Virtual Simulation Of Experimental Process Digital Twin Design And Implementation

Virtual simulation experiments will be a deep integration of modern information technology and experimental teaching,to carry out real experimental conditions do not have or the actual operation of the difficult practical activities,can effectively solve the traditional experiments "can not do,do not do,do not do well" problem.However,there are some problems in the process of virtual simulation experiments,such as lack of records of experimental process,easy to tamper with experimental results,difficulty in achieving effective digitalization of students’ experimental process and accurate assessment and management of experimental effect quality.Therefore,this thesis constructs a digital twin of the virtual simulation experiment process,realizes the mapping between the virtual simulation experiment process entity and the digital twin based on the formal modeling method,analyzes the data in the twin process by using deep learning algorithms,and records key information through blockchain technology,which can effectively record the student learning trajectory and establish a trustworthy mechanism to realize the accurate assessment of student learning effect.The main research contents of this thesis are as follows:(1)A digital twin modeling method of virtual simulation experiment process based on OOPN(Object Oriented Petri Network)is proposed.In order to effectively record the virtual simulation experimental process,it is necessary to construct its digital twin,and due to the complicated steps of the experimental process,the proposed OOPN approach is used for digital twin modeling in order to better modularize the analysis of its behavioral state variation.Firstly,the overall architecture of the digital twin of the virtual simulation experimental process and the twin mechanism are proposed,then the geometric model of the digital twin is constructed by the formal modeling method,and finally the mapping mechanism between the experimental process and the digital twin of the experimental process is better described by the proposed OOPN-based behavioral state modeling method of the virtual simulation experimental process.(2)A Transformer-based fine-grained image classification method for virtual simulation twins is proposed.In order to effectively distinguish the twin data with subtle feature differences during the virtual simulation experimental process digital twin,image classification is performed by the proposed Transformer-based virtual simulation twin fine-grained image classification method.Firstly,a new Transformer structure called CBG Transformer Block,which consists of convolution operator and multi-axis attention,is proposed to extract image features by repetitive stacking in this module.Then a dual similarity module consisting of a cosine network and a relational network is proposed in the feature comparison module to learn the feature map and take the average of the similarity scores to derive the final prediction.Finally,it is shown experimentally that the proposed method can effectively improve the classification accuracy of twin images in the digital twinning process.(3)Finally,based on the above design of the digital twin of the virtual simulation experiment process and the twin fine-grained image classification method,this thesis designs a blockchain-based digital twin system for the virtual simulation experiment process,which enables the secure sharing of key data in the digital twin system by up-chain storage for multiple parties.This system can monitor,record and analyze the whole process of virtual simulation experiments,provide early warning of abnormal situations,and record key experimental records with key information through blockchain technology,which has high practical value.