Design and Implementation of a Digital Twin for Live Petroleum Production Optimization
with corresponding sensor signals stored in columns . An example of such a file is displayed in Figure 3 , in the image titled “ Sensor Data : Flat files – csv ”.
2 . Metadata : To further represent the system which the sensor data is associated with , the metadata of the system is necessary . Examples of metadata can include but not restricted to :
● Name , location , deviation and completion data of the well
● Design data for the artificial lift equipment
● Piping and Instrumentation schematic of the well and facilities sites
● Fluid properties : PVT ( Pressure , Volume , Temperature ) data
Figure 3 highlights how the metadata may be available in various formats such as images , excel sheets , PDF documents etc . A key component of automation is to digitize the various forms of data into a uniform data store . This involves creating a template that records the quantitative details of the metadata . An example of such a digitized template of the metadata is displayed in Figure 4 . Once digitized , metadata is extractable through an API as a hierarchical data format such as JSON .
It is important to note that metadata is not usually available at a single location or from a single source . There is a significant amount of manual effort that goes into contacting the field operators , the SCADA company , the equipment manufacturers to gather this data and to digitize it . A uniform reference data store space was created on the cloud to maintain and access the raw metadata as a source of truth .
Often metadata is manually entered into the system because it is associated with the well when it is installed and does not change afterwards . When there is a maintenance event that changes a physical equipment of the well , the metadata may need to be updated depending on the changes being performed .
- 42 - March 2021