Spectral analysis and machine learning to detect offset well communication using high frequency acoustic or vibration sensing
This disclosure presents a system, method, and apparatus for preventing fracture communication between wells, the system comprising: a sensor coupled to a fracking wellhead, circulating fluid line, or standpipe of a well and configured to convert acoustic vibrations in fracking fluid in the well into an electrical signal; a memory configured to store the electrical signal; a machine-learning system configured to analyze current frequency components of the electrical signal in a window of time and to identify impending fracture communication between the well and an offset well, the machine-learning system having been trained on previous frequency components of electrical signals measured during previous instances of fracture communication between wells; and a user interface configured to return a notification of the impending fracture communication to an operator of the well.
1 . A method of predicting fracture communication between wells comprising:
providing a sensor coupled to a wellhead, circulating fluid line, or standpipe of a well and configured to convert acoustic vibrations in fluid in the well into an electrical signal in a time domain;
recording the electrical signal to a memory;
analyzing current frequency components of the electrical signal in a window of time;
identifying impending fracture communication between the well and an offset well where the current frequency components show sufficient similarity to previous frequency components measured during previous instances of fracture communication between wells; and
returning a notification of the impeding fracture communication to an operator of the well.
2 . The method of claim 1 , wherein the acoustic vibrations are caused by injecting the fracking fluid through perforations in a casing of the well under pressure in order to form subsurface fractures, the fracking fluid's flow through subsurface fractures, expansion of subsurface fractures, or a pumping truck at the offset well.
3 . The method of claim 1 , wherein the sensor samples at greater than 1 kHz.
4 . The method of claim 1 , wherein the sensor is an acoustic sensor.
5 . The method of claim 1 , further comprising converting the electrical signal in the window of time to a frequency spectrum and analyzing the current frequency components in the frequency spectrum to identify impending fracture communication between the well and the offset well.
6 . The method of claim 1 , wherein the analyzing and the identifying is performed by a machine-learning system.
7 . The method of claim 6 , wherein the machine-learning system is trained on the previous frequency components of electrical signals measured during previous instances of fracture communication between wells.
8 . The method of claim 7 , wherein the machine-learning system is trained on pressure data measured during previous instances of fracture communication between wells.
9 . The method of claim 6 , wherein the analyzing comprises monitoring increases in amplitude of the current frequency components.
10 . The method of claim 6 , further comprising analyzing the electrical signal in the time domain to identify impending fracture communication between the well and the offset well.
11 . The method of claim 6 , further comprising analyzing electrical signals from a pressure sensor at the wellhead, circulating fluid line, or standpipe of the well in addition to the electrical signal from the acoustic vibrations.
12 . The method of claim 6 , further comprising adjusting parameters of subsequent fracking operations to reduce chances of fracture communication between the well and the offset well.
13 . A method of preventing fracture communication between wells, the method comprising:
performing a fracking operation on a first well in a subterranean formation;
providing a sensor coupled to a wellhead, circulating fluid line, or standpipe of a second well and configured to convert acoustic vibrations in fluid in the second well into an electrical signal in a time domain;
recording the electrical signal to a memory;
analyzing current frequency components of the electrical signal in a window of time;
identifying impending fracture communication between the first and second wells where the current frequency components show sufficient similarity to previous frequency components measured during previous instances of fracture communication between wells; and
adjusting parameters of the fracking operation on the first well to avoid the impending fracture communication between the first and second wells.
14 . The method of claim 13 , wherein the acoustic vibrations are caused by injecting fracking fluid through perforations in a casing of the first well under pressure in order to form subsurface fractures, the fracking fluid's flow through subsurface fractures, expansion of subsurface fractures, or a pumping truck at the first well.
15 . The method of claim 13 , wherein the sensor samples at greater than 1 kHz.
16 . The method of claim 13 , wherein the sensor is an acoustic sensor.
17 . The method of claim 13 , further comprising converting the electrical signal in the window of time to a frequency spectrum and analyzing the current frequency components in the frequency spectrum to identify impending fracture communication between the first and second wells.
18 . The method of claim 13 , wherein the analyzing and the identifying is performed by a machine-learning system.
19 . The method of claim 18 , wherein the machine-learning system is trained on the previous frequency components of electrical signals measured during previous instances of fracture communication between wells.
20 . The method of claim 19 , wherein the machine-learning system is trained on pressure data measured during previous instances of fracture communication between wells.
21 . The method of claim 13 , wherein the analyzing comprises monitoring increases in amplitude of the current frequency components.
22 . The method of claim 13 , further comprising analyzing the electrical signal in the time domain to identify impending fracture communication between the first and second wells.
23 . The method of claim 13 , further comprising analyzing electrical signals from a pressure sensor at the wellhead, circulating fluid line, or standpipe of the second well in addition to the electrical signal from the acoustic vibrations.
24 . The method of claim 13 , wherein the adjusting comprises at least reducing fracking fluid pressure.
25 . The method of claim 13 , wherein the adjusting comprises at least terminating a current fracking stage.
26 . A method of training a machine-learning model to predict fracture communication between wells, the method comprising:
providing a sensor coupled to a wellhead, circulating fluid line, or standpipe of a first well and configured to convert acoustic vibrations in fluid in the first well into an electrical signal in a time domain;
recording the electrical signal to a memory;
analyzing current frequency components of the electrical signal in a window of time;
identifying, via a machine-learning system, impending fracture communication between the first well and an offset well where the current frequency components show sufficient similarity to previous frequency components measured during instances of fracture communication between wells; and
wherein the machine-learning system is trained by grouping the current frequency components with similar ones of the previous frequency components that are also associated with fracture communication between wells.
27 . The method of claim 26 , wherein the acoustic vibrations are caused by injecting fracking fluid through perforations in a casing of the offset well under pressure in order to form subsurface fractures, the fracking fluid's flow through subsurface fractures, expansion of subsurface fractures, or a pumping truck at the offset well.
28 . The method of claim 26 , wherein the sensor samples at greater than 1 kHz.
29 . The method of claim 26 , wherein the sensor is an acoustic sensor.
30 . The method of claim 26 , further comprising converting the electrical signal in the window of time to a frequency spectrum and analyzing the current frequency components in the frequency spectrum to identify impending fracture communication between the first well and the offset well.
31 . The method of claim 26 , wherein the analyzing and the identifying is performed by a machine-learning system.
32 . The method of claim 31 , wherein the machine-learning system is trained on the previous frequency components of electrical signals measured during previous instances of fracture communication between wells.
33 . The method of claim 32 , wherein the machine-learning system is trained on pressure data measured during previous instances of fracture communication between wells.
34 . The method of claim 26 , wherein the analyzing comprises monitoring increases in amplitude of the current frequency components.
35 . The method of claim 26 , further comprising analyzing the electrical signal in the time domain to identify impending fracture communication between the first well and the offset well.
36 . The method of claim 26 , further comprising analyzing electrical signals from a pressure sensor at the wellhead, circulating fluid line, or standpipe of the first well in addition to the electrical signal from the acoustic vibrations.
37 . The method of claim 26 , further comprising adjusting parameters of subsequent fracking operations to reduce chances of fracture communication between the first well and the offset well.
38 . The method of claim 37 , wherein the adjusting comprises at least reducing fracking fluid pressure.
39 . The method of claim 37 , wherein the adjusting comprises at least terminating a current fracking stage.
40 . A method of spacing wells to avoid fracture communication between the wells, the method comprising:
performing a fracking operation on a first well in a subterranean formation;
providing a sensor coupled to a wellhead, circulating fluid line, or standpipe of a second well and configured to convert acoustic vibrations in fluid in the second well into an electrical signal in a time domain;
recording the electrical signal to a memory;
analyzing frequency components of the electrical signal in a window of time;
identifying fracture communication between the first and second wells where the frequency components show sufficient similarity to previous frequency components measured during previous instances of fracture communication between wells; and
adjusting a spacing of new wells based on the identifying fracture communication between the first and second wells.