Motor remote state monitoring method based on multiple modules

The invention provides a motor remote state monitoring method based on multiple modules. Multiple types of data such as vibration, temperature, current, voltage and rotating speed of the motor are acquired through a data acquisition module of the edge calculation gateway, and threshold alarm, time domain and time-frequency domain feature value extraction and fault diagnosis are performed through an edge calculation module. A fault diagnosis model is dynamically deployed at a cloud end and an edge end through an edge cloud cooperation architecture, and an edge side fault diagnosis algorithm is upgraded through OTA. The invention also provides a motor bearing wear degree fault diagnosis method based on the above scheme. A neural network is used to establish a diagnosis model of a bearing wear state. The problem that the data type of motor state monitoring is single due to the fact that the number of supported sensor types is small, the problem that field measuring points cannot be fully covered due to the fact