US 11,755,795 B2
Detecting and mitigating flow instabilities in hydrocarbon production wells
Feng Xiao, Spring, TX (US); Neeraj R. Dani, Houston, TX (US); Ted A. Long, Spring, TX (US); Justin A. Gantt, St. Johns (CA); Zachary H. Borden, Houston, TX (US); Amr El-Bakry, Houston, TX (US); and Curtis J. Holub, Spring, TX (US)
Assigned to ExxonMobil Technology and Engineering Company, Spring, TX (US)
Filed by ExxonMobil Technology and Engineering Company, Spring, TX (US)
Filed on Sep. 18, 2018, as Appl. No. 16/133,983.
Claims priority of provisional application 62/562,015, filed on Sep. 22, 2017.
Prior Publication US 2019/0093455 A1, Mar. 28, 2019
Int. Cl. G06F 30/28 (2020.01); E21B 47/10 (2012.01); E21B 43/12 (2006.01); G06F 30/20 (2020.01); E21B 47/07 (2012.01); E21B 47/008 (2012.01); E21B 41/00 (2006.01); E21B 47/06 (2012.01); G06F 111/10 (2020.01)
CPC G06F 30/28 (2020.01) [E21B 41/00 (2013.01); E21B 43/122 (2013.01); E21B 47/008 (2020.05); E21B 47/07 (2020.05); E21B 47/10 (2013.01); G06F 30/20 (2020.01); E21B 47/06 (2013.01); G06F 2111/10 (2020.01)] 6 Claims
OG exemplary drawing
 
1. A method of detecting and mitigating flow instabilities in one or more hydrocarbon production wells, the method comprising:
retrieving real-time production data pertaining to each of the one or more hydrocarbon production wells;
using the real-time production data, identifying patterns of flow instability within the real-time production data;
generating a numerical model of transient and thermal multiphase flow in each of the one or more hydrocarbon production wells;
retrieving well test data from a database;
calibrating the numerical model using the well test data;
using the calibrated numerical model, performing a parametric study to determine how input parameters affect at least one of a flow stability or a performance of the one or more hydrocarbon production wells, wherein results from the parametric study comprise stable parameter sets and unstable parameter sets for the one or more hydrocarbon production wells;
generating a surrogate model based on the parametric study, wherein the surrogate model comprises a neural network model that is trained using a training dataset comprising the results from the parametric study comprising the stable parameter sets and the unstable parameter sets for the one or more hydrocarbon production wells, in combination with the real-time production data;
using the surrogate model to determine a type of flow instability and to determine at least one change to at least one operating condition to improve the at least one of the flow stability or the performance of the one or more hydrocarbon production wells; and
providing, to a user, an advisory to effect the at least one change to the at least one operating condition.