Decision trees and ICT for rolling bearing monitoring

In its evolution, condition monitoring must benefit from all new data processing and information and communication technologies (ICT). Thus, to improve the capacity of online monitoring of bearings, a combined use of machine learning and web technology can be implemented. This work uses web technology to expose vibration signals, measured on bearings, to a decision tree for fault detection and diagnosis. The study is undertaken on an experimental test bench from which vibration data are initially recorded to build the decision tree model. Then the system is remotely monitored and diagnosed thanks to the decision tree.