Design Of Intelligent Motor On-line Integrated Monitoring System And Research On Fault Diagnosis Of Vibration

It's unable to monitor every motor in real time and difficult to discovery and deal with sudden faults of motors in time, because the number of motors is large and installation locations of motors are dispersive. In order to acquire the early warnings of abnormal operations of the motor, a kind of intelligent motor on-line integrated monitoring system is designed to solve this problem. Common faults and early warning signs of motors are firstly analyzed in this dissertation. It is determined that the most basic and important function of the system is to measure all kinds of operation parameters of motors in real time. The artificial intelligence methods of typical fault diagnosis are discussed. The overall scheme of the system is put forward.The system uses microcontroller and sensor chips with high performance and low cost to measure parameters. Data analysis and processing are mainly dependent on the lower computer.The system is composed of four modules. The centralized measurement module has measured three-phase voltages, three-phase currents and rotational speed of the motor. The temperature measurement module has measured the temperature of the motor. The environmental temperature and humidity measurement module has measured environment temperature and relative humidity of the motor. The vibration measurement module has measured the vibration of the motor. Every module also has finished system date and time record, data storage, liquid crystal display, data transmission, alarming and other related auxiliary functions.The system is applied to the motor vibration test in this dissertation. The two kinds of vibration signals sampled in normal operation and under the circumstance that fixed screws of motor base are loose are processed by the microcontroller using FFT so as to get the amplitude-frequency characteristic curves. The combination ofcosine similarity with ratio of modulus and the combination of correlation coefficient with vector norms are analyzed and compared. The result shows that the combination of correlation coefficient with 1-norm is more suitable for the microcontroller to automatically distinguish the amplitude-frequency characteristic curves of the two kinds of working conditions. It provides a new way to discriminate whether the motor is under the circumstance that fixed screws of motor base are loose.