Predictive maintenance process used during mfr. of machine e.g. motor, ventilator - involves monitoring of groups of signals from detectors attached to machine to train and provide input to neural network

摘要

The process involves grouping functions of a machine being monitored into functional sub-groups, and associating detectors with each group. The detector signals are monitored and recorded. The data is then available to a processor. The processor has a measurement data acquisition module (1), a preprocessing module (2), a memory (4), a parameter selection module (5) permitting determination of characteristic vectors, and a module (6) that analyses these vectors using a neural net (7). The modes and levels of performance recognised from the characteristic vectors are then displayed (10). ADVANTAGE - Gives early warning of onset of failure of part of variety of machines, assisting preventative maintenance.