Page 97 - Handbook Of Multiphase Flow Assurance

92 4. Hydraulic and thermal analysis

Accuracy of the individual components of the hydraulic model such as reservoir flow
resistance, wellbore flow resistance, tree choke flow resistance and gathering flowline and
riser flow resistance all play into the overall reliability of the prediction for network flow
optimization.
Eventually the process was developed to integrate the network model within a single
steady state multiphase simulator. There are numerous commercially available steady state
simulators which can handle the task of flow network optimization.
Present day simulation is seeing a growing adoption of methods based on machine
learning algorithms. Multiple operator corporations and service companies are developing
capability in this area and deploying their solutions as field application. In the machine
learning approach, a number of simulations is performed for the individual segments and
for the whole network to determine flow hydraulic loss for a given set of conditions. With
the data set developed for the flow resistance versus operating conditions, machine learn-
ing database is then trained on a part of the data set, with the remaining part of the sim-
ulations comprising the data set kept for verification or validation of the accuracy of the
trained database prediction. Modern tools such as Python 3.7 language and Pytorch library
for parallelizing the database training in order to save time, with appropriate integrated
development environment tool such as Jupyter, may be used to implement artificial intel-
ligence for flow network optimization. Development tools keep evolving along with hard-
ware and software, so newer ones may gain acceptance with time. Recent implementations
of machine learning also deal with virtual flow metering (Andrianov, 2018), choke control,
gas lift optimization, as well as the detection of flow assurance blockages building up in the
production system.
A comparison of various artificial intelligence methods for multiphase hydraulic calcula-
tions with field data is provided by Attia et al. (2015).

References

Al-Ne’aim, S.A., Aggour, M.A., Al-Yousef, H.Y., 1995. SPE-29850, Horizontal multiphase flow correlations for large
diameter pipes and high flow rates. In: Presented at Middle East Oil Show, 11–14 March Bahrain.
Andrianov, N., 2018. A machine learning approach for virtual flow metering and forecasting. In: Proc. of 3rd IFAC
Workshop on Automatic Control in Offshore Oil and Gas Production, Esbjerg, Denmark, May 30–June 01.
Arabnejad, H., Mansouri, A., Shirazi, S.A., McLaury, B.S., 2015. SPE-174987, evaluation of solid particle erosion equa-
tions and models for oil and gas industry applications. In: Proceedings SPE ATCE Houston 28–30 September.
Attia, M., Abdulraheem, A., Mahmoud, M.A., 2015. SPE-175724, Pressure drop due to multiphase flow using four
artificial intelligence methods. In: SPE North Africa ATCE, 14–16 September, Cairo.
Aziz, K., Govier, G.W., Fogarasi, M., 1972. Pressure drop in wells producing oil and gas. J. Can. Pet. Technol.
38–48.
Beggs, H.D., Brill, J.P., 1973. A study of two-phase flow in inclined pipes. J. Pet. Technol. 25 (5), 607–617.
Blasius, P.R.H., 1913. Das Aehnlichkeitsgesetz bei Reibungsvorgangen in Flüssigkeiten. Forschungsheft 131, 1–41.
Brill, J.P., Mukherjee, H., 1999. Multiphase Flow in Wells. SPE Monograph series. vol. 17. Richardson, Texas, Society
of Petroleum Engineers.
Jorge Fernando Carrascal, 1996. Multiphase flow application to ESP pump design program. Texas Tech University.
Master of Science Thesis.
Child, H., Brauer, P., 2017. Comparing HZ Critical rate models against Marcellus field data. In: Gas well deliquifica-
tion workshop, Feb 20-22, 2017, Denver.
Coleman, S.B., Clay, H.B., McCurdy, D.G., Norris, L.H., 1991. A new look at predicting gas-well load up. J. Pet. Tech.,
329–333.