DETAILED ACTION
Claims 1-21 are presented for examination.
This office action is in response to the arguments submitted on 23-OCT-2025.
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
MPEP 609.04(b)
V. STATEMENT UNDER 37 CFR 1.97(e)
“In the alternative, a statement can be made if no item of information contained in the information disclosure statement was cited in a communication from a foreign patent office in a counterpart foreign application and, to the knowledge of the person signing the statement after making reasonable inquiry, neither was it known to any individual having a duty to disclose more than three months prior to the filing of the statement. If an inventor of the U.S. application is also a named inventor of one of the items of information contained in the IDS, the 37 CFR 1.97(e)(2) statement cannot be made for that particular item of information, and if made, will not be accepted.
The phrase "after making reasonable inquiry" makes it clear that the individual making the statement has a duty to make reasonable inquiry regarding the facts that are being stated. The statement can be made by a registered practitioner who represents a foreign client and who relies on statements made by the foreign client as to the date the information first became known. A registered practitioner who receives information from a client without being informed whether the information was known for more than three months, however, cannot make the statement under 37 CFR 1.97(e)(2) without making reasonable inquiry. For example, if an inventor gave a publication to the attorney prosecuting an application with the intent that it be cited to the Office, the attorney should inquire as to when that inventor became aware of the publication and should not submit a statement under 37 CFR 1.97(e)(2) to the Office until a satisfactory response is received. The statement can be based on present, good faith knowledge about when information became known without a search of files being made.”
Inventor Suk Kyoon Choi is cited as one of the authors in the IDS dated 12/04/2025 “SPE-228040-MS Unconventional Inflow Performance Relationship and Machine Learning Study” (hereinafter ‘publication’). The paper is a presentation for the SPE Annual Technical Conference and Exhibition held in Houston, Texas, USA, 20 - 22 October 2025. Call for proposals are typically months prior to the event. The October 2026 convention currently has proposals due March 9 2026. The publication cites many of the references provided in the IDS. The inquiry is when that inventor became aware of the publication. The Non-Final Rejection was mailed 07/23/2025. If the Applicant had submitted a proposal of the paper prior to the Non-Final Rejection and then submitted an IDS with the cited references months later, a fee would be required under 37 CFR 1.17(v) because the inventor was aware of the publication.
The information disclosure statement filed 10/09/2025 and 12/04/2025 fails to comply with the provisions of 37 CFR 1.97(a) because it lacks the appropriate fee set forth in 37 CFR 1.17(v). It has been placed in the application file, but the information referred to therein has not been considered as to the merits.
If the Applicant can provide justification for how the inventor was unaware of the publications until December 2025 while presenting in at a conference in October 2025 with proposals due months earlier, Applicant is encouraged to file a new statement with the IDS’s and the IDS’s will be reconsidered. If a new statement is not made, a fee is required under 37 CFR 1.17(v).
Interview Summary Comments
The interview summary on pg. 8 of Applicant’s Arguments/Remarks dated 10/23/2025 (hereinafter ‘Remarks’) provided all the answers that were given in the interview to the agenda. Examiner does note that the Applicant did want to follow the agenda questions only. No additional discussion was given which may have been beneficial to resolving some of the disagreements which appear to pertain to the scope of the claims.
Response to Arguments - 35 USC § 101
On pgs. 8-12 of the Remarks, Applicant argues the claims are patent eligible over 35 U.S.C. 101.
Prior to addressing any of the specific arguments, Examiner finds it important to clarify interpretation of the claims and the terms used.
Claims 1 recites “a type reservoir model for the reservoir”. This is not a known term in the art. The specification provides the “type reservoir model” uses “history machining” and “history matching” is a known term in the art. Dependent claim 10 also recites “type reservoir model information is obtained based on history matching”. Dependent claims must narrow the scope of the independent and it is reasonable to interpret the “type reservoir model” using “history matching”.
[0044] of the specification as published recites “In some implementations, the type reservoir model information may be obtained based on history matching and/or other information. For example, the type reservoir model may be developed using history matching. In some implementations, multiple reservoir models may be constructed, and a reservoir model that best matches the desired properties of the reservoir may be selected for use as the type reservoir model. In some implementations, model parameters (e.g., reservoir/fracture parameters) of the type reservoir model may be varied until the type reservoir model matches the desired properties of the reservoir (e.g., production simulated by the type reservoir model matching historical production data).”
Examiner agrees with [0044]. It describes the method as “varying until the desired properties”, which is interpreted as observing the model while adjusting input values.
However, in reviewing the arguments, there seems to be disagreement on how one of ordinary skill in the art would interpret the “type reservoir model” and the “history matching”.
The definition provided by the SLB Energy Glossary is as follows:
history matching
1. n. [Reservoir Characterization] “The act of adjusting a model of a reservoir until it closely reproduces the past behavior of a reservoir. The historical production and pressures are matched as closely as possible. The accuracy of the history matching depends on the quality of the reservoir model and the quality and quantity of pressure and production data. Once a model has been history matched, it can be used to simulate future reservoir behavior with a higher degree of confidence, particularly if the adjustments are constrained by known geological properties in the reservoir.”
Lemouzy et al., “Successful History Matching of Chaunoy Field Reservoir Behavior Using Geostatistical Modeling” [1995] demonstrates the basic concept of history matching in Fig. 12 using watercut data.
PNG
media_image1.png
500
788
media_image1.png
Greyscale
Fetkovich, “Decline Curve Analysis Using Type Curves” [1980] demonstrates curve matching and uses following bottomhole pressure in Fig. 8.
PNG
media_image2.png
452
582
media_image2.png
Greyscale
The instant application demonstrates the output in Fig. 5A and described in [0064] “FIG. 5A illustrates an example inflow performance relationship curve 500 for an oil reservoir. The inflow performance relationship curve 500 for the oil reservoir may be generated based on the production prediction (e.g., production rate, average pressure, absolute open flow) for a time point output by the trained machine learning model. The average pressure (average reservoir pressure) for the time point may form the y-intercept of the inflow performance relationship curve 500. The absolute open flow for the time point may form the x-intercept of the inflow performance relationship curve 500. The productivity index may be calculated based on the production rate, and the slope of the inflow performance relationship curve 500 at the bubble point pressure (e.g., known, assumed and/or defined by the user) may be equal to inverse of the productivity index (1/PI). The known/assumed shape or functional form of the inflow performance relationship curve may be used to generate the inflow performance relationship curve 500 between the y-intercept and the x-intercept, with the inflow performance relationship curve 500 passing through the bubble point pressure with the slope equal to inverse of the productivity index. Separate inflow performance relationship curves may be generated for separate time points to generate the transient inflow performance relationship.”
PNG
media_image3.png
384
550
media_image3.png
Greyscale
On pg. 9 of the Remarks, Applicant argues no clear identification of a single abstract idea in the claims. Examiner respectfully disagrees with this statement. All limitations have been addressed. The claim is evaluating the observed data to determine how the data should be adjusted. When performing curve matching, the data is observed to understand the previous performance and comparing it to the output. This is repeated multiple times. The training process would be similar to someone observing multiple datasets and understanding a pattern.
On pg. 10 of the Remarks, Applicant states “the Office Action merely states such simulations ‘can reasonably be done by observing an input and evaluating how other values adjust based on the input value.’ Page 4. This is factually incorrect. Production simulations for petrochemical (oil and gas) reservoirs are sophisticated computational processes that use numerical models to predict fluid flow through porous underground rock formations.” Examiner does now see how any of this information has been “factually incorrect”. The input described in dependent claim 7 is “production rate”. The input described in claim 8 is “fracture geometry”, “flowing bottom hole pressure”, and “number of fracture cluster”. These are in the alternative. Even if all of these are in combination, none of these are sophisticated computationally. A simple display of “bottom hole pressure” is found in Fig. 5A. A person of ordinary skill in the art would recognize all of these to be common inputs which can be easily plotted as shown in Applicant’s own figure. The output of the system in claim 9 is the “average pressure”, “production rate”, and “absolute open flow” at variable times. When producing an oil well, “production rate” of the well is a common output. This can again be seen in Fig. 5A of Applicant’s drawings, the “oil rate” could be interpreted as the claimed “production rate” of claim 9. Relationship between these elements is within the ability of the person of ordinary skill in the art to establish in the mind. For instance, when observing the flowing bottomhole pressure of previous data, understanding the predicted well will most likely following a same pattern.
On pg. 11 of the Remarks, the training data generation is completed by performing an evaluation. The trained machine learning model is performed based on the observation. Formatting could include arranging the data, which as these are datasets affecting production rate and these are datasets affecting average pressure. Once these datasets are grouped, a person of ordinary skill in the art could look at these different inputs to determine patterns.
Continuing on pg. 11 of the Remarks, the trained machine learning model provides a prediction, which amounts to a judgement. Applicant states “Training a machine-learning model is not a human judgment, it is an automated process in which huge amounts of data are processed to result in a model that provides a specific function.” Examiner respectfully disagrees with the automated process requires a large amount of data. The claim does not recite the amount of data required and Figs. 5A-6C do not show a large amount of data being processed. The rigidity in the plot in 6A would imply only a few dozen evaluated points. The input in the inputs selected for 6A are the fracture half length and No. of clusters. Given a handful of data sets from a field that contains well data with fracture half length and No. of clusters, it would be reasonably for a person of ordinary skill in the art to predict the output of Fig. 6A.
Pg. 12 of the Remarks ends the analysis at Step 2A, Prong One.
Applicant's arguments have been fully considered but they are not persuasive. Rejection under 35 U.S.C. 101 is maintained.
Response to Arguments - 35 USC § 103
On pg. 12 of the Remarks, Applicant argues Claims 1 and 11 patentable over the current prior are rejection under 35 U.S.C. 103 over Sankaran et al., United States Patent 11,514,216 B1 (hereinafter ‘Sankaran’) in view of Jiang et al., “Data-space inversion using a recurrent autoencoder for time-series parameterization” [published 18 Nov 2020] (hereinafter ‘Jiang’).
On pg. 12 of the Remarks, Applicant argues the limitation “wherein the trained machine learning model provides prediction of transient production in the reservoir and facilitates determination of transient inflow performance relationship for the reservoir” is not taught by Sankaran as found on pg. 15 of the Non-Final Rejection.
Regarding Sankaran, Applicant argues “The Office Action alleges Sankaran teaches these features at FIGS. 8 and 9A, and column 17, lines 19-35. This is factually erroneous. While these portions of Sankaran appear to describe calculations of transient production, including calculations that facilitate determination of transient inflow performance relationships, the approach described in Sankaran is physics-based, and does not use the technology of machine learning for these calculations”. Examiner respectfully disagrees with the Sankaran does not use machine learning. On pg. 15 of the Non-Final, the limitation “generate training data for a machine learning model” is cited to Col 24 lines 40-47 Sankaran “…The field-wide results are shown in FIG. 14 for oil forecasting. Similar results were obtained for the gas and water phase forecast. In this example, each well was forecasted with various training data windows up to a full year (e.g., 30, 60, 90, days, etc.) using the remaining "known" data as a blind hold-out sample. In this example, only wells with large histories were included, to guarantee a representative validation dataset was available...”. The machine learning model is the same “machine learning model” in both limitations. Further, the limitation “train the machine learning model using the training data” is mapped explicitly reciting “machine learning”. Col 23 lines 28-33 Sankaran "...In certain example embodiments, control unit 400 is configured to use information from database 408 to train one or many machine learning algorithms 412, including, but not limited to, artificial neural network, random forest, gradient boosting, support vector machine, or kernel density estimator..."
Further on pg. 13 of the Remarks, Applicant presents arguments regarding teaching away. “Sankaran expressly teaches away from using a trained machine learning model to perform these types of reservoir modeling for a variety of reasons at column 2, lines 37-52. The cited portions of Jiang, even if it would have been obvious to combine Jiang with Sankaran in view of the teaching away by Sankaran from such a combination, do not address these deficiencies of Sankaran with respect to the claim features reproduced above.”
column 2, lines 37-52 Sankaran
“Data-driven models such as statistical learning techniques and machine learning have been implemented as alternatives to the traditional domain-specific physics-based reservoir models. Given sufficient training data, these models can capture complex relationships and predict hydrocarbon production from a variety of data inputs, including rock-fluid properties, completion design parameters and operational factors. These models can be very powerful for various 45 applications in descriptive, predictive, and prescriptive analytics, but have their own inherent drawbacks. First, sufficient training data is required to properly capture meaningful correlations, both in quantity (large number of wells) and quality (representing variability in the input ranges). Data-driven models may only be valid within the training data features and ranges, so they tend to be very specific to a given field/basin/sector and not fully generalizable.”
MPEP 1504.03(III) "A prima facie case of obviousness can be rebutted if the applicant...can show that the art in any material respect ‘taught away’ from the claimed invention...A reference may be said to teach away when a person of ordinary skill, upon reading the reference...would be led in a direction divergent from the path that was taken by the applicant." See In re Haruna, 249 F.3d 1327, 58 USPQ2d 1517 (Fed. Cir. 2001).
Col 1 and col 2 of Sankaran cover the background of the invention. The background discusses Decline curve analysis ("DCA") (Col 1 lines 33-57), Analytical methods such as rate transient analysis ("RTA") (Col 1 line 58-Col 2 line 9), numerical models (Col 2 lines 14-36), and statistical learning techniques/machine learning (Col 2 lines 37-62). The background has surveyed several different techniques and recognized potential limitations. DCA and transient analysis (referred to as transient well performance) are both used in Sankaran throughout the invention. The background is not a divergent path, but finding a way to incorporate these models effectively. Col 2 lines 63-67 “It is desirable to have a robust and scalable method for quantifying well productivity, which can be applied in a practical manner to all wells, overcoming the limitations of decline curve analysis and conventional analytical and numerical models”. It is also noted Col 23 lines 28-33 Sankaran teaches an embodiment explicitly using machine learning cited above.
No additional arguments presented for claims 2-10 and 12-21.
Applicant's arguments have been fully considered but they are not persuasive. Rejection under 35 U.S.C. 103 is maintained.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claim 1 (Statutory Category – System)
Step 2A – Prong 1: Judicial Exception Recited?
Yes, the claim recites a mental process, specifically:
MPEP 2106.04(a)(2)(Ill) “Accordingly, the "mental processes" abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, Judgments, and opinions.”
Further, the MPEP recites “The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid (e.g., pen and paper or a slide rule) to perform the claim limitation.”
2106.04(a)(2)(I)(A) “Mathematical Relationships A mathematical relationship is a relationship between variables or numbers. A mathematical relationship may be expressed in words or using mathematical symbols. For example, pressure (p) can be described as the ratio between the magnitude of the normal force (F) and area of the surface on contact (A), or it can be set forth in the form of an equation such as p = F/A.”
2106.04(a)(2)(I)(B) “Mathematical Formulas or Equations A claim that recites a numerical formula or equation will be considered as falling within the "mathematical concepts" grouping. In addition, there are instances where a formula or equation is written in text format that should also be considered as falling within this grouping. For example, the phrase "determining a ratio of A to B" is merely using a textual replacement for the particular equation (ratio = A/B). Additionally, the phrase "calculating the force of the object by multiplying its mass by its acceleration" is using a textual replacement for the particular equation (F= ma).”
2106.04(a)(2)(I)(C) “Mathematical Calculations A claim that recites a mathematical calculation, when the claim is given its broadest reasonable interpretation in light of the specification, will be considered as falling within the "mathematical concepts" grouping. A mathematical calculation is a mathematical operation (such as multiplication) or an act of calculating using mathematical methods to determine a variable or number, e.g., performing an arithmetic operation such as exponentiation. There is no particular word or set of words that indicates a claim recites a mathematical calculation. That is, a claim does not have to recite the word "calculating" in order to be considered a mathematical calculation. For example, a step of "determining" a variable or number using mathematical methods or "performing" a mathematical operation may also be considered mathematical calculations when the broadest reasonable interpretation of the claim in light of the specification encompasses a mathematical calculation.”
generate multiple production simulations for the reservoir based on different values of input parameters for the type reservoir model;
The multiple production simulations can reasonably be done by observing an input and evaluating how other values adjust based on the input value.
generate training data for a machine learning model based on the multiple production simulations for the reservoir;
The previous evaluated simulated data can be formatted and fed into a machine learning model.
train the machine learning model using the training data, wherein the trained machine learning model provides prediction of transient production in the reservoir and facilitates determination of transient inflow performance relationship for the reservoir; and
The trained machine learning model observes the inflow and evaluates a relationship between the inflow and the production to create a prediction, which can be interpreted as a judgment. The relationship between the variables can also be interpreted as a mathematical concept.
Therefore, the claim recites a mental process and mathematical concepts.
Step 2A – Prong 2: Integrated into a Practical Solution?
MPEP 2106.05(g) Insignificant Extra-Solution Activity has found mere data gathering and post solution activity to be insignificant extra-solution activity.
The following step is merely gathering the data on elements to be used in the calculation:
obtain type reservoir model information, the type reservoir model information defining a type reservoir model for the reservoir;
MPEP 2106.05(f) Mere Instructions To Apply An Exception has found simply adding a general purpose computer or computer components after the fact to an abstract idea (e.g., a fundamental economic practice or mathematical equation) does not integrate a judicial exception into a practical application or provide significantly more.
one or more physical processors configured by machine-readable instructions to:
store the trained machine learning model in a non-transitory storage medium.
The additional elements have been considered both individually and as an ordered combination in to determine whether they integrate the exception into a practical application.
Therefore, no meaningful limits are imposed on practicing the abstract idea.
The claim is directed to the abstract idea.
Step 2B: Claim provides an Inventive Concept?
No, as discussed with respect to Step 2A, the additional limitation is mere data gathering/post solution activity (Insignificant Extra-Solution Activity) and a general purpose computer do not impose any meaningful limits on practicing the abstract idea and therefore the claim does not provide an inventive concept in Step 2B.
Further, in regards to step 2B and as cited above in step 2A, MPEP 2106.05(g) “Obtaining information about transactions using the Internet to verify credit card transactions, CyberSource v. Retail Decisions, Inc., 654 F.3d 1366, 1375, 99 USPQ2d 1690, 1694 (Fed. Cir.2011)” is merely data gathering.
The additional elements have been considered both individually and as an ordered combination in the significantly more consideration.
The claim is ineligible.
2. “The system of claim 1, wherein the one or more physical processors are further configured by the machine-readable instructions to: obtain production parameter scenario information, the production parameter scenario information defining a scenario of production parameter for the reservoir; determine transient production prediction based on inputting the scenario of production parameter for the reservoir into the trained machine learning model; and determine the transient inflow performance relationship for the reservoir based on the transient production prediction” obtaining is mere data gathering which provides the observed input for the determining elements which amount to evaluation.
3. “The system of claim 2, wherein the scenario of production parameter for the reservoir includes a time series of flowing bottom hole pressure” is mere data gathering.
4. “The system of claim 2, wherein the determination of the transient inflow performance relationship for the reservoir includes determination of separate inflow performance relationships for separate times” is further evaluation based on an additional parameter.
5. “The system of claim 4, wherein the determination of the transient inflow performance relationship for the reservoir includes determination of transient oil inflow performance relationship or transient gas condensate inflow performance relationship” is further evaluation based on an additional parameter.
6. “The system of claim 1, wherein the machine learning model includes a recurrent neural network” is specifying the method used for evaluation.
7. “The system of claim 1, wherein the training data for the machine learning model includes pairings of corresponding values of the input parameters and values of production rate” is mere data gathering of inputs.
8. “The system of claim 1, wherein the input parameters include one or more of fracture geometry, flowing bottom hole pressure, and number of fracture clusters” is mere data gathering of inputs.
9. “The system of claim 1, wherein output of the trained machine learning model includes multiple sets of average pressure, production rate, and/or absolute open flow of the reservoir at different times, wherein the multiple sets of the average pressure, the production rate, and/or the absolute open flow of the reservoir at different times includes a first set of the average pressure, the production rate, and/or the absolute open flow of the reservoir at a first time and a second set of the average pressure, the production rate, and/or the absolute open flow of the reservoir at a second time” is further specifying which parameters to consider when performing the evaluation.
10. “The system of claim 1, wherein the type reservoir model information is obtained based on history matching” is mere data gathering.
Claims 11-20 are method claims, containing substantially the same elements as system Claims 1-10, respectively, and are rejected on the same grounds under 35 U.S.C. 101 as Claims 1-10, respectively, Mutatis mutandis.
Claim 21 (Statutory Category – method)
Step 2A – Prong 1: Judicial Exception Recited?
Yes, the claim recites a mental process, specifically:
MPEP 2106.04(a)(2)(Ill) “Accordingly, the "mental processes" abstract idea grouping is defined as concepts performed in the human mind, and examples of mental processes include observations, evaluations, Judgments, and opinions.”
Further, the MPEP recites “The courts do not distinguish between mental processes that are performed entirely in the human mind and mental processes that require a human to use a physical aid (e.g., pen and paper or a slide rule) to perform the claim limitation.”
2106.04(a)(2)(I)(A) “Mathematical Relationships A mathematical relationship is a relationship between variables or numbers. A mathematical relationship may be expressed in words or using mathematical symbols. For example, pressure (p) can be described as the ratio between the magnitude of the normal force (F) and area of the surface on contact (A), or it can be set forth in the form of an equation such as p = F/A.”
2106.04(a)(2)(I)(B) “Mathematical Formulas or Equations A claim that recites a numerical formula or equation will be considered as falling within the "mathematical concepts" grouping. In addition, there are instances where a formula or equation is written in text format that should also be considered as falling within this grouping. For example, the phrase "determining a ratio of A to B" is merely using a textual replacement for the particular equation (ratio = A/B). Additionally, the phrase "calculating the force of the object by multiplying its mass by its acceleration" is using a textual replacement for the particular equation (F= ma).”
2106.04(a)(2)(I)(C) “Mathematical Calculations A claim that recites a mathematical calculation, when the claim is given its broadest reasonable interpretation in light of the specification, will be considered as falling within the "mathematical concepts" grouping. A mathematical calculation is a mathematical operation (such as multiplication) or an act of calculating using mathematical methods to determine a variable or number, e.g., performing an arithmetic operation such as exponentiation. There is no particular word or set of words that indicates a claim recites a mathematical calculation. That is, a claim does not have to recite the word "calculating" in order to be considered a mathematical calculation. For example, a step of "determining" a variable or number using mathematical methods or "performing" a mathematical operation may also be considered mathematical calculations when the broadest reasonable interpretation of the claim in light of the specification encompasses a mathematical calculation.”
obtaining a trained machine learning model, the trained machine learning model providing prediction of transient production in the reservoir and facilitating determination of transient inflow performance relationship for the reservoir,
The trained machine learn model is based on an observation of the of the production and the performance which is evaluated to determine the relationship. The relationship between the variables can also be interpreted as a mathematical concept.
wherein training data for the trained machine learning model is generated based on multiple production simulations for the reservoir, the multiple production simulations for the reservoir generated based on different values of input parameters for a type reservoir model for the reservoir;
The multiple production simulations can reasonably be done by observing an input and evaluating how other values adjust based on the input value.
determining transient production prediction based on inputting the scenario of production parameter for the reservoir into the trained machine learning model; and
Inputting the scenario parameters is observation and evaluation of those parameters.
determining the transient inflow performance relationship for the reservoir based on the transient production prediction.
The trained machine learning model observes the inflow and evaluates a relationship between the inflow and the production to create a prediction, which can be interpreted as a judgment. The relationship between the variables can also be interpreted as a mathematical concept.
Therefore, the claim recites a mental process and mathematical concepts.
Step 2A – Prong 2: Integrated into a Practical Solution?
MPEP 2106.05(g) Insignificant Extra-Solution Activity has found mere data gathering and post solution activity to be insignificant extra-solution activity.
The following step is merely gathering the data on elements to be used in the calculation:
obtaining production parameter scenario information, the production parameter scenario information defining a scenario of production parameter for the reservoir;
The additional elements have been considered both individually and as an ordered combination in to determine whether they integrate the exception into a practical application.
Therefore, no meaningful limits are imposed on practicing the abstract idea.
The claim is directed to the abstract idea.
Step 2B: Claim provides an Inventive Concept?
No, as discussed with respect to Step 2A, the additional limitation is mere data gathering/post solution activity (Insignificant Extra-Solution Activity) and a general purpose computer do not impose any meaningful limits on practicing the abstract idea and therefore the claim does not provide an inventive concept in Step 2B.
Further, in regards to step 2B and as cited above in step 2A, MPEP 2106.05(g) “Obtaining information about transactions using the Internet to verify credit card transactions, CyberSource v. Retail Decisions, Inc., 654 F.3d 1366, 1375, 99 USPQ2d 1690, 1694 (Fed. Cir.2011)” is merely data gathering.
The additional elements have been considered both individually and as an ordered combination in the significantly more consideration.
The claim is ineligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-21 are rejected under 35 U.S.C. 103 as being unpatentable over
Sankaran et al., United States Patent 11,514,216 B1 (hereinafter ‘Sankaran’) in view of
Jiang et al., “Data-space inversion using a recurrent autoencoder for time-series parameterization” [published 18 Nov 2020] (hereinafter ‘Jiang’).
Regarding Claim 1: A system for determining inflow performance relationship for a reservoir, the system comprising:
Sankaran teaches one or more physical processors configured by machine-readable instructions to: (Col 23 lines 35-41 Sankaran “…Processor 404 may include, for example, a microprocessor, microcontroller, digital signal processor (DSP), application specific integrated circuit (ASIC), or any other digital or analog circuitry configured to interpret and/or execute program instructions and/or process data. In some embodiments, processor 404 may be communicatively coupled to memory 406…”)
Sankaran teaches obtain type reservoir model information, the type reservoir model information defining a type reservoir model for the reservoir; (Col 8 lines 1-15 Sankaran teaches fluid properties, i.e. reservoir model information for the EOS model, i.e. a type reservoir model “…In certain embodiments, one or more fluid properties ( e.g., PVT properties) may be determined for a specific well or at least a portion of a reservoir. In other embodiments, one or more fluid properties may be determined for two or more wells, or across an entire reservoir. In such embodiments, for example, a field-wide equation of state (EOS) model may be used to provide fluid characterization of one or more fluid properties ( e.g., the phase and volume behavior of a fluid system). The EOS model may then be used to estimate the fluid properties at any well location based on its corresponding fluid composition, reservoir pressure, and/or temperature. In some embodiments, fluid composition may be known from samples taken across the reservoir and/or wells…”)
Sankaran teaches generate training data for a machine learning model … (Col 24 lines 40-47 Sankaran teaches the training data “…The field-wide results are shown in FIG. 14 for oil forecasting. Similar results were obtained for the gas and water phase forecast. In this example, each well was forecasted with various training data windows up to a full year (e.g., 30, 60, 90, days, etc.) using the remaining "known" data as a blind hold-out sample. In this example, only wells with large histories were included, to guarantee a representative validation dataset was available…”)
Sankaran teaches train the machine learning model using the training data, (Col 23 lines 28-33 Sankaran teaches to train “…In certain example embodiments, control unit 400 is configured to use information from database 408 to train one or many machine learning algorithms 412, including, but not limited to, artificial neural network, random forest, gradient boosting, support vector machine, or kernel density estimator…”)
Sankaran teaches wherein the trained machine learning model provides prediction of transient production in the reservoir and facilitates determination of transient inflow performance relationship for the reservoir; and (Fig. 9A and Col 17 lines 19-35 Sankaran teaches the inflow performance relationship and the productivity index with the IPR found on Fig. 9A based on days, i.e. transient “…In some embodiments, inflow performance relationship (IPR) curves represent well deliverability at a given reservoir condition, as they display expected liquid production at various pressure drawdowns. FIG. 8 shows a PI inflow performance relationship curve for a PI-based forecasting method. When combined with a vertical lift performance (VLP) curve, representing the well intake, nodal analysis may be performed to determine the well operating point. In certain embodiments, this analysis provides the basis for production optimization, informing artificial lift timing and operational decisions. IPR curves have not traditionally been used in unconventionals exhibiting prolonged transient flow, since they are only strictly valid for boundary-dominated systems in pseudo-steady state. Example approaches may define IPR curves dynamically using the calculated average reservoir pressure and liquid productivity index…”)
PNG
media_image4.png
392
680
media_image4.png
Greyscale
Sankaran teaches store the trained machine learning model in a non-transitory storage medium. (Col 23 lines 44-53 Sankaran teaches using a non-tranistory media “…Program instructions or data may constitute portions of software for carrying out fluid system modeling, as described herein. Memory 406 may include any system, device, or apparatus configured to hold and/or house one or more memory modules; for example, memory 406 may include read-only memory, random access memory, solid state memory, or disk-based memory. Each memory module may include any system, device or apparatus configured to retain program instructions and/or data for a period of time (e.g., computer-readable non-transitory media) …”)
Sankaran does not appear to explicitly disclose
generate multiple production simulations for the reservoir based on different values of input parameters for the type reservoir model;
…model based on the multiple production simulations for the reservoir;
However, Jiang teaches generate multiple production simulations for the reservoir based on different values of input parameters for the type reservoir model; (Pg. 419 left col 2nd paragraph Jiang teaches 800 realizations, i.e. multiple simulations with varying permeability, i.e. different values of input parameters “…The 2D synthetic bimodal channelized system considered here is the same as the system treated in earlier DSI studies [12, 34]. The channelized realizations are defined on a 60 × 60 grid. The size of each grid block is 25 m × 25 m × 10 m. We use Nr = 800 prior realizations. Figure 4 displays four prior realizations of log-permeability. Porosity is constant at 0.2 in all realizations. The channelized model includes sand and mud facies, with permeability variation in each facies—thus, the log-permeability is bimodal…”)
Jiang teaches …model based on the multiple production simulations for the reservoir; (Pg. 414 right col 3rd paragraph Jiang teaches a model from assimilating data multiple times, i.e. multiple productions “…ESMDA, developed by Emerick and Reynolds [5], is a popular approach for model-based history matching. ESMDA generates posterior samples by assimilating data multiple times over an ensemble with inflated error covariance…”)
Sankaran and Jiang are analogous art because they are from the same field of endeavor, well production characterization.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the generate training data for a machine learning model as disclosed by Sankaran by generate multiple production simulations for the reservoir based on different values of input parameters for the type reservoir model and model based on the multiple production simulations for the reservoir as disclosed by Jiang.
One of ordinary skill in the art would have been motivated to make this modification in order to provide model calibration as discussed in on pg. 413 by Jiang “…Model calibration is typically required before subsurface flow simulation tools can be used for prediction and optimization. This is usually accomplished through application of formal inversion procedures in which key flow parameters, such as porosity and permeability in all grid blocks in the model, are determined based on observations and a prior concept of the geological scenario. With large and geologically complex models, however, this inversion can be challenging. This is because many flow simulations may be required, which can be expensive for highly refined models, and because it can be difficult to retain geological realism in the posterior (calibrated) models…”
Regarding Claim 2: Sankaran and Jiang teach The system of claim 1, wherein the one or more physical processors are further configured by the machine-readable instructions to:
Sankaran teaches obtain production parameter scenario information, the production parameter scenario information defining a scenario of production parameter for the reservoir; (Col 9 lines 30-39 Sankaran teaches use of production data based on surface measurements, i.e. production parameter scenario information, to determine the flowing conditions and well performance, i.e. scenario of production “…In step 130, production data of a well and/or reservoir may be calculated from surface measurements (e.g., measured well data) in order to capture transient downhole flowing conditions and well performance. In some embodiments, transient well performance is characterized in unconventional wells using well production rates, PVT data or correlations, and flowing bottomhole pressures obtained as described in the prior steps…”)
Sankaran teaches
PNG
media_image5.png
500
729
media_image5.png
Greyscale
determine transient production prediction based on inputting the scenario of production parameter for the reservoir into the trained machine learning model; and (Fig. 14 and Col 24 lines 50-57 Sankaran teaches additional training data included to determine a production forecast, i.e. production prediction “…However, as shown in FIG. 14, the PIBF method performed better relative to DCA especially at early time, when production declines are yet not very well defined, achieving a median well-level error reduction from -65% to -45%. As more training data is included, the performance of both methods starts converging on similar outputs, as enough data is available to define a clear production decline trend…”)
Sankaran teaches determine the transient inflow performance relationship for the reservoir based on the transient production prediction. (Col 7 lines 29-38 Sankaran teaches the transient well performance when determining the productivity index and estimating the IPR, inflow performance relationship “…In certain embodiments, step 130 may be optional. In step 140, a productivity index is determined based, at least in part, on the transient well performance for the one or more wells. productivity index (PI). In step 150, PI is used as a base variable to perform production forecasting. In certain embodiments, production forecasting 150 is performed using PI and expected future operating conditions. In certain 35 embodiments, production forecasting 150 may be used to estimate the inflow performance relationship (IPR) curves for one or more wells to capture well deliverability…”)
Regarding Claim 3: Sankaran and Jiang teach The system of claim 2,
Sankaran teaches wherein the scenario of production parameter for the reservoir includes a time series of flowing bottom hole pressure. (Col 9 lines 23-30 Sankaran teaches determining the bottom hold pressure over time, i.e. time series “…In certain embodiments, the flow model 200 is used to calculate bottomhole pressure over time. In certain embodiments, the flow model 200 is configured to calculate BHP for all periods in the life of the well. In some embodiments, BHP may be continuously calculated in real-time to provide automatic production surveillance and monitoring…”)
Regarding Claim 4: Sankaran and Jiang teach The system of claim 2,
Sankaran teaches wherein the determination of the transient inflow performance relationship for the reservoir includes determination of separate inflow performance relationships for separate times. (Col 10 lines 53-59 Sankaran teaches the separate inflow performance relationships of oil, water, and gas “…qte is total equivalent rate, qo is oil rate, qw is water rate, qg is gas rate, Rs is solution gas-oil ratio at a pressure of interest, BO is oil formation volume factor at a pressure of interest, Bw is water formation volume factor at a pressure of interest, and Bg is gas formation volume factor at a pressure of interest…”)
Regarding Claim 5: Sankaran and Jiang teach The system of claim 4,
Sankaran teaches wherein the determination of the transient inflow performance relationship for the reservoir includes determination of transient oil inflow performance relationship or transient gas condensate inflow performance relationship. (Col 17 line 67 – Col 18 lines 1-5 Sankaran teaches using gas condensates “…As a starting point, either for dry-gas or gas-condensate cases, appropriate PVT correlations can be used, for example in the case of gas condensate where the specific gravity is estimated from both gas and liquid specific gravities, considering the separator conditions…”)
Regarding Claim 6: Sankaran and Jiang teach The system of claim 1,
Jiang teaches wherein the machine learning model includes a recurrent neural network. (Pg. 416 left col 1st paragraph Jiang teaches recurrent neural network “…Here, we address this problem using a recurrent autoencoder, which is an autoencoder based on recurrent neural networks (RNNs)…”
Further, Pg. 416 left col 2nd paragraph Jiang teaches “…Long short-term memory [10] is an efficient method within the class of RNNs…”)
Regarding Claim 7: Sankaran and Jiang teach The system of claim 1,
Sankaran teaches wherein the training data for the machine learning model includes pairings of corresponding values of the input parameters and values of production rate. (Col 4 lines 38-42 Sankaran teaches a well production rate with an associated flowing pressure, i.e. input parameter where the production and pressure form a pair “…determining a transient Productivity Index (PI) for the unconventional reservoir, based, at least in part, on the one or more fluid properties and measured well data for the one or more wells, wherein the measured well data includes a well production rate and a well flowing pressure…”)
Regarding Claim 8: Sankaran and Jiang teach The system of claim 1,
Sankaran teaches wherein the input parameters include one or more of fracture geometry, flowing bottom hole pressure, and number of fracture clusters. (Col 10 lines 40-44 Sankaran teaches flowing bottom hole pressure “…Q is cumulative production volume, q is instantaneous production rate, P; is initial reservoir pressure, and pwf is flowing bottomhole pressure…”)
Regarding Claim 9: Sankaran and Jiang teach The system of claim 1,
Jiang teaches wherein output of the trained machine learning model includes multiple sets of average pressure, production rate, and/or absolute open flow of the reservoir at different times, (Pg. 419 right col last paragraph Jiang teaches a production rate at different time series, i.e. different times “…The simulation results include the water injection rate (WIR) for injectors I1 and I2, and the water production rate (WPR) and the oil production rate (OPR) for producers P1, P2, and P3. Thus, we have eight quantities (time series) of interest; i.e., NQoI = 8…)
Jiang teaches wherein the multiple sets of the average pressure, the production rate, and/or the absolute open flow of the reservoir at different times includes a first set of the average pressure, the production rate, and/or the absolute open flow of the reservoir at a first time and a second set of the average pressure, the production rate, and/or the absolute open flow of the reservoir at a second time. (Fig. 3 and pg. 418 right col last paragraph Jiang teaches a production rate with two times shown, 210 and 450 days for the first and second time“…The lower row in Fig. 3 shows cross-plots for water production rates in another well (P3) at two different times—210 and 450 days. Because water rate again increases monotonically in time for this well, all points in the reference prior simulation results (left) would be expected to fall above the dashed 45◦ line, and this is indeed observed…”)
PNG
media_image6.png
346
664
media_image6.png
Greyscale
Further, col 7 lines 38-49 Sankaran teaches determining the average pressure based on the performance “…This workflow 100 and embodiments thereof are discussed further below. In some embodiments, the method of the present disclosure may be used to determine one or more indicators or metrics ( e.g., PI, drainage volume, instantaneous recovery ratio, average reservoir pressure), which may be used to normalize and compare well performance. Additionally, in certain embodiments, average reservoir pressure and PI may be determined and used to represent well deliverability (as a dynamic IPR curve) at various field conditions (past and future) that can be used for drawdown management, artificial lift planning and optimization…”)
Regarding Claim 10: Sankaran and Jiang teach The system of claim 1,
Jiang teaches wherein the type reservoir model information is obtained based on history matching. (Pg. 419 left col 1st paragraph Jiang teaches history matching “…In this section, we present history matching results using the DSI framework for a 2D channelized system and a 3D Gaussian model. The posterior results from DSI, using a range of treatments, are compared to reference rejection sampling (RS) results. We assess the quality of the various DSI results, including correlations, for several quantities of interest…”)
Claims 11-20 are method claims, containing substantially the same elements as system claims 1-10, respectively, and are rejected on the same grounds under 35 U.S.C. 103 as Claims 1-10, respectively, Mutatis mutandis.
Regarding Claim 21: A method for determining inflow performance relationship for a reservoir, the method comprising:
Sankaran teaches obtaining production parameter scenario information, the production parameter scenario information defining a scenario of production parameter for the reservoir; (Col 8 lines 1-15 Sankaran teaches fluid properties, i.e. reservoir model information for the EOS model, i.e. a type reservoir model “…In certain embodiments, one or more fluid properties ( e.g., PVT properties) may be determined for a specific well or at least a portion of a reservoir. In other embodiments, one or more fluid properties may be determined for two or more wells, or across an entire reservoir. In such embodiments, for example, a field-wide equation of state (EOS) model may be used to provide fluid characterization of one or more fluid properties ( e.g., the phase and volume behavior of a fluid system). The EOS model may then be used to estimate the fluid properties at any well location based on its corresponding fluid composition, reservoir pressure, and/or temperature. In some embodiments, fluid composition may be known from samples taken across the reservoir and/or wells…”
Further, Col 9 lines 30-39 Sankaran teaches use of production data based on surface measurements, i.e. production parameter scenario information, to determine the flowing conditions and well performance, i.e. scenario of production “…In step 130, production data of a well and/or reservoir may be calculated from surface measurements (e.g., measured well data) in order to capture transient downhole flowing conditions and well performance. In some embodiments, transient well performance is characterized in unconventional wells using well production rates, PVT data or correlations, and flowing bottomhole pressures obtained as described in the prior steps…”)
Sankaran teaches obtaining a trained machine learning model, the trained machine learning model providing prediction of transient production in the reservoir and facilitating determination of transient inflow performance relationship for the reservoir, (Fig. 9A and Col 17 lines 19-35 Sankaran teaches the inflow performance relationship and the productivity index with the IPR found on Fig. 9A based on days, i.e. transient “…In some embodiments, inflow performance relationship (IPR) curves represent well deliverability at a given reservoir condition, as they display expected liquid production at various pressure drawdowns. FIG. 8 shows a PI inflow performance relationship curve for a PI-based forecasting method. When combined with a vertical lift performance (VLP) curve, representing the well intake, nodal analysis may be performed to determine the well operating point. In certain embodiments, this analysis provides the basis for production optimization, informing artificial lift timing and operational decisions. IPR curves have not traditionally been used in unconventionals exhibiting prolonged transient flow, since they are only strictly valid for boundary-dominated systems in pseudo-steady state. Example approaches may define IPR curves dynamically using the calculated average reservoir pressure and liquid productivity index…”)
PNG
media_image4.png
392
680
media_image4.png
Greyscale
PNG
media_image5.png
500
729
media_image5.png
Greyscale
Sankaran teaches determining transient production prediction based on inputting the scenario of production parameter for the reservoir into the trained machine learning model; and (Fig. 14 and Col 24 lines 50-57 Sankaran teaches additional training data included to determine a production forecast, i.e. production prediction “…However, as shown in FIG. 14, the PIBF method performed better relative to DCA especially at early time, when production declines are yet not very well defined, achieving a median well-level error reduction from -65% to -45%. As more training data is included, the performance of both methods starts converging on similar outputs, as enough data is available to define a clear production decline trend…”)
Sankaran teaches determining the transient inflow performance relationship for the reservoir based on the transient production prediction. (Col 7 lines 29-38 Sankaran teaches the transient well performance when determining the productivity index and estimating the IPR, inflow performance relationship “…In certain embodiments, step 130 may be optional. In step 140, a productivity index is determined based, at least in part, on the transient well performance for the one or more wells. productivity index (PI). In step 150, PI is used as a base variable to perform production forecasting. In certain embodiments, production forecasting 150 is performed using PI and expected future operating conditions. In certain 35 embodiments, production forecasting 150 may be used to estimate the inflow performance relationship (IPR) curves for one or more wells to capture well deliverability…”)
Sankaran does not appear to explicitly disclose
wherein training data for the trained machine learning model is generated based on multiple production simulations for the reservoir, the multiple production simulations for the reservoir generated based on different values of input parameters for a type reservoir model for the reservoir;
However, Jiang teaches wherein training data for the trained machine learning model is generated based on multiple production simulations for the reservoir, the multiple production simulations for the reservoir generated based on different values of input parameters for a type reservoir model for the reservoir; (Pg. 419 left col 2nd paragraph Jiang teaches 800 realizations, i.e. multiple simulations with varying permeability, i.e. different values of input parameters “…The 2D synthetic bimodal channelized system considered here is the same as the system treated in earlier DSI studies [12, 34]. The channelized realizations are defined on a 60 × 60 grid. The size of each grid block is 25 m × 25 m × 10 m. We use Nr = 800 prior realizations. Figure 4 displays four prior realizations of log-permeability. Porosity is constant at 0.2 in all realizations. The channelized model includes sand and mud facies, with permeability variation in each facies—thus, the log-permeability is bimodal…”)
Further, Pg. 414 right col 3rd paragraph Jiang teaches a model from assimilating data multiple times, i.e. multiple productions “…ESMDA, developed by Emerick and Reynolds [5], is a popular approach for model-based history matching. ESMDA generates posterior samples by assimilating data multiple times over an ensemble with inflated error covariance…”)
Sankaran and Jiang are analogous art because they are from the same field of endeavor, well production characterization.
It would have been obvious for one of ordinary skill in the art before the effective filing date of the claimed invention to have combined the determining transient production prediction based on inputting the scenario of production parameter for the reservoir into the trained machine learning model as disclosed by Sankaran by wherein training data for the trained machine learning model is generated based on multiple production simulations for the reservoir, the multiple production simulations for the reservoir generated based on different values of input parameters for a type reservoir model for the reservoir as disclosed by Jiang.
One of ordinary skill in the art would have been motivated to make this modification in order to provide model calibration as discussed in on pg. 413 by Jiang “…Model calibration is typically required before subsurface flow simulation tools can be used for prediction and optimization. This is usually accomplished through application of formal inversion procedures in which key flow parameters, such as porosity and permeability in all grid blocks in the model, are determined based on observations and a prior concept of the geological scenario. With large and geologically complex models, however, this inversion can be challenging. This is because many flow simulations may be required, which can be expensive for highly refined models, and because it can be difficult to retain geological realism in the posterior (calibrated) models…”
Conclusion
Claims 1-21 are rejected.
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOHN E JOHANSEN whose telephone number is (571)272-8062. The examiner can normally be reached M-F 9AM-3PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Emerson Puente can be reached at 5712723652. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/JOHN E JOHANSEN/Examiner, Art Unit 2187