Construction method of digital twin system in production workshop of special transformer

In order to solve the problems of weak real-time control, poor decision-making ability, non intuitiveprocess monitoring and unclear operation situation in the special transformer production workshop. A new architect-ure of digital twin is discussed, the relationship and content of physical workshop layer, multi-source data mapping layer, digital twin model layer and digital twin application layer are generally expounded. A modeling and express-ion method of digital twin based on "5 dimension-4 perspective", Engine with high performance computing power and technology of immersive visualization are described, the modeling and development of digital twin system is divided into three parts, 3D modeling, data integration and computing power development. A deep fusion method of knowledge and digital twin is studied based on principle of tree’s growth. At first, a design method of index system for digital twin workshop is proposed by analogy with the hierarchical structure and growth principle of tree. From the perspective of workshop scale, index is divided into workshop level, production line level, station level and equipment level, which constitute the main branches of the index tree model vertically. From the perspe-ctive of the life cycle of production activities, depend on the basis of the index tree model’s main branches, the indexs are derived into planning parameters, process parameters, work order parameters, quality parameters, equipm-ent parameters and so on, the subdivision branches of the detailed tree model are formed. Repeat the steps until a complete tree model of digital twin workshop is built. Then, simulate photosynthesis and transportation of tree’s organic products, formally describe the deep fusion mechanism of knowledge and digital twin in transformer works-hop. Based on the big data of transformer workshop operation, complex and diverse intelligent computing units bu-ilt based on mathematical formulas, machine learning algorithms and deep learning models, to mine knowledge and identify information from the messy big data of workshop operation. Further more, the basic index and lower dimensional knowledge will be highly abstracted and fused through iterative calculation of aggregation nodes, it canbe converged to the trunk of the digital twin workshop’s index system step by step. An operation reliability analy-sis model of digital twin system for special transformer workshop based on composite matter element information entropy is also proposed. Combined with analytic hierarchy process and correlation entropy method, a composite matter-element model was built to realize real-time monitoring of the operation situation of the transformer workshopby analyzing the reliability of the workshop. Finally, taking a special transformer workshop as the application case,a prototype digital twin system is developed, the rationality and effectiveness are verified.