Ultrasonics 38 (2000) 604–608 www.elsevier.nl/locate/ultras Some aspects of AE application in tool condition monitoring Krzysztof Jemielniak * Warsaw University of Technology, Narbutta 86, 02-524 Warsaw, Poland Abstract Acoustic emission (AE ) is rather a well-known form of non-destructive testing. In the last few years the technology of the AE measurement has been expanded to cover the area of tool condition monitoring. The paper presents some experience of Warsaw University of Technology ( WUT) in such applications of AE. It provides an interpretation of common AE signal distortions and possible solutions to avoid them. Furthermore, a characteristic study of several different AE and ultrasonic sensors being used in WUT is furnished. Evaluation of the applicability of some basic measures of acoustic emission for tool condition monitoring is also presented in the paper. Finally paper presents a method of the catastrophic tool failure detection in turning, which uses symptoms other than the direct magnitude AE RMS signal. The method is based on the statistical analysis of the distributions of the AE RMS signal. © 2000 Elsevier Science B.V. All rights reserved. Keywords: Acoustic emission; Tool condition monitoring 1. Introduction piezoelectric AE sensor is usually placed as close as possible to the cutting zone. Low frequency noise com- Rising labour cost makes production automation an ponents, which are inevitably present in AE signal, are important priority in the major industrial countries. One considered to be not correlated with tool’s condition of the most important factors limiting the progress in and hence useless. In addition, they can be of high introduction of unattended machine tool operation is amplitude forcing usage of lower signal amplification. tool condition monitoring ( TCM ). Numerous different Therefore those components should be eliminated (high- phenomena can be employed for this purpose and a pass filtered) at the earliest possible stage of signal variety of sensor types are available on the market, e.g. processing, just after inevitable buffering. Sometimes the Refs. [1–4]. Despite a growing number of publications AE signal is then fed through a low-pass filter to get rid presenting new, elegant methods of raw AE signal of the high frequency noise components due to electric processing, the methods have not been applied in indu- sparks, etc. or to avoid aliasing. The raw AE signal strial practice so far, where rather simple methods (AE raw ) can be demodulated in the form of the mean prevail, utilising the demodulated (usually AE RMS ) signal value or RMS (AE RMS ) to obtain a low frequency [3]. However, existing tool condition monitoring sys- variable, so it can be recorded or processed with conven- tems, based on AE measurements, are still not consid- tional, less expensive signal processing equipment. ered reliable enough. Therefore it seems worthwhile to Fig. 2 presents AE raw signal obtained from the broad go back to some basic consideration concerning the AE band transducer Bru ¨ el & Kjær 8312 without filtering, signal processing and evaluation. when cutting carbon steel 45 with conventional carbide P30 worn out insert. The long view (Fig. 2(a)) shows three bursts of nearly the same maximum amplitude. In 2. AE signal processing and interpretation demodulated signal (analogue AE RMS , Fig. 2(b)) the third burst reaches the highest value. Fragments marked 2.1. Measuring chain with black rectangles under both figures are shown in Fig. 2(c), and their amplitude spectra are presented in The typical procedure of AE signal processing in Fig. 2(d). The figures reveal the existence of a dominant metal cutting follows the pattern shown in Fig. 1. The low frequency component. It was caused by some mechanical disturbances, and consequently should be * Tel.: +48-22-6608-656; fax: +48-22-8490285. E-mail address: k.jemielniak@wip.pw.edu.pl ( K. Jemielniak) filtered out as irrelevant to tool wear. Of course in the 0041-624X/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved. PII: S0041-624X(99)00195-X