Design And Method Research Of Intelligent Manufacturing System Based On CPS Koji Making

With the rapid development of the Internet and artificial intelligence,great changes have taken place in the manufacturing industry.Typically,the intelligent manufacturing technology represented by Industry 4.0 has been widely used in industrial production.CPS system,as a comprehensive computing system,network communication systems,sensing systems and control systems of multi-dimensional complex system,is specific way to achieve intelligent manufacturing system.In order to establish the intelligent manufacturing system applied in liquor production,a system based on CPS is being investigated,which is designed to solve the problems of high degree of manual intervention and low production efficiency in the traditional production process,thus,the digital and intelligent koji process can be achieved.The main contents of this paper is as below:(1)Research and improvement on CPS system algorithm.The online parameter measurement method based on artificial neural network optimization is proposed to solve the problem such as longer measuring time and untimely systematic feedback in koji process.By comparing the characteristics of different algorithms,Elman neural network is chosen as the main algorithm for parameter prediction.Based on Elman neural network,DH-Elman neural network with double hidden layer and output layer self-feedback is suggested.It can be seen from algorithm comparison experiment that the proposed parameter prediction model based on DH-Elman neural network has better prediction results and higher accuracy.(2)The system architecture design of CPS intelligent koji process.The system architecture of intelligent koji process based on CPS system architecture is proposed(three layers of CPS system architecture such as perceptual layer,network layer and application layer),which can guide the system design and implementation.According to the proposed CPS system architecture,hierarchical design is conducted.The design of system perception layer proposes four kinds of sensing node design schemes to complete the comprehensive sensing of the koji parameters,namely sensor sensing node,RFID sensing node,Ethernet PLC sensing node and intelligent measurement device sensing node.On the system network layer,a heterogeneous network model of CPS intelligent koji is investigated and the network model is simulated by Packet Tracer to verify the feasibility of the networking model.On the system application layer,the software design process of production process real-time awareness system based on B/S Architecture is proposed and the key code of DH-Elman algorithm based on Java language is designed.(3)Design of algorithm service platform based on RPC architecture.Aiming at the shortcomings that the traditional algorithm service is highly coupled in the system business,this paper proposes the design of RPC algorithm service platform,which makes DH-Elman neural network algorithm service Independent.The algorithm prediction by remote service invocation can greatly improve the system software scalability.(4)The implementation and application of koji intelligent manufacturing system based on CPS architecture.According to the above system design ideas,the implementation of system hardware and software is conducted,and the actual operation of the system is evaluated.Finally,the existing problems and the direction of further improvement are summarized.