In order to provide the flow assurance engineer with a more general workflow to assess prediction uncertainties we are in this work proposing a new method to specify flow model uncertainties. First we have reviewed a comprehensive database of laboratory measurements of primarily pressure drop and holdup. Datasets are grouped according to flow conditions. The most significant flow model parameters for each group are identified and then for a subset of the database tuned to each individual measurement simultaneously. Results from sensitivity analysis and uncertainty quantification are discussed. Furthermore, predicted flow model error estimates are verified with one specific measurement series. The results confirm the approach and show that model errors vary significantly within the operational envelope dependent on the actual flow conditions. Finally the obtained flow model parameter distributions can be applied as part of a standard workflow to provide an uncertainty analysis of a flow assurance system comprising both design parameters and inherent flow model errors in one step.

This content is only available via PDF.
You can access this article if you purchase or spend a download.