Back to results
US 12,529,811 B2Granted
Detection and quantification of sand flows in a borehole
AssigneeHalliburton Energy Services, Inc.
InventorsAtchyuta Ramayya Venna, Batakrishna Mandal, John Philip Granville, Federico Combis Lucas, Jr.
Life of this patent
Estimating the term…
Abstract
Systems, methods, and computer-readable media are provided for detecting sand in a production flow. An example method can include receiving acoustic field data generated by at least one acoustic sensor on a downhole tool lowered into a borehole of a production flow. The method can further include inputting the acoustic field data into an acoustic sand detection model and generating a sand flow signal based on output on the acoustic sand detection model.
Claims20
A method comprising:
- storing a plurality of datasets, wherein:
- each respective dataset of the plurality of datasets is labeled with a respective label that identifies a sand flow rate and an orifice size, and
- a first set of label data of the respective labels identifies that a first dataset of the plurality of datasets is associated with a first sand flow rate and a first orifice size;
- controlling movement of at least one acoustic sensor along a borehole based on operation of a conveyance, wherein the movement of the at least one acoustic sensor is controlled based on operation of a well management system that includes a memory, one or more processors that execute instructions out of the memory, at least one input coupled to the at least one acoustic sensor, and at least one output communicatively coupled to the conveyance;
- sensing acoustic field data by the at least one acoustic sensor when the at least one acoustic sensor is controlled by the conveyance to move to borehole locations that include a first borehole location, wherein the acoustic field data is associated with a production flow of the borehole and the acoustic field data is received by the one or more processors of the well management system via the at least one input;
- performing a plurality of simulations to generate a machine learning sand detection model based on each of the respective datasets being labeled with the respective labels that identify the sand flow rate and the orifice size, wherein a first simulation of the plurality of simulations is performed according to parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- training the machine learning sand detection model with a first portion of a designated number of data subsets to improve operation of the machine learning sand detection model;
- testing the machine learning sand detection model with a second portion of the designated number of data subsets to validate the improved operation of the machine learning sand detection model;
- identifying that the acoustic field data corresponds to the first dataset based on operation of the machine learning sand detection model and the first simulation being performed according to the parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- generating a sand flow signal associated with a current sand flow rate at the first location of the borehole based on the first set of label data that corresponds to the first sand flow rate and the first orifice size;
- and
- performing at least one action of changing the current sand flow rate, scaling back production, plugging a production zone, and repairing or replacing a piece of equipment based on the sand flow signal.
9 more elements
A system comprising:
- a downhole tool that includes at least one sensor;
- a conveyance that lowers the downhole tool into a borehole;
- one or more processors of a well management system that receive data from the at least one sensor via an input of the well management system;
- and
- at least one computer-readable storage-medium having stored therein instructions which, when executed by the one or more processors, cause the system to:
- store a plurality of datasets, wherein:
- each respective dataset of the plurality of datasets is labeled with a respective label that identifies a sand flow rate and an orifice size, and
- a first set of label data of the respective labels identifies that a first dataset of the plurality of datasets is associated with a first sand flow rate and a first orifice size;
- control movement of at least one acoustic sensor along a borehole based on operation of a conveyance, wherein the movement of the at least one acoustic sensor is controlled based on the one or more processors being communicatively coupled to the at least one acoustic sensor and based on at least one output of the well management system being communicatively coupled to the conveyance;
- receive acoustic field data sensed by the at least one sensor when the at least one acoustic sensor is controlled by the conveyance to move to borehole locations that include a first location of the borehole, wherein the acoustic field data is associated with a production flow of the borehole;
- perform a plurality of simulations to generate a machine learning sand detection model based on each of the respective datasets being labeled with the respective labels that identify the sand flow rate and the orifice size, wherein a first simulation of the plurality of simulations is performed according to parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- train the machine learning sand detection model with a first portion of a designated number of data subsets to improve operation of the machine learning sand detection model;
- test the machine learning sand detection model with a second portion of the designated number of data subsets to validate the improved operation of the machine learning sand detection model;
- apply the machine learning sand detection model to the acoustic field data to identify that the acoustic field data corresponds to the first dataset based on operation of the machine learning sand detection model and the first simulation being performed according to the parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- generate a sand flow signal associated with a current sand flow rate at the first location of the borehole based on the first set of label data that corresponds to the first sand flow rate and the first orifice size;
- and
- performing at least one action of changing the current sand flow rate, scaling back production, plugging a production zone, and repairing or replacing a piece of equipment based on the sand flow signal.
14 more elements
A non-transitory computer-readable storage medium comprising:
- instructions stored on the non-transitory computer-readable storage medium, the instructions, when executed by one or more processors, cause the one or more processors to:
- store a plurality of datasets, wherein:
- each respective dataset of the plurality of datasets is labeled with a respective label that identifies a sand flow rate and an orifice size, and
- a first set of label data of the respective labels identifies that a first dataset of the plurality of datasets is associated with a first sand flow rate and a first orifice size;
- control movement of at least one acoustic sensor along a borehole based on operation of a conveyance, wherein the movement of the at least one acoustic sensor is controlled based on operation of a well management system that includes the one or more processors, at least one input coupled to the at least one acoustic sensor, and at least one output communicatively coupled to the conveyance;
- receive acoustic field data sensed by the at least one acoustic sensor when the at least one acoustic sensor is controlled by the conveyance to move to borehole locations that include a first location of the borehole, wherein the acoustic field data is received via at least one input coupled to the at least one acoustic sensor;
- perform a plurality of simulations to generate a machine learning sand detection model based on each of the respective datasets being labeled with the respective labels that identify the sand flow rate and the orifice size, wherein a first simulation of the plurality of simulations is performed according to parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- train the machine learning sand detection model with a first portion of a designated number of data subsets to improve operation of the machine learning sand detection model;
- test the machine learning sand detection model with a second portion of the designated number of data subsets to validate the improved operation of the machine learning sand detection model;
- apply the machine learning sand detection model to the acoustic field data to identify that the acoustic field data corresponds to the first dataset based on the machine learning sand detection model and the first simulation being performed according to the parameter settings that correspond to the first sand flow rate and the first orifice size of the first set of label data;
- generate a sand flow signal associated with a current sand flow rate at the first location of the borehole based on the first set of label data that corresponds to the first sand flow rate and the first orifice size;
- and
- performing at least one action of changing the current sand flow rate, scaling back production, plugging a production zone, and repairing or replacing a piece of equipment based on the sand flow signal.
10 more elements
Dependencies are inferred from the claim text.