The Data Centric Architecture of a Factory Digital Twin
anticipated incremental expansion , and can serve to ensure robustness of design for a distribution of demand scenarios ( see Miller et al 2010 [ 13 ]). An FDT supporting facility retrofit will answer many of the same questions but using the existing plant as the starting point . An FDT supporting planning will answer questions pertaining to customer demand satisfaction timing , needed raw material ordering , expected equipment load and labor levels .
An FDT used for detailed scheduling will specify equipment activity sequencing and timing to accomplish the planned production . Using FDTs for facility fit depends on having an FDT for each possible facility at which a new product can be launched . This network of FDTs can be used to determine which of the plants can best undertake the product launch under a range of scenarios . Considering a network of FDTs brings up many tantalizing prospects for driving corporate profitability and detailed cross company supply chain cooperation both in normal times and exceptional times – for example emergency production during a pandemic or a conflict .
The next section reviews the Resource Task Network ( RTN ) and describes how it must be extended for custom process physics . The following two sections underscore the central importance of timeline management to a FDT and presents a unique innovation for optimizing over the timeline . The following sections review real world FDT characteristics and a step-by-step approach for developing an FDT . Three examples are then presented , and the paper concludes with a discussion of FDT value , a fundamental challenge and a key practical lesson learned . In short , the paper flows from the RTN data to its use in FDT timeline optimization and implementation to examples and concludes with lessons learned .
2 STRUCTURING DATA TO DEVELOP AN FDT : THE RESOURCE TASK NETWORK ( RTN )
The Resource Task Network ( RTN ) underlies the FDT technology described in the paper . The RTN provides a technology independent means of describing processes and organizing manufacturing data and the presented technology provides the ability to optimize manufacturing operations based upon it . From an intuitive perspective , RTN data can be categorized as follows :
• Master – products , equipment , shift-patterns , and other fundamental data
• Recipe – process activities and bill-of-materials
• Process constraints – data to describe physics that must be obeyed
• Demand – orders , forecasts
• Strategy – safety stock , customer priorities
• State – inventories , activity status
• Output – alerts , decision making information
The life cycle of the data moves from ( i ) conception – what questions of interest will the FDT answer , ( ii ) implementation – identify the data sources , ( iii ) operation – connect to demand and state data and incorporate the FDT into the business process .
100 February 2025