This paper presents and discusses comparisons between field data from large-diameter, multiphase pipelines and mechanistic two-phase flow analysis methods. The 10 sets of field data used in the data comparisons are publicly available in the A.G.A. Multiphase Pipeline Data Bank. Pipeline diameters range from 10 inches to 36 inches, and lengths range from 5.5 to 226 miles. For most of these lines, pressure drop and overall holdup data are available. These lines offer a credible set of data for analysis comparisons. The calculation methods used for the field data comparisons include mechanistic methods for flow regime prediction and for the calculation of the pressure drop and holdup in each flow regime. These analysis methods have been selected based upon a thorough review of numerous two-phase analyses found in the literature. We validated the methods against laboratory experiments at large pipe size and high gas density (including Creare experiments at 3.5 and 7-hich diameter as well as the SINTEF test facility). Pipeline Research Committee documents describe the analysis methods and their integration into a framework for pipeline predictions. Calculations show that the pipelines described here tend to be predominantly in the stratified flow regime. For the operating pipelines, best-estimate calculations show that pressure drops are predicted within 25% for this set of pipelines, illustrating good scaling of the mechanistic predictions with pipe size. Total liquid holdup is overpredicted by about 0.5% of the pipe volume at low holdup values (z 1%) and underpredicted by 5% of the pipe volume at high holdup values (s 15%). A useful feature of the mechanistic analysis approach is the ability to offer credible estimates of uncertainties. For example, sensitivity calculations involving interfacial shear quantify the range of uncertainty in the predictions of both laboratory and field data.
Multiphase flow technology is important to the design and operation of many new commercial gas-oil pipelines because most proven reserves are offshore. Running multiphase lines is less costly than separate gas and oil lines because of reduced line lay costs and the elimination of processing facilities on platforms. In the design of production systems, knowledge of multiphase flow is important for sizing pipes, sizing slug catchers, and estimating pigging schedules. Numerous correlations have been developed previously in order to predict the steady-state pressure drop and holdup in commercial pipelines. Many of these empirical equations correlate experimental data in laboratory pipes at small size and low gas density, but they are applied to commercial pipelines at large size and high gas density. There is a plethora of comparisons of these correlations against Commercial pipeline data. For example, see the papers by Bamette (1987), Cuniiffe (1978), Furukawa (1987), Goodreau (1979), and Gregory (1983). GeneraUy, the range of the predictions in the correlations is sufficiently broad that one of them matches the data reasonably well.