Design And Implementation Of Fault Diagnosis System For Industrial Internet Of Things Based On Machine Learning

The planned maintenance mode of traditional industrial equipment maintenance causes problems such as untimely equipment maintenance and high maintenance costs.The application of artificial intelligence technology provides a new way for equipment fault diagnosis.At present,most researches on intelligent device fault diagnosis are limited to algorithm theory and limited to open data sets,which leads to the limited data dimension of the model and the lack of practical operability.In this paper,industrial data acquisition technology and fault diagnosis algorithm based on machine learning are studied according to the actual application scenario of industrial Internet of Things.A fault diagnosis system for intelligent equipment of industrial Internet of Things is designed and developed for rolling bearing components.The main work of this paper includes:(1)The industrial IoT data acquisition system and equipment fault diagnosis algorithm are designed.The paper analyzes the requirements of the industrial loT equipment fault diagnosis system,and designs the system hardware organization structure,data transmission network architecture and system software architecture for the equipment access,local gateway and cloud platform of the data acquisition system.In this paper,data mining technology and machine learning algorithm are used to study the characteristics of five types of sensing data of rolling bearings in practical systems,and a fault diagnosis algorithm based on cyclic neural network pre-trained with boosting tree is proposed.(2)The industrial IoT data acquisition system and equipment fault diagnosis algorithm are realized.The paper studies industrial IoT transmission protocol and industrial sensing technology,and adopts industrial data acquisition technology and big data technology to realize the functions of collecting,transmitting,storing and displaying multi-dimensional sensor data of rolling bearings.The fault diagnosis algorithm based on cyclic neural network pre-trained with boosting tree uses boosting tree to pre-train the collected multi-dimensional sensor data,and the features are combined and reduced at the same time,which reduces the complexity of the cyclic neural network model downstream of the model and meets the requirements of industrial IoT equipment fault diagnosis with high real-time requirements.(3)In this paper,the function test and performance analysis of the data acquisition system of the industrial Internet of Things and the proposed equipment fault diagnosis algorithm are carried out.The data acquisition system tests the functions of data display,data storage and fault diagnosis.The test results show that the system is stable and can diagnose equipment faults in real time.In this paper,the performance of fault diagnosis algorithm based on cyclic neural network pre-trained with boosting tree is analyzed in terms of tree model parameters,multi-dimensional data characteristics and model comparison.The experimental results show that the proposed fault diagnosis algorithm has achieved significant improvement in accuracy.