Abstract
It is known that Ferrography analytical method can find the unusual state of the bearing surface of the rotating machine with a higher sensitivity than that of other monitoring methods, such as the measurement of frictional force, frictional temperature and rotational speed. This method can detect the abnormal wear particles, such as cutting wear particles, spherical particles, sever wear particles and black oxide particles, before the frictional force and temperature caused the unusual change and rotational speed decreased. However this method requires considerable operator skill to distinguish the wear particles. We applied image processing techniques and artificial Neural Networks using a computer to Ferrography analytical method. It was shown that our proposed system could distinguish the wear particles well.
