Page 122 - Intelligent Digital Oil And Gas Fields
P. 122

Data Filtering and Conditioning                               89


              of reconciliation are that it can only solve mismatches when there is pressure
              fluctuation, mass or density variations, water cut changes, temperature
              change, etc. Reconciliation cannot be used when an abnormal situation
              occurs, such as the pipeline leakage, pump wear, pump blockage, equipment
              failure, bottleneck in pipeline, etc., which must be attended to immediately.

              3.2.5.1 Reconciliation Method: Example
              An oil well is equipped with a Coriolis meter measuring in real-time liquid
              rate (Well C). The Coriolis meter has a tolerance of 10%; the average liquid
              rate in Well C is 850bbl/d; the total uncertainty is  85bbl/d. Well B is pro-
              ducing 1300bbl/d using a trained virtual meter. The production engineer
              mentions that the confidence interval is 6% because physics-based calibra-
              tion was used to calibrate the virtual metering; therefore, the total uncer-
              tainty is  78bbl/d. Both wells are measured to a portal-mobile MPFM,
              which has a tolerance of 3%. The total sum of independent points is
              2150bbl/d, but the MPFM measures 2265bbl/d (+115bbl/d above the
              manual sum). Fig. 3.4 describes the wells.
                 The method describes:
                 Uncertainty calculation (U t ): U t ¼y (1  %Tol) .





























              Fig. 3.4 Data reconciliation method with three points of measurement: coriolis, virtual,
              and multiphase flowmeters.
   117   118   119   120   121   122   123   124   125   126   127