A fuzzy approach to the condition monitoring of a packaging plant

The packaging of manufactured products attracts relatively little attention from research or benefits from the opportunities available through conditi on monitoring. The technical requirements of packaging can easily be undere stimated in the overall production process. A widely used method of reducin g waste relies upon the knowledge of an experienced mechanic to set up the machine before production runs and to correct misalignments should problems occur. This approach has the following drawbacks: non-availability of expe rtise, the relevant practical experience takes time to acquire and the redu ced staffing levels present in the current economic climate. This study aim s to develop an efficient, hybrid method for capturing the machine informat ion, by means of fuzzy logic to decode the intelligence supplied from an RT CM system. Developed from the min-max fuzzy inference method, the present a pproach replaces the maximum selection technique for output set truncation, by a matrix of truncation values derived from all rule outcomes. A suitabl y generic method has been employed allowing the developed techniques to be utilised for a wide range of manufacturing problems. The work has lent supp ort to the use of fuzzy condition monitoring as a reliable and inexpensive means of reducing wastage and maintenance overheads in the packaging indust ry. (C) 2001 Published by Elsevier Science B.V.