Artificial Neural Network based identification system for abnormal vibration of motor rotating disc system

Abstract

This paper reports an early work of machinery fault detection system module development. The system is developed and employed on a mechanical platform having a series of 3 aluminum rotating discs with unbalanced rotating mass to simulate an abnormal condition of a real machinery. This detection system is intended to have a capability of either to give an early warning due to an abnormality of the machine vibration or to localize the position of such abnormality among the discs. Artificial Neural Network (ANN) method is used to determine and to localize the abnormality by utilizing the vibration data. The method utilizes 3 features of time domain and 2 feature frequency domain signal characteristics. After the ANN was trained, this detection system was able to identify the plant condition of 90% accuracy.