Using Data Analytics for Virtual Metering of Multiphase Flow

Value Added by Xodus

When clients are looking to get more visibility of their operating assets, attention is often focussed on the requirement to install expensive permanent instrumentation. Xodus is becoming known as the providers of expert engineering and data science to provide uniquely accurate solutions for flow prediction and other physical properties. Our ability to know when a data-based model or physics-based model can be used to save costs associated with installation of physical instrumentation is what makes operators think of Xodus when they need the most fit for purpose and high value solutions. 


The Operations team for a gas and condensate producing asset wished to consider virtual metering of the export streams to supplement the flow measurement from an existing export flow meter which had been producing erratic readings at normal export rates to one of two parallel export trunklines. Investigations had been conducted in the past which revealed that no viable strap-on style ultrasonic instruments were available to measure the application due to high CGR and piping geometry of the 36” and 42” export header/risers. Therefore development of a validated and reliable virtual meter was proposed by the Operations team to predict gas rates and condensate rates to each trunkline. Of primary concern was measurement during a mode of operation where a cross-over between the two trunklines is open, allowing for a pressure balance to govern the direction and amount of flow to each trunkline. 

Disciplines Involved

Data Science, Production Assurance, Solution

  • Data extraction and data analysis of relevant independent process parameters around the existing export flow meter 
  • Multiple Linear Regression in Python of historical data from process parameters to predict gas flow rate to one of the export trunklines and produce a virtual meter 
  • Create a replica of the export system in OLGA and simulate the entire operating envelope of conditions to determine the condensate allocation to both trunklines for any set of conditions 
  • Embed virtual meters of gas flow and condensate flow to each trunkline in the operations surveillance platform 
  • Create a dashboard to monitor deviation of flow prediction over time 


  • Validated virtual meters for gas flow to each export line to provide a reliable measurement of flow when the existing meter is intermittent 
  • Visibility of condensate export bias to each trunkline to improve understanding of liquids in the export trunklines 
  • Understanding of direction and amount of flow across the cross-over between trunklines 
  • Visibility of drift in the actual or virtual metering system by way of quantitative comparison 
Published Date
03 Apr 2022
Production 3
Share this article