Artificial Intelligence Tools in Diagnostics of Machine Tool Drives*

Abstract

Diagnostics of machine tool drives aided with an expert system and neural network and allowing for a variety of drive conditions and diagnostic signals is discussed. An actual solution is presented of a drive self-diagnosing system consisting in on-line temperature- and power-based monitoring supplemented by detailed off-line diagnostics backed by Al tools and knowledge bases and invoked in need only. The detailed diagnostics is based on power and acoustic noise measurements and involves data base propagation, a customized diagnosing algorithm, a mechanism of automatic inferring using fuzzy logic procedures and simulation of the inferring mechanism by a neural network. The exposition is completed by a working example of drive diagnostics application in a machining centre.