Research And Application Of Deep Learning In Bldc Production Line Monitoring And Fault Diagnosis System

Brushless DC motor(BLDC)is becoming more and more popular with its superior characteristics.It has been widely used in various fields from civil to military.There are more and more enterprises producing brushless DC motor.In order to ensure the safety and intelligent production of Brushless DC motor production line,BLDC motor is widely used in many fields,It is of great significance to develop a monitoring and fault diagnosis system for brushless DC motor production line.Based on mqtt new Internet communication protocol and deep belief network(DBN),this paper develops a set of Brushless DC motor production line monitoring and fault diagnosis system,which realizes the operation status monitoring of the production line equipment and fault diagnosis of the main production equipment.Firstly,the function and function of each equipment in BLDC production line are analyzed;This paper analyzes the structural and functional requirements of the monitoring and fault diagnosis system,and designs the overall structure of the system according to the structural and functional requirements of the production line.The system is divided into five parts: equipment layer,data acquisition layer,data transmission layer,cloud server layer and monitoring layer.The functions of each layer are analyzed,and the design ideas and principles of the system are elaborated.Secondly,taking the main equipment of BLDC production line as the diagnosis target,the principle and structure of deep confidence network are analyzed,as well as its characteristics of adaptive fault feature extraction and independent of signal processing technology;Based on the traditional deep belief network model,the adaptive learning rate is improved.The improved method is applied to the fault diagnosis of the main production equipment winding machine of BLDC production line,and the performance of the model is tested by training model simulation.Then,the software design and communication design of BLDC production line monitoring and fault diagnosis system are given.The standard mqtt Internet of things communication protocol is applied to the communication of the system.Using the accuracy and rapidity of mqtt protocol,the operation data transmission of each equipment in BLDC production line is realized,and the specific communication implementation steps and processes are given,It includes the construction of mqtt server and the implementation of mqtt subscription / publication client;Based on the software development technology and platform characteristics,the system selects the.Net framework,uses C # high-level language programming,the whole system uses threetier architecture,and selects SQL server relational database according to the system data size;The software structure of the system is designed,which is mainly divided into three core parts: user management module,status monitoring module and fault diagnosis module,and the development ideas of each module are given.Finally,the BLDC production line monitoring and fault diagnosis system is deployed on cloud platform.According to the design of front-end system software structure and requirements,the user module,status monitoring module and fault diagnosis module are implemented,and the practicability of the system is tested online.