| 摘要 |
<p>PURPOSE: To make on-line monitoring on the real time basis for the trouble possibility of a mechanical part by equipping modules for data acquisition, diagnosing technique, machine modeling, database, data control, etc., and introducing a trouble diagnostic system based upon the ARTMAP neuro-circuit network architecture. CONSTITUTION: A machine diagnosing system senses likelihood of a failed condition using a neuro-circuit network, expert system, physical model, and fuzzy theory. The function of each module of this diagnosing system includes data acquisition 410 consisting of collection of vibration, pressure and temp. signal from the machine and a diagnosing technique 420 to perform on-line trouble sensing and diagnosis wherein various techniques such as neuro-circuit network, fuzzy theory, etc., are used. A machine modeling 430 provides data of neuro- circuit for a failure in common bearings and gears, forms a trouble reasoning mechanism, and distinguishes the failed condition. A database 440 stores the information about repair records, while a system control 460 controls and coordinates the inter-module activities.</p> |
