extraction of rules for faulty bearing classification by a neuro-fuzzy approach

in this paper, a classification system of faulty bearings based on a neuro-fuzzy approach is presented. the vibration signals in the frequency domain produced by the faulty bearings will be taken as the inputs to the classification system. in this sense, it is an essential characteristic for the used neuro-fuzzy approach, the possibility of taking a great number of inputs. the system consists of several neuro-fuzzy systems for determining different bearing status, along with a measurement equipment of the vibration spectral data. in this paper, a special attention is focused on the analysis of the rules obtained by the final neuro-fuzzy system. in fact, a rule extraction process and an interpretation rule process is discussed. several trials have been carried out, taking into account the vibration spectral data collected by the measurement equipment, where satisfactory results have been achieved.