Research On Key Technologies Of Cyber-physical Systems For Intelligent Workshop

The intelligent workshop is an important place to practice intelligent manufacturing concept.Its operation mode is to acquire the relevant dynamic and static information in the workshop entity through the sensing function and map it into the digital twin model corresponding to the equipment entity.With strong computing power to deal with massive data and precision model,the system is able to issue precise commands to devices to improve production efficiency and quality.The intelligent workshop is featured with three key characteristics: sensing,calculation and feedback control,of which sensing serves as the foundation of the other two characteristics.The sensing function relies on the widely deployed sensor networks,processing equipment and other sensing sources in the workshop,involving the interaction between the physical world and the information world.The implementation of sensing function requires the knowledge of Cyber-physical systems(CPS).In this thesis,the sensing model of intelligent workshop is established,and the industrial Ethernet technology and sensor related chronological order reconstruction technology are explored.Finally,via the sensing function,the thesis solved the tool state synchronization and processing parameter identification problems related to digital twins of NC machie.The research content of the thesis includes the following four aspects:1.Research on sensing function related model of CPS in intelligent workshop.Obtaining the dynamic and static information of the processing devices in a workshop is the key to the realization of sensing function.Based on the OPC UA standard,a sensing-oriented information model for workshop is established to locate a certain device in the Ethernet and to describe the data items of the device.The sensing function model is established based on the information model,which enables the function to be provided on time,on demand and by role.2.Research on OPC UA TSN and sensor related chronological order reconstruction technology.Various multi-source heterogeneous network flows exist in the connection layer of the intelligent workshop in CPS.When the network is congested,the real-time transmission of high-priority network flows cannot be guaranteed.Time Sensitive Network(TSN)technology can provide different quality transmission services according to the priority of network flows,thus ensuring the real-time performance of high-priority network flows.However,the static scheduling algorithm in TSN is unable to guarantee the transmission quality of low-priority network flows.The thesis proposes a dynamic gate scheduling algorithm to reconcile the transmission quality between the higher and lower network flows.The algorithm is able to improve the transmission quality of low-priority network flows while ensuring the quality of high-priority ones represented by OPC UA control flows.In the intelligent workshop CPS,the time instant at which the event starts is referred to as chronological order,which provides a reference for determining the state of the system.The direct detection of the events for their respective sensors require the sensors to have strict timing and high sampling frequency.In this thesis,a chronological order reconstruction method based on the integrity of events is proposed.The method distinguishes the chronological order demanding a comparative lower sensor sampling frequency.3.Research on the modelling of machining state of CNC machine tools.During the operation of the CNC machine,the tool exists different states.Although traditional state machine is able to describe the determined state of the tool,it is unable to handle the critical ones.A method based on Fuzzy Finite State Machine(FFSM)is proposed to model critical processing state of CNC machine tools.The model quantifies the critical attributes of the tool through "state superposition" mechanism and cancels system delay by predicting the future state of the tool,which assists the synchronization between tool entity and its virtual counterpart in digital twin.The thesis gives a detailed discussion on modelling steps,parameter optimization and prediction mechanism,with the outcomes consisting of three parts: 1)the qualitative analysis of state transition rules and input variables;2)a genetic algorithm based method for quantitative optimization of model parameters;3)reveal of the prediction mechanism when tools are in steady and critical state respectively.Finally,the results show that within a certain prediction step,the prediction errors of the proposed model is less than that of the low-order linear autoregressive model(LAM).Simultaneously,the model has the ability to predict machine state continuously.4.Research on the depth of processing plane for CNC machine tool.In the digital twin construction problem of CNC machine,the system not only needs to acquire the dynamic and static data of the machine itself,but also the parameters of the parts it has processed,which includes the depth of the milling plane.It is necessary to obtain a higher parameter identification rate at a lower cost using the existing sensing service.The thesis proposes an application for production parameter analysis.Firstly,the depth of rectangular milling plane is achieved by using processed acceleration signal in the feed direction during machining.The methods of signal processing consist of adaptive filtering,wavelet packet decomposition and fuzzy logic.Secondly,the thesis studied how the key parameters of the acceleration sensor influence the performance of the application,configured the industrial acceleration sensor with proper parameter and conducted verification experiment in the proposed system,obtaining a higher application performance at a smaller hardware price.