The Data Centric Architecture of a Factory Digital Twin
Figure
5-1 : An example Gantt chart .
In the example Gantt Chart above , the Y-axis is the set of modeled equipment , and the X-axis is over the time horizon of interest starting with now and out to a point in the future . The Activity Span of Table 5-1 corresponds to the time span of the Gantt chart . The projected inventory levels of raw materials , intermediates , and product are also predicted in the bottom plot of Figure 5-1 using the same X-axis as the above Gantt chart . The Gantt chart text simply provides the names of the activities and materials and is not relevant for purposes of this paper . The key point is that the Gantt Chart is a useful output of the FDT and communicates behavior over the timeline .
2 . Based upon routine usage of the starting model in 1 , the model can be incrementally expanded to include more detail . Again , the emphasis should be to add only detail that is deemed necessary to address additional questions of interest . Balancing the additional detail is the effort of maintaining the needed data . Perhaps the initial FDT only predicts equipment activity , and the refinement adds labor usage required by each task / activity .
3 . After a short period of time over which a useful working model is developed , the predicted activity of the FDT can be published to the manufacturing community – see Figure 5-2 . An accurate prediction of the future allows community members to best plan their individual activity to maximize the business impact . The results of the FDT can be published on an intranet website and require minimal training . A blog can develop around a manufacturing activity to collaboratively trouble-shoot or update the community on actual start time and other key process information . A database of predicted versus actual process performance results in continuous improvement .
Journal of Innovation 109