Application of the artificial immune systems for bearings diagnostic in servomotors

Nowadays in industries we can assist to a technological evolutionary phase in design of machineries, from asynchronous motor-based (AMB) actuation to servomotor-based (SMB) actuation. This new type merges together the functions of a motor, a gearbox and a cam into a unique element which is the servomotor. The main advantages are the removal of complex mechanical components like gearboxes and cams which are subjected to wear, need maintenance, etc…, an improved easiness in the setup changing and an increased acceptable complexity of the motion profile. The main drawback regards the diagnostics activity, e.g. on bearings, where classical methods based on the research of the fault frequencies in the signal spectrum cannot be applied anymore, due to the consistent variability in the speed of the shaft. This paper tackles the bearing diagnostic in servomotors by means of an unsupervised learning approach: the artificial immune system, which has been developed and applied with success in the field of computer security, and – as the name suggests – its aim is to mimic the behavior of the human immune system which is able to recognize health hazards, like virus, even if never seen before