DAS data processing to identify fluid inflow locations and fluid type
A method of identifying inflow locations along a wellbore comprises obtaining an acoustic signal from a sensor within the wellbore, determining a plurality of frequency domain features from the acoustic signal, and identifying, using a plurality of fluid flow models, a presence of at least one of a gas phase inflow, an aqueous phase inflow, or a hydrocarbon liquid phase inflow at one or more fluid flow locations. The acoustic signal comprises acoustic samples across a portion of a depth of the wellbore, and the plurality of frequency domain features are obtained across a plurality of depth intervals within the portion of the depth of the wellbore. Each fluid flow model of the plurality of fluid inflow models uses one or more frequency domain features of the plurality of the frequency domain features, and at least two of the plurality of fluid flow models are different.
1 . A method of developing a fluid flow model for a wellbore, the method comprising:
performing a plurality of flow tests, wherein each flow test comprises introducing one or more fluids of a plurality of fluids into a flowing fluid within a conduit at predetermined locations, wherein the plurality of fluids comprises a hydrocarbon gas, a hydrocarbon liquid, an aqueous fluid, or a combination thereof;
obtaining an acoustic signal from a sensor within the conduit for each flow test of the plurality of flow tests at a plurality of length intervals, wherein the acoustic signal comprises acoustic samples across a portion of the conduit including the predetermined locations and wherein a plurality of frequency domain features are determined for each length interval of the plurality of length intervals from a frequency domain representation of the acoustic signal in the time domain;
determining the plurality of frequency domain features from the acoustic signal for each of the plurality of fluid flow tests, wherein the plurality of frequency domain features are obtained across the portion of the conduit including the predetermined locations; and
training a plurality of fluid flow models using, as inputs, the plurality of frequency domain features for the plurality of the tests at each length interval of the plurality of length intervals and an identification of the one or more fluids introduced into the conduit, where a first fluid flow model of the plurality of fluid flow models is different than a second fluid flow model of the plurality of fluid flow models.
2 . The method of claim 1 , further comprising:
validating the plurality of fluid flow models using the acoustic signals from one or more of the tests and the predetermined locations of the plurality of tests.
3 . The method of claim 1 , wherein the conduit comprises a flow loop, and wherein the flowing fluid comprises an aqueous fluid, a hydrocarbon fluid, a gas, or a combination thereof.
4 . The method of claim 1 , wherein the flowing fluid comprises a liquid phase, a multi-phase mixed liquid, or a liquid-gas mixed phase.
5 . The method of claim 1 , wherein the plurality of fluid flow models comprises logistic regression models, and wherein training the plurality of fluid flow models comprises:
providing the one or more frequency domain features to a first logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise a hydrocarbon gas;
providing the one or more frequency domain features to the first logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise a hydrocarbon gas; and
determining a first multivariate model using the one or more frequency domain features as inputs, wherein the first multivariate model defines a relationship between a presence and an absence of the hydrocarbon gas in the one or more fluids.
6 . The method of claim 5 , wherein training the plurality of fluid flow models comprises:
providing the one or more frequency domain features to the second logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise an aqueous fluid;
providing the one or more frequency domain features to the second logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise an aqueous fluid; and
determining a second multivariate model using the one or more frequency domain features as inputs, wherein the second multivariate model defines a relationship between a presence and an absence of the aqueous fluid in the one or more fluids.
7 . The method of claim 6 , wherein training the plurality of fluid flow models comprises:
providing the one or more frequency domain features to a third logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise a hydrocarbon liquid;
providing the one or more frequency domain features to the third logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise a hydrocarbon liquid; and
determining a third multivariate model using the one or more frequency domain features as inputs, wherein the third multivariate model defines a relationship between a presence and an absence of the hydrocarbon liquid in the one or more fluids.
8 . The method of claim 1 , wherein the one or more frequency domain features comprise a normalized variant of the spectral spread (NVSS) and a normalized variant of the spectral centroid (NVSC).
9 . The method of claim 8 , further comprising:
providing the acoustic signals from one or more of the plurality of flow tests and the predetermined locations of the one or more of the plurality of flow tests to each of the first multivariate model, the second multivariate model, and the third multivariate model;
determining a presence or absence of at least one of the gas in the one or more fluids, the aqueous fluid in the one or more fluids, or the hydrocarbon liquid in the one or more fluids based on an output of each of the first multivariate model, the second multivariate model, and the third multivariate model; and
validating the fluid flow model using at least a portion of the plurality of flow tests, the predetermined locations of the plurality of inflow tests, and the presence of at least one of the gas in the one or more fluids, the aqueous fluid in the one or more fluids, or the hydrocarbon liquid in the one or more fluids as determined from the first multivariate model, the second multivariate model, and the third multivariate model.
10 . The method of claim 1 , wherein the plurality of fluid flow models is developed using a supervised learning algorithm.
11 . The method of claim 1 , wherein the flowing fluid within the conduit has a flow regime selected from the group consisting of: laminar flow, plugging flow, slugging flow, annular flow, turbulent flow, mist flow, and bubble flow.
12 . The method of claim 1 , wherein the sensor is disposed along the length of the conduit, and wherein the acoustic signal is indicative of an acoustic source along a length of the conduit.
13 . A system of developing a fluid flow model for a wellbore, the system comprising:
a processor; and
a memory comprising a flow application, wherein the flow application, when executed on the processor, configures the processor to:
receive an acoustic signal from a sensor within a conduit for each flow test of a plurality of flow tests at a plurality of length intervals, wherein each flow test comprises introducing one or more fluids of a plurality of fluids into a flowing fluid within a conduit at predetermined locations, wherein the plurality of fluids comprise a hydrocarbon gas, a hydrocarbon liquid, an aqueous fluid, or a combination thereof, wherein the acoustic signal comprises acoustic samples across a portion of the conduit including the predetermined locations, and wherein a plurality of frequency domain features are determined for each length interval of the plurality of length intervals from a frequency domain representation of the acoustic signal in the time domain;
determine the plurality of frequency domain features from the acoustic signal for each of the plurality of fluid flow tests, wherein the plurality of frequency domain features are obtained across the portion of the conduit including the predetermined locations; and
train a plurality of fluid flow models using, as inputs, the plurality of frequency domain features for the plurality of the tests at each length interval of the plurality of length intervals and an identification of the one or more fluids introduced into the conduit, where a first fluid flow model of the plurality of fluid flow models is different than a second fluid flow model of the plurality of fluid flow models.
14 . The system of claim 13 , wherein the processor is further configured to:
validate the plurality of fluid flow models using the acoustic signals from one or more of the tests and the predetermined locations of the plurality of tests.
15 . The system of claim 13 , wherein the conduit comprises a flow loop, and wherein the flowing fluid comprises an aqueous fluid, a hydrocarbon fluid, a gas, or a combination thereof.
16 . The system of claim 13 , wherein the flowing fluid comprises a liquid phase, a multi-phase mixed liquid, or a liquid-gas mixed phase.
17 . The system of claim 13 , wherein the plurality of fluid flow models comprises logistic regression models, and wherein the processor is further configured to:
provide the one or more frequency domain features to a first logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise a hydrocarbon gas;
provide the one or more frequency domain features to the first logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise a hydrocarbon gas; and
determine a first multivariate model using the one or more frequency domain features as inputs, wherein the first multivariate model defines a relationship between a presence and an absence of the hydrocarbon gas in the one or more fluids.
18 . The system of claim 17 , wherein the processor is further configured to:
provide the one or more frequency domain features to the second logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise an aqueous fluid;
provide the one or more frequency domain features to the second logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise a aqueous fluid; and
determine a second multivariate model using the one or more frequency domain features as inputs, wherein the second multivariate model defines a relationship between a presence and an absence of the aqueous fluid in the one or more fluids.
19 . The system of claim 18 , wherein the processor is further configured to:
provide the one or more frequency domain features to a third logistic regression model of the logistic regression models corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids comprise a hydrocarbon liquid;
provide the one or more frequency domain features to the third logistic regression model corresponding to one or more flow tests of the plurality of inflow tests where the one or more fluids do not comprise a hydrocarbon liquid; and
determine a third multivariate model using the one or more frequency domain features as inputs, wherein the third multivariate model defines a relationship between a presence and an absence of the hydrocarbon liquid in the one or more fluids.
20 . The system of claim 13 , wherein the processor is further configured to train plurality of fluid flow models using a supervised learning algorithm.