Research On Visualization Monitoring And Fault Warning System Of Rotating Machinery Based On Digital Twins

With the rapid development of information technology,countries around the world have explored the path of intelligent and digital transformation of manufacturing industry.The problems of incomplete view of the whole life cycle monitoring of machinery and equipment,poor real-time data transmission,low accuracy and efficiency of fault status warning in the process of intelligent manufacturing are still the key technical problems to be faced in the field of machinery and equipment health status monitoring.In response to the above problems,this paper proposes a method of visualization monitoring and fault warning of rotating machinery based on digital twin technology,aiming to realize visualization monitoring and real-time fault warning of rotating machinery with the help of 3D modeling,data communication,deep learning and other multidisciplinary technologies.The main research contents are as follows.(1)The overall scheme design of the rotating machinery visualization monitoring and fault warning system.Aiming at the key issues in the research of visualization monitoring and fault warning of rotating machinery,the design requirements and design principles of the system are summarized,and the design scheme of overall framework and functional modules of the rotating machinery visualization monitoring and fault warning system based on the five-dimensional architecture of the digital twin model are determined.(2)Research on the construction method of high-fidelity digital twin model of rotating machinery.A bearing-gearbox rotating machinery experimental rig was built.The structure of the digital twin model was analyzed from the perspectives of Geometric model,behavioral model and fault evolution model,and the strategy of constructing,assembling and optimizing the geometric model of the rotating machinery test rig was proposed based on the principles of digital twin model construction.Based on the principle of spatial geometric transformation of the digital twin model,we analyzed the implementation method of the behavioral model and constructed a high-fidelity digital twin model of rotating machinery with complete mapping.(3)Research on the mapping method of the digital twin model of the rotating machine test rig.In order to obtain the real-time twin data of the rotating machine test rig and provide data support for the rotating machinery visualization monitoring and fault warning system construction,a multi-sensor data acquisition system is built for the twin data requirements,and a data acquisition scheme based on Lab VIEW DAQmx driver is designed.The real-time transmission and storage of twin data are realized by building a database and data transmission channel.The construction method of digital twin behavior model is proposed,and the construction of 3D simulation scene and the real-time accurate mapping of digital twin model are realized based on Unity3 D development engine.(4)Research on fault diagnosis method based on capsule network.In order to achieve accurate diagnosis and early warning of rotating machinery faults,a fault diagnosis model construction method based on capsule network is proposed.Faulty parts of rolling bearings and planetary gearboxes are made,and their composite fault data sets are constructed as model training samples for the diagnosis of single and coupled faults of rolling bearings and planetary gears.The convolutional neural network fault diagnosis model is constructed using the same parameters for comparison tests,and the accuracy of the fault diagnosis model is verified by analyzing the confusion matrix,t-SNE visualization results and comparison experimental results output from the capsule network model based on accuracy assessment index.(5)Development of visualization monitoring and fault warning system for rotating machinery.For the fluency,applicability and compatibility requirements of the rotating machinery visualization monitoring and fault warning system,the system development environment was configured,four functional modules of user registration and login,equipment status monitoring,3D model visualization and fault warning were built,each functional module was integrated and developed based on the hybrid B/S and C/S architecture.