Research And Application Of Digital Twin Technology For Metal Furniture Production Line

Due to its features of virtual-real integration and high visualization,digital twin technology has attracted significant interest from academics both domestically and internationally in the context of the global promotion of the deep integration of industry and information technology.In this paper,digital twin technology was applied to the workshop,and a digital twin workshop model oriented to the metal furniture production line was proposed in order to address the issues of poor visualization and low information integration in the current metal furniture production workshop informatization process.Using data from a real workshop,the model was built and confirmed.First,a digital twin workshop model for metal furniture manufacturing workshops was created based on the prevalent digital twin model and integrated with the traits of metal furniture production workshops.Next,the model’s various components were investigated.To address the issue of the variety of communication protocols and challenges in data collection due to the different workshop equipment,an OPC UA-based data collection solution was proposed in the information gathering and storage section.This solution enables the description of workshop equipment information and unified data collection.To develop a virtual workshop with high fidelity and a good visualization effect where human-computer interaction was realized,the Unity 3D-based workshop visualization modeling was researched for the virtual workshop section.Petri net was used to simulate the data transmission process in the section regarding information interaction between virtual and physical workshops,which confirmed the accuracy and efficiency of the transmission process.Afterwards,to address the problem of the opaque production schedule in the workshop,a real-time production schedule prediction method in the digital twin workshop system was investigated,and a production schedule prediction model based on SAE-BPNN was proposed,using a stacked autoencoder to feature select the high-dimensional feature data before inputting the key feature data into the BP neural network to complete the prediction.The model was then trained and validated with historical shop floor data to illustrate its effectiveness.Finally,a digital twin workshop system was designed and developed for a metal manufacturing enterprise,and the key functions of the system were demonstrated and validated,proving that the system could effectively improve the integration of workshop information and achieve a comprehensive visual control of the workshop production process.