Method for generating predicted ultrasonic measurements from sonic data
A method, computer program product, and computing system are provided for receiving sonic data associated with an inner casing of a well. Predicted ultrasonic data associated with an outer casing of the well may be generated based upon, at least in part, a nonlinear regression model and the received sonic data associated with the inner casing of the well.
1. A computer-implemented method for predicting ultrasonic data associated with a well, the method comprising:
receiving, using one or more processors, sonic data associated with an inner casing of the well; and
generating predicted ultrasonic data associated with an outer casing of the well based upon, at least in part, a nonlinear regression model and the received sonic data associated with the inner casing of the well, wherein the nonlinear regression model is generated using one or more neural networks, wherein generating the nonlinear regression model using the one or more neural networks comprises: receiving measured ultrasonic data associated with the outer casing of the well at a plurality of well depths; identifying corresponding sonic data at the plurality of well depths from the received sonic data associated with the inner casing of the well, thus defining a plurality of sonic data and ultrasonic data pairs defined for the plurality of well depths, and training the one or more neural networks with the plurality of sonic data and ultrasonic data pairs defined for the plurality of well depths.
2. The computer-implemented method of claim 1 , wherein generating the nonlinear regression model using the one or more neural networks includes:
calibrating the trained one or more neural networks for generating predicted ultrasonic data on one or more of a different well and a different well depth, via transfer learning.
3. The computer-implemented method of claim 1 , further comprising:
estimating an inner casing standoff from the outer casing of the well based upon, at least in part, the measured ultrasonic data.
4. The computer-implemented method of claim 1 , wherein the one or more neural networks include at least one of:
one or more convolutional neural networks; and
one or more fully-connected neural networks.
5. The computer-implemented method of claim 1 , wherein the received sonic data includes a plurality of sonic waveforms generated via a sonic scanning tool.
6. The computer-implemented method of claim 1 , wherein the predicted ultrasonic data is selected from a group consisting of: a plurality of ultrasonic flexural attenuation (UFAK) measurements, acoustic impedance (AIBK), and high frequency sonic cement bond logging (CBL).
7. A computing system including one or more processors and one or more memories configured to perform operations comprising:
receiving sonic data associated with an inner casing of the well; and
generating predicted ultrasonic data associated with an outer casing of the well based upon, at least in part, a nonlinear regression model and the received sonic data associated with the inner casing of the well, wherein the computing system is configured to generate the nonlinear regression model using one or more neural networks by: receiving measured ultrasonic data associated with the outer casing of the well at a plurality of well depths, identifying corresponding sonic data at the plurality of well depths from the received sonic data associated with the inner casing of the well, thus defining a plurality of sonic data and ultrasonic data pairs defined for the plurality of well depths, and training the one or more neural networks with the plurality of sonic data and ultrasonic data pairs defined for the plurality of well depths.
8. The computing system of claim 7 , wherein generating the nonlinear regression model using the one or more neural networks includes:
calibrating the trained one or more neural networks for generating predicted ultrasonic data on one or more of a different well and a different well depth, via transfer learning.
9. The computing system of claim 7 , further configured to perform operations comprising:
estimating an inner casing standoff from the outer casing of the well based upon, at least in part, the measured ultrasonic data.
10. The computing system of claim 7 , wherein the one or more neural networks include at least one of:
one or more convolutional neural networks; and
one or more fully-connected neural networks.
11. The computing system of claim 7 , wherein the received sonic data includes a plurality of sonic waveforms generated via a sonic scanning tool.
12. The computing system of claim 7 , wherein the predicted ultrasonic data includes but are not limited to a plurality of ultrasonic flexural attenuation (UFAK) measurements, acoustic impedance (AIBK) measurements, and high frequency sonic cement bond logging (CBL) measurements.