Design And Implementation Of SAMP Platform Operation And Maintenance Analysis System Of CAS Based On Digital Twin

With the rapid increase of large-scale and high-precision instruments,it is particularly important to use modern information technology to manage large-scale instruments scientifically and informationally.The instrument and equipment sharing management system(SAMP)established by the Chinese Academy of Sciences,which realize scientific and informational management of instrument and equipment,and promotes the opening and sharing of instrument and equipment resources.However,there are still some problems in the instrument and equipment platform based on the SAMP system of CAS,including faults not be predicted actively,network state is unstable and hard to telemaintenance.Aiming at solving these problems,based on SAMP instrument and equipment sharing management platform of Chinese Academy of Sciences,this thesis combines digital twin technology with instrument and equipment sharing platform to design and implement an operation and maintenance analysis system of SAMP platform of Chinese Academy of Sciences based on digital twin,promoting this system's architecture,operation mechanism,and implementation scheme.The operation and maintenance analysis system of SAMP platform of CAS based on digital twin divided into four parts,including physical platform,virtual platform,platform twin data and platform service module.These four parts correspond to the physical layer,model layer,data layer and system service layer.This system establishes the corresponding digital twin platform model through two-way real-time interaction between physical platform and virtual platform,and achieved data integration and integration among physical platform,virtual platform and platform service module.Facing platform network layer,the system used online monitoring and fault detection method based on Simple Network Management Protocol,achived online monitoring and fault early warning of transmission status of network layer.The system realized data acquisition,model building and online monitoring for platform-oriented instrument and equipment management unit,using LSTM to model the internal voltage data,memory operation data and CPU occupancy data,further more,to predict their operation status.In the process of experiment,two methods of single predictive value prediction and multi-predictive value prediction are set up to compare the models.The results show that the LSTM single predictive value prediction model for maximum voltage,memory operation state and CPU occupancy state is better.The Mean Square of Error(MSE)are 0.0034,0.0022 and 0.0039 respectively.To sum up,the system can predict the running trend of SAMP platform and the possible risks in the platform effectively by analyzing the platform twin data.The system can ensure the operation of the SAMP platform while enabling maintenance personnel to monitor the operation of the platform remotely,facilitate the remote maintenance of the platform,timely correct the problems arising in the operation of the platform,improve the efficiency of remote maintenance of the platform,reduce the maintenance cost of the platform,and also provide the foundation for the establishment and identification of digital twin laboratory.