Early warning analysis of online vibration fault characteristics of motor base screw loosening based on similarity measurement theory

Abstract

When small- and medium-sized motors are running, abnormal vibration phenomenon caused by various reasons often occur, and continuous development may cause serious motor failure. Therefore, online monitoring and warning for motor vibration is very important. Among them, the looseness of the base fixing screw is the most common and typical cause of abnormal vibration of the motor. Based on the similarity measurement theory, a new method for predicting the fault characteristics of the motor base screw loosening online vibration is proposed in this paper. Firstly, the causes of motor vibration signals are analyzed, and the vibration characteristics of different causes are classified. The vibration fault characteristics of the pedestal fixing screws are mainly studied. The vibration signal generated by the looseness of the base fixing screw is collected by the designed real-time online vibration monitoring system. Then, a fast Fourier transform of the vibration signal is carried out to obtain the corresponding amplitude and frequency characteristic curve. By using the two schemes of ‘Cosine similarity and modulus ratio combination’ and ‘Correlation coefficient combined with vector norm,’ the vibration fault characteristics are analyzed and compared, the vibration fault characteristic signal is extracted, and the vibration reason is determined and alerted by combining fault characteristic classification. Finally, this method is used to compare and test the working condition of the motor frame fixing screw loosening and the normal operation of the motor. The experimental results show that the algorithm which combines correlation coefficient with 1-norm has the advantages of simple algorithm and high accuracy. The method also provides new solutions and ideas for online warning of abnormal vibration caused by other reasons.