Patent Application
20190138963
Data is often collected on heavy equipment such as construction equipment or drilling rigs, for instance, those involved in oil and gas exploration and production. Such data may include surface-based measurements about the equipment for any given day or taken at any time interval, such as rig location (e.g., Global Position Satellite (GPS), TRS (township, range, and section information), latitude, and longitude), well American Petroleum Institute (API) number, well surface location, a commodity basin location, well lateral length, an expected ultimate recovery at the commodity basin location, geography at the project location, basin distance from current rig position, well or target formation, formation depth, and commodity type (oil or gas). Additional data about the drilling site or process may be measured and recorded (commonly known as measurements while drilling (MWD)) and may include daily oil price, production data, permit approval date and expiration, permit depth, rig spud date, rig release date, permit location, and state regulations affecting the distance between wells or drilling spacing units (DSU). Recorded data may further include information about the equipment itself, such as operator name and information, rig name or rig identification (ID), rig capabilities such as drill horse power, maximum movement speed, and other rig capabilities.
Users of heavy equipment and the companies interested in their operation typically use human intelligence to analyze a terrain or a given portion of land to determine how long it will take to drill a preset amount of wells in a predetermined size of the land. However, due to limitations in using such techniques, the results have been disappointing and require a lengthy amount of time to obtain. Such results include the timeline of how long it will take to complete a task, e.g., drill a preset amount of wells for a predetermined size of terrain, and require human subjective intervention.
The following presents a simplified summary of the disclosure in order to provide a basic understanding of some aspects of the disclosure.
The systems and methods disclosed herein may, among other things, allow for the determination of: a number of rigs used in a potential project for a third party operator, number of wells dug in a potential project, well density for a third party operator, and rig movement for a particular operator, time of movement, and/or time to complete a task, all determined by a self-learning manner, to predict the total time to complete a potential project, e.g., drill a preset amount of wells in a defined area of terrain or move to a specific location or area to drill a well.
In embodiments, the method described herein is a predictive method based on previous human decisions that were made at the beginning and during a previous project involving heavy equipment (e.g., drilling wells). The method is faster, more efficient, more flexible, and improves computer technology due to the contribution of the self-learning aspect of the disclosure. To be able to accurately predict the total time for a project, the systems and methods discussed below calculate for a potential project: a rig count, well density, quantity of wells, rig movement, time to complete drilling of all wells, and time of rig movement. The method presents a prediction potentially without foreknowledge of the number of heavy equipment items, or rig movement that will be used and without knowledge of how many wells will be ultimately dug. The method uses historical data to create a model that is regularly updated as additional data is gathered.
More specifically, in an embodiment, a self-learning method to assist in determining a total time to complete a potential project involves a plurality of historical heavy equipment information. The method is implemented on one or more computing systems, which includes a set of rules for: (1) receiving a request for an estimated total time to complete a potential project from a user; (2) accessing a first portion of the historical heavy equipment information for heavy equipment for the operator; (3) calculating how a particular number of rigs (i.e. one) will move in the potential project by the operator based on the first portion of historical heavy equipment information; (4) accessing a second portion of the historical heavy equipment information for the operator; (5) providing an updated rig movement path to be used in the potential project based on the second portion of the historical heavy equipment information; (6) predicting a total estimated time for the potential project utilizing the updated rig movement to be used in the potential project; and (7) sending, in response to the request, the estimated total time to complete the potential project.
In step (1), the potential project includes a quantity of a terrain, or acreage owned by an operator, an operator, and at least one geological property of the terrain. In step (2), the first portion of the historical heavy equipment information includes information concerning how the heavy equipment moved from a first position to a second position by the operator for a first historical project. In step (4), the second portion of the historical heavy equipment information includes information concerning the heavy equipment moved from the second position to a third position in a historical project.
In another embodiment, a self-learning system comprises a processor; an application programing interface that communicates with a heavy equipment database containing a plurality of historical heavy equipment data that is updated at a predetermined time interval; a storage medium for a storing portion of the historical heavy equipment data relating to a specific operator and rig movement; a calculating device for calculating, upon request by a user, a total time to complete a potential project based on the portion of historical heavy equipment information; and an output device for displaying the calculated total time for the potential project. The request by the user comprises information about the potential project including a quantity of a terrain, an operator, and geological property of the terrain. The calculating device updates the calculated total time for the potential project calculation with the data that is updated at the predetermined time interval.
In yet a further embodiment, a non-transitory computer readable medium with computer executable instructions stored thereon, executed by a digital processor to perform a self-learning method, includes instructions for: (1) receiving a request from a user for an estimated total time to complete a potential project, the request including information about the potential project including a quantity of a terrain, an operator, and geological property of the terrain; (2) accessing a first portion of the historical heavy equipment information for heavy equipment for the operator, wherein the first portion of the historical heavy equipment information includes information concerning the movement of the heavy equipment owned by the operator for a first historical project; (3) calculating how a particular number of rigs (i.e. one) will move in the potential project by the operator based on the first portion of historical heavy equipment information; (4) accessing a second portion of the historical heavy equipment information for the operator; (5) providing an updated rig movement path to be used in the potential project based on the second portion of the historical heavy equipment information; (6) predicting a total estimated time for the potential project utilizing the updated rig movement to be used in the potential project; and (7) sending, in response to the request, the estimated total time to complete the potential project.
The drawings constitute a part of this specification and may include exemplary embodiments of the present disclosure and illustrate various objects and features thereof.
A further understanding of the disclosure may be recognized by reference to the accompanying drawing in which:
Reference is made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. It is to be understood that other embodiments may be utilized and structural and functional changes may be made. Moreover, features of the various embodiments may be combined or altered. As such, the following description is presented by way of illustration only and should not limit in any way the various alternatives and modifications that may be made to the illustrated embodiments. In this disclosure, numerous specific details provide a thorough understanding of the subject disclosure. It should be understood that aspects of this disclosure may be practiced with other embodiments not necessarily including all aspects described herein, and the like.
The structure 102 may include a processor 106, which may be in data communication with a network interface 108, an input device 110, an output device 112, and a memory 114. Processor 106 represents one or more digital processors. Network interface 108 may be implemented as one or both of a wired network interface and a wireless network interface, as is known in the art. The input device 110 may include a keyboard, a mouse, a stylus pen, buttons, knobs, switches, and/or any other device that may allow a user to provide an input to the system 100 via the structure 102. In some embodiments, the input device 110 may comprise a media port (such as a USB port or a SD or microSD port) to allow for media (e.g., a USB drive, a SD or micro SD drive, a laptop memory, a smart phone memory, and the like) to be communicatively coupled to the structure 102. The output device 112 may include one or more visual indicators (e.g., a display, touch screen), audible indicators (e.g., speakers), or any other such output device now known or subsequently developed. In some embodiments, at least a part of the input device 110 and the output device 112 may be combined.
Although shown within the structure 102, memory 114 may be, at least in part, implemented as network storage that is external to the structure 102 and accessed via the network interface 108. The memory 114 may house software 116, which may be stored in a transitory or non-transitory portion of the memory 114. Software 116 includes machine readable instructions that are executed by processor 106 to perform the functionality described herein. In some example embodiments, the processor 106 may be configured through particularly configured hardware, such as an application specific integrated circuit (ASIC), field-programmable gate array (FPGA), and the like, and/or through execution of software (e.g., software 116) to perform functions in accordance with the disclosure herein.
The software 116 may include a self-learning project time determination tool 118, which may be configured to estimate a total time to complete a project including multiple drill sites or it may be used to estimate how long an operator will take to get from a starting location to a secondary location as will be further discussed below. The tool 118 may, in some embodiments, be an application 158, such as a mobile application configured for an Android, Apple, or other device, or as a computer application 158 configured for a mobile computer 134, such as a desktop, laptop, or mobile computer, and the like.
The mobile computer 134 includes a processor 138 in communication with memory 142. In one embodiment, computer 134 is a stationary computer. A user 136 may download the application or program 158 onto computer 134 that enables the computer 134 to communicate with the structure 102 via API 132B. The application 158 is software stored in a non-transitory portion of memory 142, and includes machine readable instructions that are executed by processor 138 to improve functionality of computer 134 and to allow communication with structure 102. As discussed herein, in embodiments, application 158 may provide a graphical user interface 160 that prompts the structure 102 to initiate the total time project determination tool 118. Alternately or additionally, in some embodiments, the tool 118 may be accessible over a network 140B (e.g., over the internet via a password protected or other website, over an intranet, and the like).
The tool 118 may include one or more of a data acquisition module 120, a rig count module 122, a well density module 124, a rig movement module 126, a graphical user interface (GUI) module 130, and API 132A, 132B, each of which are described in more detail herein. A module may be a portion of a computer program and may include instructions for performing a particular function.
The structure 102, via the API 132A, may selectively communicate over a network 140A with a heavy equipment database 150. The heavy equipment database 150 may be a storage medium, such as an optical hard drive, for storing heavy equipment information as data. Historic heavy equipment data may come in the form of records of rig positions (e.g., based on GPS positioning) which may be at discrete time intervals (e.g., daily, hours, by the minute, etc.). Each data point represents a rig position including, for example, degrees of latitude and longitude and may further include a township, range, and section (TRS) description, as well as other relevant information (e.g. heavy equipment type, rig capabilities, etc.). As discussed herein, the tool 118 may be used to import or receive the heavy equipment data and store the data in the memory 114 or database 150.
The structure 102 may be in communication with a solitary computer 134 being used by a user 136 (shown in dashed lines in
In one embodiment, the structure 102 is an online structure 102 which, using protocol 119 and application programming interface (API) 132B, may communicate over a wireless network 140B with the computer 134, such as a desktop computer, mobile computer, a laptop, notebook, tablet, smartphone, et cetera, with which a user 136 interacts. Protocol 119 may be any known internet protocols, such as, IPv6, IPv8, and the like used in the art now or those yet to be known.
Self-learning methods of estimating a total project time are now described with reference to
To begin, in step 202, the method 200 is initiated by activating the total time determination tool 118 by the user 136 (e.g. through the input device 110). The user 136 may be interacting with a GUI 160 initiated by the GUI module 130, as discussed above. The user 136 may define the potential project including the terrain and geological features and the desired operator to evaluate.
In step 204, an amount of rigs or rig number is predicted or determined. In one embodiment, the prediction may be made by the method 400 as discussed below (
In step 206, well density per drilling spacing unit (DSU) is predicted or determined by method 500 as discussed below (
In step 208, having estimated the number of rigs at step 204 (e.g., from method 400 in
A final determination (or estimation) needed in order to more accurately predict the total time to complete a project is how long it will take the operator to dig the determined amount of wells given the number of rigs being used for the project (or based on the estimated number of rigs). To estimate how long it will take, the method 200 must predict how the heavy equipment will move, as each move affects the overall timing of the project. For example, rigs moving in a straight line (or relatively straight line) may move quicker than rigs that follow a more non-linear path. Accordingly, in step 210, the rig movement is predicted by method 600 as discussed below (
Finally, in step 212, the program predicts a total time to complete the project via method 700 as discussed below (
Thus, the method 200 involves receiving known data (if available) or estimating, based on historical information, data for several variables known to have an effect on the timing it takes an operator to develop a predetermined tract of land. If the variable is known (i.e. rig count), then the sub-method concerning calculating that variable may be skipped. It is understood that each variable may include a prompt for a user 136 to input the variable, if known, known into the tool 118. The sub-methods of estimation are now described below, with reference to
Beginning with
In step 302, the sub-method 300 begins. In embodiments, the data acquisition module 120 is activated by the user 136. In step 304, the historical data from the heavy equipment database 150 (which, as noted, may be multiple databases) for every job or project previously tracked for each operator is downloaded from the heavy equipment database 150. Heavy equipment data may currently be retrieved from already existing databases such as Drilling Info (DI) or IHS Market.
In step 306, the data is filtered by the operator or the company that owns and utilizes the heavy equipment, to be used for the project and stored in memory 114. Once the heavy equipment data is grouped by operator, in step 308, the heavy equipment (or rigs) owned by those operators can be sorted based on specific rig data, which may include a time stamp (e.g. GPS latitude and longitude on 10/1/16). Once the data is sorted by rig or heavy equipment, then in step 310, a state table may be created (or updated, as the case may be) concerning each rig. A state table may show each change in state for each particular rig over a given period (e.g., for the time that records for the operator exists, or a subset thereof). Such state changes may include variables that may be tracked as data and that a drilling rig site operator may wish to review including but not limited to rig count in an area, well API numbers, rig location, well lateral length, permit date changes, rig spud dates, rig release dates, formation of well, and formation depth. The state table is saved to memory 114. As sub-method 300 is repeated, and for each repetition, only new data, if available, will be retrieved and the state table will be updated with the newest information. Due to the size of the data to be transferred, optionally, this sub-method 300 may only retrieve data for a single operator at step 304. Accordingly, the sub-method 300 would be repeated for every desired operator.
The sub-method 300 may be repeated daily or at a set time interval (e.g., hourly, weekly, and the like) in which data is updated in the heavy equipment database 150 to monitor the state changes for all operators and to update prediction models, as will be discussed further below.
Moving on,
The sub-method 400 is activated by the rig count module 122 in step 402. In step 404, an initial estimate for rig count is made. This estimate could be as simple as one rig or an average of the rig count over the total amount of projects for a given operator (computed from state table of historical data stored in memory 114) for example. Alternately, the estimate may be based on a number of rigs used for a similarly sized tract of land by the operator for a previous project.
At step 406, the sub-method 400 self-learns (or updates the estimates) by taking the initial estimate from step 404 and modeling it against the actual numbers in the historical data. Step 406 may include modeling in the form of linear regression, logarithmic regression, and/or non-linear regression using known methods (e.g., least square approach, least absolute deviation approach, random forest regression, boosted tree model, etc.). Regression may be used to fit a predictive model to observed rig count historical heavy equipment data and variables (e.g., the w variables as discussed above). W variables may be related to rig count, depending on the operator. Non-linear and/or linear regression can be applied to quantify the strength of the relationship between rig count and the w variables. One or more linear and/or non-linear regression methods may be used in step 406 to determine which modeling approach fits the data the best. The modeling can include subsets and not just one historical data set to estimate another data set. For example, four data sets may be used as a subset and may be modeled in order to estimate a fifth data set to compare the different modelling approaches.
After developing a fitted model, the model can be used to make a prediction of the value of rig number or count. In step 408, if further historical data exists in memory 114 or database 150 that would aid in developing the accuracy of rig number, then step 406 is repeated with the most recent historical heavy equipment data set for the current project (or another project) for the same operator.
It should be understood that the sub-method 400 may be repeated in its entirety for each operator. Step 408 is repeated until the historical data no longer includes relevant updated data. For example, a project may be finalized when all wells are dug and the commodity has been completely depleted or for all intents and purposes depleted and all the rigs at the location are moved off site to the next project (e.g. rig count is zero). No additional data may be received, because no further action is occurring. It may also be that the rig has reached a predetermined destination, after which location data is no longer updated.
The sub-method then moves to step 410, where a new rig number is determined. The sub-method ends at step 412. Therefore, the sub-method 400 is constantly learning with each iteration of the method 200, and the sub-method will re-determine the rig number with the latest data from the sub-method 300.
With reference now to
At step 506, the sub-method 500 updates the estimate by taking the initial estimate and modeling it against the actual numbers in the historical data for a given finalized project. Step 506 may include modeling in the form of linear regression, logarithmic regression, and/or non-linear regression using known methods (e.g., least square approach, least absolute deviation approach, random forest regression, boosted tree, etc.). Regression may be used to fit a predictive model to observed well density historical heavy equipment data and the w variables discussed above. W variables may be related to well density or not depending on the operator. Linear and non-linear regression can be applied to quantify the strength of the relationship between well density and the w variables. Some of these w variables may change as the project goes on and some will not be known at the initial outset of a project (e.g., well lateral length or distance of well from current rig position if there are no wells dug yet). Accordingly, the sub-method 500 may be completed with unknown w variables, which may be ignored (e.g., given a value of zero). The known w variables may be collected from the historical data from the heavy equipment database 150 as discussed in sub-method 300.
One or more linear and/or non-linear regression methods may be used in step 506 to determine which modeling approach fits the data the best. The modeling can include subsets and not just one historical data set to estimate another data set. For example, four data sets may be used as a subset which may be modeled to estimate a fifth data set to compare the different modelling approaches.
After developing a fitted model, the model can be used to make a prediction of the value of well density. In step 508, if further historical data exists in memory 114 or database 150 that would aid in further developing the accuracy of well density, then step 506 is repeated with the most recent historical heavy equipment data set for that single operator. As more data is accumulated, the accuracy of the method 500 is increased. Step 508 is repeated until the historical data no longer includes relevant updated data or data that would assist in predicting a well density of an operator.
The model determines a quantity for well density at step 510, and the method ends at step 512. Sub-method 500 may be repeated at the conclusion of each operator project, or at set time intervals, to learn and iteratively adapt based on the data retrieved from sub-method 300.
Moving on,
The sub-method 600 updates by taking the selection 804 and modeling the selection against the actual candidate location 804 selection in the historical data at step 606 for a given project. Step 606 may include modeling in the form of linear regression, logarithmic regression, and/or non-linear regression using known methods, (e.g., least square approach, least absolute deviation approach, random forest, boosted tree, etc.). Regression may be used to fit a predictive model to observed heavy equipment movement historical heavy equipment data and x variables. X variables that may be considered by the sub-method may include, but are not limited to: distance to candidate position from current position, number of candidate locations, cumulative production features, distance of other heavy equipment from candidate locations, proximity of target location to infrastructure elements (e.g., pipeline), estimated ultimate recovery of candidate positions, oil price at the time of selection, geological properties of the candidate positions, target formation at the candidate position, and permit expiration date for candidate positions. X variables may be related to rig movement, depending on the operator. Linear and non-linear regression can be applied to quantify the strength of the relationship between rig movement and the x variables.
One or more linear and/or non-linear regression methods may be used in step 606 to determine which modeling approach fits the data the best. The modeling can include subsets and not just one historical data set to estimate another data set. For example, four data sets may be used as a subset which may be modeled to estimate a fifth data set to compare the different modelling approaches.
Some of the x variables may change as the project continues and the sub-method 600 will self-learn from this. If the sub-method 600 is run with unknown x variables, these x variables are ignored (e.g., given a value of zero). The known x variables may be collected from the latest upload of iterative data from the heavy equipment database 150 as discussed regarding sub-method 300.
After developing a fitted model, the model can be used to make a prediction of the value of rig movement. In step 608, if further historical data exists that would aid in further developing the accuracy of rig movement, then step 606 is repeated with the most recent historical heavy equipment data set.
It should be understood that this sub-method 600 may be repeated in its entirety for one or more operators. Step 606 is repeated until the historical data no longer includes a relevant updated data or data that would assist in predicting rig movement of an operator. A prediction of rig movement for the entire project or from a starting location to a second location is determined at step 610. The method ends at step 612. Sub-method 600 may be repeated at the conclusion of each operator project, or at set time intervals, to learn and iteratively adapt based on the retrieved data from sub-method 300.
Referring now to
One or more linear and/or non-linear regression methods may be used in step 706 to determine which modeling approach fits the data the best. It is foreseen that the modeling can include subsets and not just one historical data set to estimate another data set. For example, four data sets may be used as a subset which may be modeled to estimate a fifth data set to compare the different modelling approaches.
In step 708, if further historical data exists in memory 114 that would aid in further developing the accuracy of Ttransport, then step 706 is repeated with the next historical heavy equipment data set of the next candidate location for that single operator. Step 706 is repeated until the historical data no longer includes data helpful in determining Ttransport. The time to transport a rig from location or position 800 to position 810 (
In step 712, an initial estimate for time of completing a task (e.g. dig a well) (Ttask) is made. This estimate could be as simple as the averages based on z variables, which may be a subset of w variables, such as: well depth, well lateral length, well location, geological properties at the well location, and rig capabilities for the heavy equipment of the operator (e.g., drill horse power, drill speed, and the like). The heavy equipment to be used is not known, so this portion of the sub-method 700 will be an estimation based on historical data of rigs previously used by the operator. Furthermore, when given specific geological properties of a candidate location, it is unknown how wide the well lateral length will be or how deep the well will be dug, so an estimation based on the historical data of how the rigs have performed under similar conditions (if available) will be used in step 712 to determine Ttask.
Step 714 updates by modeling the Ttask estimate against the actual time to complete drilling of wells in the historical data for a given finalized project. The step 714 may include modeling in the form of linear regression, logarithmic regression, and/or non-linear regression using known methods (e.g., least square approach, least absolute deviation approach, etc.). Regression may be used to fit a predictive model to observed time to complete single well historical heavy equipment data, and z variables. The z variables are presumed to impact Ttask, and linear regression may be applied to quantify the strength of the relationship between Ttask and the z variables.
One or more linear and/or non-linear regression methods may be used in step 714 to determine which modeling approach fits the data the best. The modeling can include subsets and not just one historical data set to estimate another data set. For example, four data sets may be used as a subset which may be modeled to estimate a fifth data set to compare the different modelling approaches.
In step 716, if more historical data exists in memory 114 that would aid in further developing the accuracy of Ttask, then step 714 is repeated with the next historical heavy equipment data set of the next candidate location for that single operator.
In step 718, the time to complete individual wells (Ttask) is determined. In step 720, the total time (Ttotal) for a potential well project can be estimated by the summing of all Ttask for each well and the Ttransport for each candidate location and each rig. Ttotal may also be a known quantity in the historical data, and can be comparably checked for accuracy (e.g., within the same order of magnitude). Therefore, in step 720, the calculated Ttotal is taken as an initial estimate.
The sub-method 700 continues at step 722 to update by taking the Ttotal estimate and then modeling the Ttotal estimate against the actual time to complete a finalized project in the historical data. Step 722 may include modeling in the form of linear regression, logarithmic regression, and/or non-linear regression using known methods (e.g., least square approach, least absolute deviation approach, etc.). The variables that are presumed to impact Ttotal may include: Ttask and Ttransport for a single operator. The closer the determinations of Ttask in step 718 and Ttransport in step 710 to the true answers in the historical data, the closer that Ttotal will also be. The final recursive step 724 may adjust for those small errors in seen the variables, if appropriate.
In step 724, if more historical data exists in memory 114 that would aid in further developing the accuracy of Ttotal, then step 722 is repeated. It should be understood that this step 722 may require the additional steps of the sub-method 700 to be repeated in their entirety. Step 724 may be just a quick double check to see if the data in method 300 has just been updated. Step 724 is repeated until the historical data no longer includes data helpful in determining Ttotal.
In step 726, the total time to complete a potential project with a given terrain and a specific operator is estimated, and the process ends at step 708. This solution may output to the user 136 via the output device 112 or GUI 160. The user 136 may utilize this knowledge, to better select an operator for a given piece of land or terrain with known geological properties. The total time determination taking into account the known geological properties of the terrain. It should be understood that this method 700 may be repeated in its entirety for one or more operators.
Many different arrangements of the various components depicted, as well as components not shown, are possible without departing from the spirit and scope of the present disclosure. Embodiments of the present disclosure have been described with the intent to be illustrative rather than restrictive. Alternative embodiments will become apparent to those skilled in the art that do not depart from its scope. A skilled artisan may develop alternative means of implementing the aforementioned improvements without departing from the scope of the present disclosure. Further, it will be understood that certain features and subcombinations may be of utility and may be employed within the scope of the disclosure. Further, various steps set forth herein may be carried out in orders that differ from those set forth herein without departing from the scope of the present methods. This description shall not be restricted to the above embodiments.
With reference to
Those of skill in the art shall understand that the graphical results may be relayed to a user 136, e.g., via the output device 112 or the GUI 160.
