Drilling intelligence guidance system for guiding a drill

1 . A non-transitory computer readable medium storing a computer readable program that causes a processor to:

receive, by a drill assembly machine learning model, a set of parameters characterizing sensor data from a plurality of sensors corresponding to drilling operations of a drill assembly for boring the Earth, and a condition of the drill assembly, wherein the condition is one of a failure condition or a non-failure condition;

select, by the drill assembly machine learning model, a subset of parameters of the set of parameters that are related to the condition of the drill assembly, wherein the drill assembly machine learning model uses a historical record of the set of parameters and a set of conditions of the drill assembly collected over time to determine whether a parameter of the set of parameters is related to a condition of the set of conditions;

apply, by the drill assembly machine learning model, weights to each parameter of the subset of parameters and weights to different levels of each parameter of the subset of parameters, producing a weighted subset of parameters; and

adjust, by a drill controller engine, operational parameters of a drilling controller for the drill assembly based on a computed damage index (DI) that is based on the weighted subset of parameters and represents a risk of failure of the drill assembly.

2 . The medium of claim 1 , wherein the each parameter of the set of parameters is weighted based on each condition of the set of conditions and the weighting of each parameter is used to determine an associated stress level, and wherein an increase in stress level is related to an increase in risk of failure of the drill assembly.

3 . The medium of claim 1 , wherein the computer readable program further causes the processor to:

analyze, by the drill assembly machine learning model, data stored with each parameter of the set of parameters to determine stress levels associated with each parameter over time; and

determine, by the drill assembly machine learning model, levels of each parameter that increase stress level of a component of the drill assembly, wherein the weights applied to the different levels of each parameter of the subset of parameters are related to the increase in stress level, and each level of the levels corresponds to a range of values for a respective parameter.

4 . The medium of claim 1 , wherein the computer readable program further causes the processor to:

match the computed DI to a plurality of DIs from the historical record; and

categorize the risk of failure of the drill assembly as low, medium or high based on the matching.

5 . The medium of claim 4 , wherein:

a low risk level corresponds to a DI below a sub-optimal threshold;

a medium risk level corresponds to a DI above the sub-optimal threshold and below a failure threshold; and

a high risk level corresponds to a DI above the failure threshold.

6 . The medium of claim 1 , wherein the computer readable program further causes the processor to:

predict a future risk level at a future point in time based on DIs calculated for two or more previous instances from the historical record.

7 . The medium of claim 1 , wherein the drill assembly machine learning model comprises a random forest decision tree.

8 . The medium of claim 1 , wherein the subset of parameters relates to a condition of a motor of the drill assembly and comprises differential pressure, number of rotary-to-slide transitions, back reaming time and time in hole during drilling operations.

9 . The medium of claim 1 , wherein the computer readable program further causes the processor to:

aggregate the set of parameters and the condition of the drill assembly into a historical record over time, wherein the historical record comprises a first class of parameters correlating to non-failure conditions of the drill assembly and a second class of parameters correlating to failure conditions of the drill assembly; and

balance the historical record by generating synthetic examples of failure conditions for the second class of parameters using a Synthetic Minority Oversampling Technique.

10 . A system for analyzing drilling operations, comprising:

a processor; and

a memory storing instructions that, when executed by the processor, cause the processor to:

receive, at a drill assembly machine learning model, a set of parameters characterizing sensor data from a plurality of sensors corresponding to drilling operations of a drill assembly for boring the Earth, and a condition of the drill assembly;

select, using the drill assembly machine learning model, a subset of parameters of the set of parameters that are related to the condition of the drill assembly, wherein the drill assembly machine learning model uses a historical record of the set of parameters and a set of conditions of the drill assembly collected over time to determine whether a parameter of the set of parameters is related to a condition of the set of conditions;

apply, using the drill assembly machine learning model, weights to each parameter of the subset of parameters and weights to different levels of each parameter of the subset of parameters, producing a weighted subset of parameters; and

adjust, by a drill controller engine, operational parameters of a drilling controller for the drill assembly based on a computed damage index (DI) that is based on the weighted subset of parameters and represents a risk of failure of the drill assembly.

11 . The system of claim 10 , wherein the instructions further cause the processor to:

aggregate the set of parameters and associated drilling conditions into the historical record over time; and

train the drill assembly machine learning model using the historical record.

12 . The system of claim 11 , wherein the historical record comprises:

a first class of parameters correlating to non-failure conditions of the drill assembly; and

a second class of parameters correlating to failure conditions of the drill assembly;

wherein the historical record is balanced by generating synthetic examples of failure conditions for the second class of parameters.

13 . The system of claim 12 , wherein the instructions further cause the processor to:

oversample the second class of parameters using a Synthetic Minority Oversampling Technique to generate the synthetic examples of the failure conditions.

14 . The system of claim 10 , wherein the instructions further cause the processor to:

match the computed DI to a plurality of DIs from the historical record; and

categorize the risk of failure of the drill assembly as low, medium or high based on the matching.

15 . The system of claim 10 , wherein the instructions further cause the processor to:

predict a future risk level at a future point in time based on DIs calculated for two or more previous instances from the historical record.

16 . The system of claim 10 , wherein the drill assembly machine learning model comprises a random forest decision tree.

17 . A computer-implemented method for analyzing drilling operations, comprising:

receiving, by a processor, a set of parameters characterizing sensor data from a plurality of sensors corresponding to drilling operations of a drill assembly for boring the Earth, and a condition of the drill assembly;

selecting, by a drill assembly machine learning model executed by the processor, a subset of parameters of the set of parameters that are related to the condition of the drill assembly, wherein the drill assembly machine learning model uses a historical record of the set of parameters and a set of conditions of the drill assembly collected over time to determine whether a parameter of the set of parameters is related to a condition of the set of conditions;

applying, by the drill assembly machine learning model, weights to each parameter of the subset of parameters and weights to different levels of each parameter of the subset of parameters, producing a weighted subset of parameters; and

adjusting, by a drill controller engine, operational parameters of a drilling controller for the drill assembly based on a computed damage index (DI) that is based on the weighted subset of parameters and represents a risk of failure of the drill assembly.

18 . The method of claim 17 , further comprising:

matching the computed DI to a plurality of DIs from the historical record; and

categorizing the risk of failure of the drill assembly as low, medium or high based on the matching.