Damage Diagnosis of Noncircular Gear Tooth Surface Using Machine Learning

Abstract

In a noncircular gear mechanism, the contact surfaces of the tooth are interchanged during its rotational motion because of its unequal speeds. The loads acting on the tooth of noncircular gears include impact loads, so this mechanism is more easily damaged than that of circular gears. Therefore, it is necessary to detect damage to gear tooth at an early stage and take appropriate preventive actions before the entire machine or equipment is damaged. This study describes an attempt to diagnose the type of damage by machine learning using the data mining tool Weka, based on tooth surface images of super duralumin (A7075) noncircular gears taken by a camera, in order to detect tooth surface damage caused by excessive tooth surface contact.