Back to results
US 12,662,918 B2Granted
System for dynamic optimization of back pressure on a fluid well
AssigneeRoam-Ai LLC
InventorDavid L. Benham
Life of this patent
Estimating the term…
Abstract
On an oil well equipped with an electric submersible pump (ESP), a valve combination coupled with a controller running a machine learning model provides for continuous valve actuator setting adjustment of a back pressure control valve (e.g., valve actuator settings updated at intervals on the order of 100 ms), and/or adjustment of a variable speed drive (VSD), to achieve tight control around a tubing pressure set point for the well. The system for ESP control may be network-connected, e.g. to the Internet, and pump and well data and the external tubing pressure set point may be communicated through the network.
Claims13
An electric submersible pump (ESP) control system providing dynamic optimization of back pressure on a fluid well, comprising:
- an ESP disposed in a wellbore of the fluid well and contacting well fluids thereof,
- a pump motor disposed in the wellbore configured to mechanically drive the ESP responsive to electrical power provided thereto to lift the well fluids to a surface of the fluid well;
- a power cable that extends from the surface of the fluid well to the pump motor to provide the electrical power thereto;
- a drive connected between a power source and the power cable and operable to vary electrical power provided to the pump motor via the power cable;
- a wellhead at the surface of the fluid well configured to pass well fluids pumped from the fluid well to sales and/or storage via a flowline;
- an actuated pressure control valve connected between the wellhead and the flowline that is operable to continuously vary a flow rate of the well fluids passing through the flowline;
- a controller for the pressure control valve configured to control a degree open of the valve based on a valve actuator setting;
- a pressure transducer that transmits pressure signals from the wellhead to the controller for the actuated pressure control valve;
- wherein
- the controller is configured to receive a pressure set point from a machine learning model output;
- receive pressure signals from a pressure transducer on the wellhead;
- and control a pressure control valve actuator setting continuously to minimize error around the pressure set point in accordance with the machine learning model;
- wherein
- the machine learning model comprises:
- a telemetry sub-model using telemetry data obtained from a plurality of sensors in communication with the ESP to produce an actuator setting;
- and a production sub-model using output from the telemetry sub-model and oil, gas and water production data from the well to produce back pressure set point values;
- and wherein
- the model is characterized by physical constraint parameters applied to predictions made by the telemetry model, wherein the constraint parameters limit optimization of predicted back pressure set points to values within defined boundaries, and wherein the constraint parameters are set manually or adjusted dynamically based on real-time well data.
15 more elements
An electric submersible pump (ESP) system providing dynamic optimization of back pressure on a fluid well, comprising:
- an ESP disposed in a wellbore of the fluid well and contacting well fluids thereof;
- a pump motor disposed in the wellbore configured to mechanically drive the ESP responsive to electrical power provided thereto to lift the well fluids to a surface of the fluid well;
- a power cable that extends from the surface of the fluid well to the pump motor to provide the electrical power thereto;
- a variable speed drive (VSD) connected between a power source and the power cable and operable to vary electrical power provided to the pump motor via the power cable;
- a wellhead at the surface of the fluid well that is configured to pass well fluids pumped from the fluid well to sales and/or storage via a flowline;
- an actuated pressure control valve connected between the wellhead and the flowline;
- a controller for the VSD configured to control pump frequency based on a VSD setting;
- a pressure transducer that transmits pressure signals from the wellhead to the controller for the VSD;
- wherein
- the controller is configured to receive a back pressure set point from a machine learning model output;
- receive pressure signals from a pressure transducer on the wellhead;
- and control the VSD setting at intervals based on the machine learning model output;
- wherein
- the controller is configured to receive the back pressure set point from an edge computing device configured for network connectivity, and the edge computing device is configured to run a machine learning model for optimizing the back pressure set point;
- and wherein
- the machine learning model comprises:
- a telemetry sub-model using telemetry data obtained from a plurality of sensors in communication with the ESP to produce a VSD set point;
- and a production sub-model using output from the telemetry sub-model and oil, gas and water production data from the well to produce a back pressure set point value;
- and wherein
- the machine learning model is characterized by inclusion of physical constraint parameters within which optimization of VSD setpoints occurs, wherein the constraint parameters ensure that predicted VSD set point settings remain within predefined operating limits for the system, and wherein the constraint parameters are set pre-programmed or adaptively adjusted based on the telemetry data.
17 more elements
Dependencies are inferred from the claim text.