Sonic log synthesis
Aspects of the subject technology relate to systems, methods, and computer-readable media for synthesizing sonic logs for downhole environments. Data associated with a wellbore in a formation is accessed. A model configured to identify one or more characteristics associated with sound traveling through one or more formation is accessed. The data is applied to the model to predict a characteristics of the formation that is identifiable through sonic logging.
1 . A method comprising:
accessing data associated with a wellbore in a formation;
accessing a model configured to identify one or more characteristics associated with sound traveling through one or more formations;
applying the data to the model to predict a characteristic of the formation that is identifiable through sonic logging;
synthesizing sonic logs based on the predicted characteristic of the formation by substituting the predicted characteristic of the formation for the data missing from the sonic logs;
selecting a type of drill bit for drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs; and
controlling a speed of the drill bit while drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs.
2 . The method of claim 1 , wherein the characteristic of the formation that is identifiable through sonic logging includes one of compressional wave slowness of a portion of the formation, shear wave slowness of the portion of the formation, a ratio of the compressional wave slowness of the portion of the formation to the shear wave slowness of the portion of the formation, Young's modulus of the portion of the formation, Poisson's ratio of the portion of the formation, and vibrational types associated with the portion of the formation.
3 . The method of claim 1 , wherein the characteristic of the formation that is identifiable through sonic logging serves as a substitute for data missing from a sonic log of the formation.
4 . The method of claim 1 , wherein the data associated with the wellbore includes drilling data associated with operation of a drill string in drilling the wellbore and the drilling data comprises torque, weight-on-bit, mud pumping rate, rotating speed, drilling fluid pumping pressure, rate of penetration, peak axial vibrations, absolute average vibrations, lateral vibrations, average pitch, average yaw, depth of cut downhole, and downhole temperature.
5 . The method of claim 1 , wherein the model is trained with drilling data associated with drilling one or more wellbores and sonic log data generated for formations around the one or more wellbores.
6 . The method of claim 5 , wherein the drilling data comprises values of a plurality of drilling variables, the method further comprising:
identifying redundant variables amongst the plurality of drilling variables based on correlation between the plurality of drilling variables;
remove one of corresponding redundant variables in the identified redundant variables to generate modified drilling data; and
training the model based on the modified drilling data.
7 . The method of claim 5 , wherein the drilling data comprises data across a specific range, the method further comprising:
identifying a subset of the specific range in which to confine the drilling data;
confining the drilling data to the subset of the specific range to generate modified drilling data; and
training the model based on the modified drilling data.
8 . The method of claim 5 , further comprising:
mapping the drilling data across varying depth levels of the wellbore to generate mapped drilling data;
remove outliers in the mapped drilling data across the varying depth levels through a moving filter applied across the varying depth levels to generate modified mapped drilling data;
interpolating the drilling data across different drilling variables and the varying depth levels to further generate the modified mapped drilling data; and
training the model based on the modified mapped drilling data.
9 . The method of claim 5 , further comprising:
mapping the drilling data across varying depth levels of the wellbore to generate mapped drilling data;
mapping the sonic log data across the varying depth levels of the wellbore to generate mapped sonic log data;
interpolating the mapped sonic log data with the mapped drilling data to create corresponding drilling data points and sonic log data points at different depth levels of the varying depth levels of the wellbore; and
training the model based on the corresponding drilling data points and sonic log data points.
10 . The method of claim 9 , further comprising:
identifying a corresponding first drilling data point and a first sonic log data point at a specific depth of the corresponding drilling data points and sonic log data points;
accessing one or more adjacent drilling data points and one or more adjacent sonic log data points at one or more adjacent depths to the specific depth;
training the model based on:
the corresponding first drilling data point and the first sonic log data point, the one or more adjacent drilling data points, and the one or more sonic log data points; and
a parameter that the one or more adjacent drilling data points and the one or more adjacent sonic log data points are related to the first drilling data point and the first sonic log data point based on depth.
11 . The method of claim 5 , further comprising:
scaling the drilling data and the sonic log data to generate scaled drilling data and scaled sonic log data; and
training the model based on the scaled drilling data and the scaled sonic log data.
12 . The method of claim 1 , wherein the data associated with the wellbore comprises log data of either or both the wellbore itself and a neighboring wellbore.
13 . The method of claim 1 , wherein the model is trained with log data that is distinct from sonic log data for either or both the wellbore itself and a neighboring wellbore.
14 . The method of claim 13 , wherein the log data that is district from the sonic log data comprises caliper log data, neutron log data, gamma ray log data, deep resistivity log data, medium resistivity log data, shallow resistivity log data, photoelectric factor and density log data, or a combination thereof.
15 . A system comprising:
one or more processors; and
at least one computer-readable storage medium having stored therein instructions which, when executed by the one or more processors, cause the one or more processors to:
access data associated with a wellbore in a formation;
access a model configured to identify one or more characteristics associated with sound traveling through one or more formations;
apply the data to the model to predict a characteristic of the formation that is identifiable through sonic logging;
synthesizing sonic logs based on the predicted characteristic of the formation by substituting the predicted characteristic of the formation for the data missing from the sonic logs;
select a type of drill bit for drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs; and
control a speed of the drill bit while drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs.
16 . The system of claim 15 , wherein the data associated with the wellbore includes drilling data associated with operation of a drill string in drilling the wellbore and the drilling data comprises torque, weight-on-bit, mud pumping rate, rotating speed, drilling fluid pumping pressure, rate of penetration, or a combination thereof.
17 . The system of claim 15 , wherein the model is trained with drilling data associated with drilling one or more wellbores and sonic log data generated for formations around the one or more wellbores.
18 . The system of claim 15 , wherein the data associated with the wellbore comprises log data of either or both the wellbore itself and a neighboring wellbore.
19 . The system of claim 15 , wherein the model is trained with log data that is distinct from sonic log data for either or both the wellbore itself and a neighboring wellbore.
20 . A non-transitory computer-readable storage medium having stored therein instructions which, when executed by one or more processors, cause the one or more processors to:
access data associated with a wellbore in a formation;
access a model configured to identify one or more characteristics associated with sound traveling through one or more formations;
apply the data to the model to predict a characteristic of the formation that is identifiable through sonic logging;
synthesizing sonic logs based on the predicted characteristic of the formation by substituting the predicted characteristic of the formation for the data missing from the sonic logs;
select a type of drill bit for drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs; and
control a speed of the drill bit while drilling the wellbore through the formation based on the predicted characteristic and the synthesized sonic logs.