Placement in a Modern CNC Lathe

Publication Details

Authors: Ratava Juho, Lohtander Mika, Varis Juha

Publication year: 2016

Language: English

Title of parent publication: The 7th International Swedish Production Symposium SPS16 Conference Proceedings

Volume number: 1

Issue number: 1

JUFO level of this publication: 0

Open Access: Not an Open Access publication


The state-of-the-art cutting equipment is connected to the Internet or corporate intranet. The equipment will be increasingly monitored for various phenomena to predict the maintenance needs of the machine or to improve the cutting process itself. Sensor data is needed to control the process, especially in the unmanned production of challenging workpieces. New uses for recorded process information include predictive maintenance, managerial purposes and Industrial Internet related services. Sensor placement is of a limited impact if sensors will be used in research and development. However, for day-to-day commercial use the practicality of sensor placement has a great impact on whether the technology can be used at all. In turning with a multiple tool machine, the placement of the sensors greatly affects the use of the lathe, depending on the arrangement of the tool magazine. Ideally, the sensors are placed as close to the tool insert as possible so that the recorded data needs to travel the minimum distance and cross a minimum amount of interfaces between parts. However, with a wired setup, placing the sensor close to the tool all but prevents the changing of the tool without a complex wiring setup or more extensive modifications to the lathe. Alternatively, a wireless communication package could be used, as long as issues with transmission delay and electromagnetic interference are solved. When placed further away from the moving parts, such as into the carriage, the signal-noise ratio of the sensor degrades while sensor wiring can be considerably simplified. In addition to different signal attenuation to different parts of the machine, different parts are bound to have different characteristic frequencies. The characteristic frequencies affect which frequency ranges can be used in acquiring data. In this paper, the effect of sensor placement on measuring acoustic emission and radial vibration caused by the turning process is experimentally studied with piezoelectric sensors using a Doosan Puma 2500Y CNC lathe with a revolving turret type tool magazine. An acoustic emission sensor is used to study the attenuation of high-frequency bursts in the structure of the lathe, and a pair of acceleration sensors are used to compare measurable radial acceleration signals between a turret attached tool holder and the carriage. The signals are additionally compared with cutting sound, which is not as dependent on microphone positioning but is considerably more prone to external interference. The study shows that the measurable signals in the data degrade considerably. However, even in conditions of very high interference and thus low signal-noise ratio, applying pattern detection methods and sensor fusion allows recovery of some information, with applications in chatter detection and continuous chip detection demonstrated.

  • Vmax (01/01/2015 - 31/03/2016)

Last updated on 2017-27-03 at 11:34