A pattern recognition based vibration monitoring module for machine tools

✍ Scribed by P. Bartal; L. Monostori

• Publisher: Elsevier Science
• Year: 1988
• Tongue: English
• Weight: 433 KB
• Volume: 4
• Category: Article
• ISSN: 0736-5845
• DOI: 10.1016/0736-5845(88)90018-x

✦ Synopsis

The unmanned operation of manufacturing cells and systems requires automated monitoring equipment, as a substitute for human beings in their supervisory capacity. Concentrating on the frequency domain analysis of measured signals, this paper presents a short summary of the available spectral and cepstral features applicable for describing the different states of the machine or the process to be monitored. Pattern recognition techniques are suggested for the classification phase of the monitoring process. Regarding the requirements of a vibration monitoring module, integrated into a complex, multipurpose machine tool monitoring system, a comparison is made between the different computational devices in performance, complexity and intelligence. The hardware and software architecture of the two-processor based module--incorporating a DSP and a 16 bit microprocessor in parallel architecturenis discussed.