DETAILED ACTION
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 09/26/2025 has been entered. Claims 1, 7, 9, and 13 have been amended. Claims 2, 8, and 14 are cancelled. No new claims have been added. Claims 1, 3-7, 9-13, and 15 are currently pending.
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 .
Response to Amendment
Regarding the amendments to the claims, the newly-added matter has been evaluated against the disclosure. Sufficient support is contained within the specification. Examiner agrees that no new matter has been added to the claims.
Response to Arguments
35 U.S.C. § 101
Regarding the rejection under 35 U.S.C. § 101, applicant has amended the claims to incorporate additional elements. Applicant asserts that the amended claims are not directed to mental process or mathematical evaluation.
Applicant argues that training a physics constrained machine learning model is a patent-eligible limitation that does not recite mathematical concepts. After further consideration, the examiner agrees that the limitation does not explicitly recite a mathematical concept and accordingly the treatment of that limitation as a mathematical concept has been withdrawn.
Applicant further argues that generating a predicted well mass flow data does not recite a judicial exception as a mathematical concept. After further consideration, examiner agrees that no mathematical concept is explicitly recited. Applicant further argues, however, that the prediction cannot be performed practically in the human mind. However, Examiner asserts that a prediction can be performed by a human practically in the human mind. The courts do not distinguish between mental processes performed in the human mind and those which are performed using a computer. The claimed language appears to be enabling the performance of a mental process by leveraging a computing system that is specified at a high level of generality.
Applicant additionally argues that the amended independent claims are not directed to extra solution activity because the activity of adjusting the physical control elements of a well system provides a practical application of improving operational efficiency over prior systems. Examiner disagrees. The specificity by which is provided in the claims to link the recited mental process is not sufficient enough to integrate the judicial exception into a practical application. The element of adjusting the control mechanism to maintain well mass flow performance based on the predicted well mass flow data recites the idea of an outcome without necessarily providing details as to how the solution to the problem is accomplished. Flow performance is allegedly maintained due to an adjustment of the control mechanism based on the predicted data. However, it is unclear how the predicted well mass flow data corresponds to the adjustment of the mechanism and the maintenance of performance. In order for the predictive aspect of the claim to be integrated into a practical application, the claim must recite a particular solution to solving a problem- i.e. how is the well mass flow data used in the adjustment of the control mechanism to achieve the maintained performance? Stating a desired outcome based on predicted data that is not clearly linked to the physical control is not enough. Accordingly, the element amounts to applying the value from the mental process to result in an idea of a solution.
For the reasons stated in this response, and in conjunction with the update rejection provided in this action, the rejection under 35 U.S.C. § 101 is maintained.
35 U.S.C. § 103
Regarding the rejections to the claims under 35 U.S.C. § 103, Applicant argues that the combination of references would not have rendered obvious the claimed invention and that the Examiner has not adequately set forth reasonable rationale as to how the combination of references would yield the claimed invention.
In response to applicant's argument that the examiner has combined an excessive number of references, reliance on a large number of references in a rejection does not, without more, weigh against the obviousness of the claimed invention. See In re Gorman, 933 F.2d 982, 18 USPQ2d 1885 (Fed. Cir. 1991). With this in mind, the Examiner has still considered Applicant’s points regarding the combination of prior art referenced with respect to the amended claimed matter.
Applicant argues that Liu does not fairly disclose the coupling of well mass flow data of the well system and modeled well mass flow data as a coupled input dataset. Applicant further argues that Liu is silent with respect to using a combination of data sources to train the neural network. Examiner agrees with this argument. Liu does disclose the coupling of measured data and a flow model to form the physics constrained machine learning model, however does not exactly disclose the coupling of the data sources as a singular input dataset. Liu also discloses only using measured data as training data. However, after further search and consideration, a new grounds of rejection has been set forth to incorporate the teachings of Karra, wherein Karra explicitly points out the simultaneous integration of measured data and modeled data as training input for a physics constrained machine learning model. In considering the obviousness of the combination of references, Examiner has provided substantial rationale as to why the references would be combined and asserts that the combination is not merely a matter of hindsight to the claimed invention. Accordingly, the claimed matter is still sufficiently disclosed in the prior art.
Applicant further argues that claim limitation denoted (iv) in the remarks document dated 09/06/2025 is not sufficiently disclosed by the prior art of record and provides rebuttal argument as to what is and is not disclosed with regard to the particular limitation. After further consideration, Examiner agrees that the previously-relied upon art does not fairly suggest the entirety of the claimed matter because the references do not suggest or clearly indicate the output of a predictive model having higher temporal density than that data generated by a physics-based model. In light of the explanation and clarity provided by the applicant, Examiner has searched the prior art for the claimed matter, and a new grounds of rejection has been set forth with consideration to the reference Kashinath which discloses the generation of sparse data and the utilization of super resolution techniques to obtain data with higher temporal resolution. Accordingly, the claimed matter is still sufficiently disclosed by the prior art.
Applicant argues that the remaining references are directed to solving different problems and do not cure the deficiencies of any allegedly uncovered claim limitations. Applicant’s arguments have been considered and are moot based on the new grounds set forth in this action. Examiner contends that prior art references directed towards solving different problems may still be considered analogous art if the claimed matter is reasonably pertinent to the problem faced by the inventor, even if the references are directed towards solving different problems. Nonetheless, in response, Examiner has provided more comprehensive rationale as to how the prior art references are reasonably pertinent to the claimed invention such that one having skill would reasonably arrive at the claimed invention.
The claims remain rejected under 35 U.S.C. § 103 for the reasons stated in this response and in conjunction with the rejection provided in this action.
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, 3-7, 9-13, and 15 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The following section follows the 2019 Patent Eligibility Guidance (PEG) for analyzing subject matter eligibility:
Step 1 - Statutory Category:
Step 1 of the PEG analysis entails considering whether the claimed subject matter falls within the four statutory categories of patentable subject matter identified by 35 U.S.C. 101 (process, machine, manufacture, or composition of matter).
Step 2A Prong 1 - Judicial exception:
In Step 2A Prong 1, examiners evaluate whether the claim recites a judicial exception (an abstract idea, law of nature, or a natural phenomenon).
Step 2a Prong 2 - Integration into a practical application:
If claims recite a judicial exception, the claim requires further analysis in Step 2A Prong 2. In Step 2A Prong 2, examiners evaluate whether the claim as a whole integrates the exception into a practical application.
Step 2B - Significantly More:
If the additional elements identified in Step 2A Prong 2 do not integrate the exception into a practical application, then the claim is directed to the recited judicial exception and requires further analysis under Step 2B- Significantly More.
As noted in the MPEP 2106.05(II): The identification of the additional element(s) in the claim from Step 2A Prong 2, as well as the conclusions from Step 2A Prong 2 on the considerations discussed in MPEP 2106.05(a) -(c), (e), (f), and (h) are to be carried over. Claim limitations identified as Insignificant Extra-Solution Activities are further evaluated to determine if the elements are beyond what is well -understood, routine, and conventional (WURC) activity, as dictated by MPEP 2106.05(II).
Independent Claims:
Claim 1:
Step 1: Claim 1 and its dependent claims 3-6 are directed to a method which falls within one of the four statutory categories of a process.
Step 2A Prong 1: Claim 1 recites a judicial exception, noted in bold:
coupling, … the well mass flow data of the well system and the modeled well mass flow data as a coupled input data set; The claim limitation can be reasonably read to entail evaluating two data streams to combine the data stream for use as input. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. For example, a human being can take two data streams and correlate the data to each other to generate a combined input dataset by writing the data values on a piece of paper. Accordingly, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process.
generating, … and the coupled input data set, predicted well mass flow data using the real-time well mass flow data; The claim limitation can be reasonably read to entail evaluating data to make a prediction of well mass flow data. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. Though this claim recites the use of a processor and a digital twin manager with a physics constrained machine learning model, these computing components are recited at a high level of generality such that the claim generally reads to using generic computing components to perform a mental process. The courts do not distinguish between mental processes performed entirely in the human mind and those performed on a computer. As such, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process
Therefore, the claim recites a judicial exception.
Step 2A Prong 2: Additional elements were identified and are noted in italics.
obtaining, by a processor of a digital twin manager and based on a predetermined monitoring criterion, well mass flow data of the well system;- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
obtaining, by the processor of the digital twin manager, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes:- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
a valve module that corresponds to and emulates a control mechanism of the well system;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; - This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
and a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
training, by the processor of the digital twin manager, a physics constrained machine learning model using one or more machine learning algorithms based on the coupled input data set;- This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for invoking the use of generic computing components to perform an existing process- that is to predict a value.
obtaining, by the processor of the digital twin manager, real-time well mass flow data of the well system; - This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
and adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on the predicted well mass flow data, - This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for reciting the equivalent of “apply it” with regard to the abstract idea in a generic manner. The element recites the idea of an outcome without necessarily providing details as to how the solution to the problem is accomplished. Flow performance is allegedly maintained due to an adjustment of the control mechanism based on the predicted data. However, it is unclear how the predicted well mass flow data corresponds to the adjustment of the mechanism and the maintenance of performance and accordingly the element amounts to applying the value from the mental process to result in an idea of a solution.
wherein the predicted well mass flow data has a time resolution that is greater than a time resolution of the real-time well mass flow data from the well system and captures non- linear dynamics behavior of the well system, and- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising time resolution details for data
wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem., - This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising these specific components of the well system
The courts have found that merely including instructions to implement an abstract idea on a computer or merely using a computer as a tool to perform an abstract idea or generically reciting the words “apply it” with regard to the judicial exception (Mere Instructions to Apply an Exception (MPEP 2106.05(f))); adding insignificant extra- solution activity to the judicial exception (Insignificant Extra Solution Activity (MPEP 2106.05(g))); and generally linking the use of a judicial exception to a particular technological environment or field of use (Field of Use and Technological Environment (MPEP 2106.05(h))) does not integrate the judicial exception into a practical application.
When viewed independently and within the claim as a whole, the additional element does not appear to integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities. Additional elements identified otherwise and conclusions from Step 2A Prong 2 are carried over for evaluating if the claim, as a whole, amounts to an inventive concept that is significantly more than the judicial exception. The following elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)):
obtaining, by a processor of a digital twin manager and based on a predetermined monitoring criterion, well mass flow data of the well system;
obtaining, by the processor of the digital twin manager, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes:
obtaining, by the processor of the digital twin manager, real-time well mass flow data of the well system;
The above limitations have been identified as insignificant extra solution activity. When read in light of the specification and under broadest reasonable interpretation, these claim limitations entail receiving data over a network. The courts have recognized this computer functionality and well-understood, routine and conventional activity when claimed in a merely generic manner such as within this claim.
The courts have found that simply appending insignificant extra solution activities that are well-understood, routine, and conventional activities to the judicial exception does not qualify the limitations as “significantly more” than the recited judicial exception. The remaining additional elements were identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) and Field of Use and Technological Environment (MPEP 2106.05(h)), as stated previously. The courts have found that merely using a computer as a tool to perform a mental process and generally linking the use of a judicial exception to a particular technological environment does not qualify the limitations as “significantly more” than the recited judicial exception.
With the additional elements viewed independently and as part of the ordered combination, the claim as a whole does not appear to amount to significantly more than the recited judicial exception because the claim is using generic computing components recited at a high level of generality and functioning in their normal capacity in conjunction with well-understood, routine, and conventional activity to enable the performance of a task that can practically be performed within the human mind or using pen and paper as an assistive physical aid. Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 U.S.C. 101.
Claim 7:
Step 1: Claim 7 and its dependent claims 9-12 are directed to a well system which falls within one of the four statutory categories of a machine.
Step 2A Prong 1: Claim 7 recites a judicial exception, noted in bold:
couples the well mass flow data of the well system and the modeled well mass flow data as a coupled input data set; The claim limitation can be reasonably read to entail evaluating two data streams to combine the data stream for use as input. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. For example, a human being can take two data streams and correlate the data to each other to generate a combined input dataset by writing the data values on a piece of paper. Accordingly, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process.
generates, … and the coupled input data set, predicted well mass flow data using the real-time well mass flow data; and The claim limitation can be reasonably read to entail evaluating data to make a prediction of well mass flow data. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. Though this claim recites the use of a processor and a digital twin manager with a physics constrained machine learning model, these computing components are recited at a high level of generality such that the claim generally reads to using generic computing components to perform a mental process. The courts do not distinguish between mental processes performed entirely in the human mind and those performed on a computer. As such, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process
Therefore, the claim recites a judicial exception.
Step 2A Prong 2: Additional elements were identified and are noted in italics.
a well site; -This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for describing the field of use by which the judicial exception is applied
a physics-based modeling server that outputs modeled well mass flow data for the well site based on a physics-based model; and -This claim has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for describing the field of use by which the judicial exception is applied
a digital twin manager, coupled to the physics-based modeling server and the well site, that includes a processor, Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the limitation invokes generic computing components recited at a high level of generality to perform an existing process
wherein the processor of the digital twin manager: Mere Instructions to Apply an Exception (MPEP 2106.05(f)) because the limitation invokes generic computing components recited at a high level of generality to perform an existing process
obtains, based on a predetermined monitoring criterion, well mass flow data of the well site;- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
obtains modeled well mass flow data for the well site using the physics-based model, wherein the physics-based model includes:- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
a valve module that corresponds to and emulates a control mechanism of the well site;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; and - This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
trains a physics constrained machine learning model using one or more machine learning algorithms based on the coupled input data set;- This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for invoking the use of generic computing components to perform an existing process- that is to predict a value.
obtains real-time well mass flow data of the well site; and - This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
wherein the control mechanism of the well site is adjusted to maintain well mass flow performance of the well site based on the predicted well mass flow data,- This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for reciting the equivalent of “apply it” with regard to the abstract idea in a generic manner. The element recites the idea of an outcome without necessarily providing details as to how the solution to the problem is accomplished. Flow performance is allegedly maintained due to an adjustment of the control mechanism based on the predicted data. However, it is unclear how the predicted well mass flow data corresponds to the adjustment of the mechanism and the maintenance of performance and accordingly the element amounts to applying the value from the mental process to result in an idea of a solution.
wherein the predicted well mass flow data has a time resolution that is greater than a time resolution of the real-time well mass flow data from the well site and captures non-linear dynamics behavior of the well site, and - This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising time resolution details for data
wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem.- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising these specific components of the well system
The courts have found that merely including instructions to implement an abstract idea on a computer or merely using a computer as a tool to perform an abstract idea or generically reciting the words “apply it” with regard to the judicial exception (Mere Instructions to Apply an Exception (MPEP 2106.05(f))); adding insignificant extra- solution activity to the judicial exception (Insignificant Extra Solution Activity (MPEP 2106.05(g))); and generally linking the use of a judicial exception to a particular technological environment or field of use (Field of Use and Technological Environment (MPEP 2106.05(h))) does not integrate the judicial exception into a practical application.
When viewed independently and within the claim as a whole, the additional element does not appear to integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities. Additional elements identified otherwise and conclusions from Step 2A Prong 2 are carried over for evaluating if the claim, as a whole, amounts to an inventive concept that is significantly more than the judicial exception. The following elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)):
obtains, based on a predetermined monitoring criterion, well mass flow data of the well site;
obtains modeled well mass flow data for the well site using the physics-based model, wherein the physics-based model includes:
obtains real-time well mass flow data of the well site; and
The above limitations have been identified as insignificant extra solution activity. When read in light of the specification and under broadest reasonable interpretation, these claim limitations entail receiving data over a network. The courts have recognized this computer functionality and well-understood, routine and conventional activity when claimed in a merely generic manner such as within this claim.
The courts have found that simply appending insignificant extra solution activities that are well-understood, routine, and conventional activities to the judicial exception does not qualify the limitations as “significantly more” than the recited judicial exception. The remaining additional elements were identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) and Field of Use and Technological Environment (MPEP 2106.05(h)), as stated previously. The courts have found that merely using a computer as a tool to perform a mental process and generally linking the use of a judicial exception to a particular technological environment does not qualify the limitations as “significantly more” than the recited judicial exception.
With the additional elements viewed independently and as part of the ordered combination, the claim as a whole does not appear to amount to significantly more than the recited judicial exception because the claim is using generic computing components recited at a high level of generality and functioning in their normal capacity in conjunction with well-understood, routine, and conventional activity to enable the performance of a task that can practically be performed within the human mind or using pen and paper as an assistive physical aid. Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 U.S.C. 101.
Claim 13:
Step 1: Claim 13 and its dependent claim 15 are directed to a non-transitory computer readable medium storing instructions executable by a computer processor which falls within one of the four statutory categories of a manufacture.
Step 2A Prong 1: Claim 13 recites a judicial exception, noted in bold:
coupling, …, the well mass flow data of the well system and the modeled well mass flow data as a coupled input data set; The claim limitation can be reasonably read to entail evaluating two data streams to combine the data stream for use as input. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. For example, a human being can take two data streams and correlate the data to each other to generate a combined input dataset by writing the data values on a piece of paper. Accordingly, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process.
generating, … and the coupled input data set, predicted well mass flow data using the real-time well mass flow data ; and The claim limitation can be reasonably read to entail evaluating data to make a prediction of well mass flow data. This task can be performed within the human mind or using a pen and paper as an assistive physical aid. Though this claim recites the use of a processor and a digital twin manager with a physics constrained machine learning model, these computing components are recited at a high level of generality such that the claim generally reads to using generic computing components to perform a mental process. The courts do not distinguish between mental processes performed entirely in the human mind and those performed on a computer. As such, this claim limitation includes the recitation of the judicial exception of abstract ideas of a mental process
Therefore, the claim recites a judicial exception.
Step 2A Prong 2: Additional elements were identified and are noted in italics.
obtaining, by the computer processor, well mass flow data of a well system based on a predetermined monitoring criterion;- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
obtaining, by the computer processor, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes:- This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
a valve module that corresponds to and emulates a control mechanism of the well system;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; and - This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the use of the judicial exception to a technological environment containing this component
training, by the computer processor, a physics constrained machine learning model using one or more machine learning algorithms based on the coupled input data set;- This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for invoking the use of generic computing components to perform an existing process- that is to predict a value.
obtaining, by the computer processor, real-time well mass flow data of the well system; - This limitation has been identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) of mere data gathering
adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on the predicted well mass flow data,- This limitation has been identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) for reciting the equivalent of “apply it” with regard to the abstract idea in a generic manner. The element recites the idea of an outcome without necessarily providing details as to how the solution to the problem is accomplished. Flow performance is allegedly maintained due to an adjustment of the control mechanism based on the predicted data. However, it is unclear how the predicted well mass flow data corresponds to the adjustment of the mechanism and the maintenance of performance and accordingly the element amounts to applying the value from the mental process to result in an idea of a solution.
wherein the predicted well mass flow data has a time resolution that is greater than a time resolution of the real-time well mass flow data from the well system and captures non- linear dynamics behavior of the well system, and- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising time resolution details for data
wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem.- This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for linking the judicial exception to this particular technological environment comprising these specific components of the well system.
The courts have found that merely including instructions to implement an abstract idea on a computer or merely using a computer as a tool to perform an abstract idea or generically reciting the words “apply it” with regard to the judicial exception (Mere Instructions to Apply an Exception (MPEP 2106.05(f))); adding insignificant extra- solution activity to the judicial exception (Insignificant Extra Solution Activity (MPEP 2106.05(g))); and generally linking the use of a judicial exception to a particular technological environment or field of use (Field of Use and Technological Environment (MPEP 2106.05(h))) does not integrate the judicial exception into a practical application.
When viewed independently and within the claim as a whole, the additional element does not appear to integrate the judicial exception into a practical application.
Step 2B: As discussed in Step 2A Prong 2, additional elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)) which must be further evaluated to determine if they are beyond WURC activities. Additional elements identified otherwise and conclusions from Step 2A Prong 2 are carried over for evaluating if the claim, as a whole, amounts to an inventive concept that is significantly more than the judicial exception. The following elements were identified as Insignificant Extra Solution Activity (MPEP 2106.05(g)):
obtaining, by the computer processor, well mass flow data of a well system based on a predetermined monitoring criterion
obtaining, by the computer processor, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes:
obtaining, by the computer processor, real-time well mass flow data of the well system;
The above limitations have been identified as insignificant extra solution activity. When read in light of the specification and under broadest reasonable interpretation, these claim limitations entail receiving data over a network. The courts have recognized this computer functionality and well-understood, routine and conventional activity when claimed in a merely generic manner such as within this claim.
The courts have found that simply appending insignificant extra solution activities that are well-understood, routine, and conventional activities to the judicial exception does not qualify the limitations as “significantly more” than the recited judicial exception. The remaining additional elements were identified as Mere Instructions to Apply an Exception (MPEP 2106.05(f)) and Field of Use and Technological Environment (MPEP 2106.05(h)), as stated previously. The courts have found that merely using a computer as a tool to perform a mental process and generally linking the use of a judicial exception to a particular technological environment does not qualify the limitations as “significantly more” than the recited judicial exception.
With the additional elements viewed independently and as part of the ordered combination, the claim as a whole does not appear to amount to significantly more than the recited judicial exception because the claim is using generic computing components recited at a high level of generality and functioning in their normal capacity in conjunction with well-understood, routine, and conventional activity to enable the performance of a task that can practically be performed within the human mind or using pen and paper as an assistive physical aid. Therefore, the claim does not include additional elements, alone or in combination that are sufficient to amount to significantly more than the recited judicial exception.
Conclusion: Based on this rationale, the claim has been deemed to be ineligible subject matter under 35 U.S.C. 101.
Dependent Claims:
Examiner notes limitations identified as judicial exceptions are indicated in italicized bold and limitations identified as additional elements are indicated using italics.
Claim 3
Step 1: Regarding dependent claim 3, the judicial exception of independent claim 1 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Step 2A Prong 1: Claim 3 does not recite any additional recitations of judicial exceptions.
Step 2A Prong 2: Claim 3 additionally recites the elements wherein the physics-based model emulates components of well mass flow behavior of the well system that are below a predetermined frequency, and and wherein the physics constrained machine learning model is trained to predict components of the well mass flow behavior of the well system that are above, below, and include the predetermined frequency. These limitations have been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for describing the technological environment by which the judicial exception is executed. The courts have ruled generally linking the use of a judicial exception to a particular technological environment or field of use does not integrate the judicial exception into a practical application. With the additional element viewed in conjunction with the other limitations, the claim as a whole does not appear to integrate the judicial exception into a practical application.
Step 2B: The courts have found that limitations that amount to generally linking the use of the judicial exception to a particular technological environment or field of use are not enough to qualify the claim as significantly more than the abstract idea. Therefore, the claim does not include additional elements, alone or in the ordered combination that are sufficient to amount to significantly more than the recited judicial exception.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 4
Step 1: Regarding dependent claim 4, the judicial exception of independent claim 1 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Step 2A Prong 1: Claim 4 does not recite any additional judicial exceptions.
Step 2A Prong 2: Claim 4 additionally recites the limitation wherein the physics constrained machine learning model is trained based on at least six months of the well mass flow data, and wherein the well mass flow data includes data for both normal operational conditions and shut- down conditions. This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for describing the technological environment by which the judicial exception is executed. The courts have ruled generally linking the use of a judicial exception to a particular technological environment or field of use does not integrate the judicial exception into a practical application. With the additional element viewed in conjunction with the other limitations, the claim as a whole does not appear to integrate the judicial exception into a practical application.
Step 2B: The courts have found that limitations that amount to generally linking the use of the judicial exception to a particular technological environment or field of use are not enough to qualify the claim as significantly more than the abstract idea. Therefore, the claim does not include additional elements, alone or in the ordered combination that are sufficient to amount to significantly more than the recited judicial exception.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 5
Step 1: Regarding dependent claim 5, the judicial exception of independent claim 1 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Step 2A Prong 1: Claim 5 does not recite any additional judicial exceptions.
Step 2A Prong 2: Claim 6 additionally recites the limitation wherein the physics constrained machine learning model is obtained using a machine learning algorithm selected from a group consisting of a Levenberg-Marquardt algorithm, a Gauss-Newton algorithm, a steepest descent algorithm, and an artificial neural network. This limitation has been identified as Field of Use and Technological Environment (MPEP 2106.05(h)) for describing the technological environment by which the judicial exception is executed. The courts have ruled generally linking the use of a judicial exception to a particular technological environment or field of use does not integrate the judicial exception into a practical application. With the additional element viewed in conjunction with the other limitations, the claim as a whole does not appear to integrate the judicial exception into a practical application.
Step 2B: The courts have found that limitations that amount to generally linking the use of the judicial exception to a particular technological environment or field of use are not enough to qualify the claim as significantly more than the abstract idea. Therefore, the claim does not include additional elements, alone or in the ordered combination that are sufficient to amount to significantly more than the recited judicial exception.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 6
Step 1: Regarding dependent claim 6, the judicial exception of independent claim 1 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Step 2A Prong 1: Claim 6 additionally recites the limitation wherein the physics constrained machine learning model uses a misfit function which includes a well mass flow prediction error, and which can reasonably be read to entail a mathematical formula that includes an error metric as the misfit function including a flow prediction error. The claim further recites wherein the well mass flow prediction error is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME. which are further recitations of mathematical formulas in their respective word forms. Therefore, this claim limitation includes the recitation of the judicial exception of abstract ideas as a mathematical concept.
Step 2A Prong 2 & Step 2B: Claim 6 does not recite any additional elements that would integrate the judicial exceptions into a practical application nor amount to significantly more than the recited judicial exceptions.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 9
Step 1: Regarding dependent claim 9, the judicial exception of independent claim 7 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Claim 9 appears to be substantially similar to claim 3 and is accordingly rejected under the same rationale.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 10
Step 1: Regarding dependent claim 10, the judicial exception of independent claim 7 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Claim 10 appears to be substantially similar to claim 4 and is accordingly rejected under the same rationale.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 11
Step 1: Regarding dependent claim 11, the judicial exception of independent claim 7 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Claim 11 appears to be substantially similar to claim 5 and is accordingly rejected under the same rationale.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 12
Step 1: Regarding dependent claim 12, the judicial exception of independent claim 7 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Claim 12 appears to be substantially similar to claim 6 and is accordingly rejected under the same rationale.
This claim is not eligible subject matter under 35 U.S.C. 101.
Claim 15
Step 1: Regarding dependent claim 15, the judicial exception of independent claim 13 is further incorporated. The claim falls within the corresponding statutory category as stated previously.
Claim 15 appears to be substantially similar to claim 3 and is accordingly rejected under the same rationale.
This claim is not eligible subject matter under 35 U.S.C. 101.
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, 3-7, 9-13, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Liu (US 2021/0124087 A1), hereinafter referred to as Liu, in view of Benish (US 2009/0198477 A1), hereinafter referred to as Benish, in view of Karra (Karra, S., Ahmmed, B., Mudunuru, M., “AdjointNet: Constraining machine learning models with physics-based codes”, September 8 2021, arXiv), hereinafter referred to as Karra, Kashinath (Kashinath, K., Mustafa, M., Albert, A., Wu, J-L. Jiang, C., Esmaeilzadeh, S., Azizzadenesheli, K., Wang, R., Chattopadhyay, A., Singh, A., Manepalli, A., Chirila, D., Yu, R., Walters, R., White, B., Xiao, H., Tchelepi, H. A., Marcus, P., Anandkumar, A., Hassanzadeh, P. and Prabhat, “Physics-informed machine learning: case studies for weather and climate modeling” , February 15, 2021, Phil. Trans. R. Soc. A., Volume 379, Issue 2194), hereinafter referred to as Kashinath, Rashid (US 9031674 B2)), hereinafter referred to as Rashid, and Lakshmanan (Charan Lakshmanan et al. "Artificial Intelligence Based Optimization of Gathering System", Abu Dhabi International Petroleum Exhibition & Conference, Abu Dhabi, UAE, November 2020.), hereinafter referred to as Lakshmanan.
Regarding claim 1, Liu discloses (except the limitations surrounded by brackets ([[..]])) A method of managing a well system, comprising: A method is disclosed for predicting production rates for hydrocarbon reservoirs ((Liu, ¶21) "The implementations disclosed herein describe methods and systems to integrate the physics of fluid flow with machine learning, or physics-constrained machine learning, to predict production rates for an unconventional hydrocarbon reservoir as a function of time.")
obtaining, by a processor of a digital twin manager and based on a predetermined monitoring criterion, well mass flow data of the well system; When read in light of the specification (¶19), a digital twin manager includes hardware and software with specified functionality equivalent to that of a computer and thus has been interpreted as such for purposes of this examination. A computer system is leveraged to execute the methods (See Liu Figure 1). Data is received from a hydrocarbon reservoir as well log data to the computer system, wherein a well log contains a record of physical measurements and the intervals of logging practice for obtaining the log are the criteria for obtaining the data ((Liu, ¶23) "Data measuring of logging instruments at the hydrocarbon reservoir 128 are communicably coupled to the computer system 100 to transmit the data 132 over a network such as the Internet 628 or the local network 622 illustrated and described in more detail later with reference to FIG. 6."); ((Liu, ¶25) " Well logging refers to the practice of making a detailed record (well log data) of physical measurements made by instruments lowered into a hydrocarbon reservoir."). The computer system comprises a computer processor that carries out the methodology ((Liu, ¶29) " The computer processors 108 are computer hardware used to perform the methods disclosed herein."). The received data includes fluid properties including density and flow which may be leveraged to compute the well mass flow rate ((Liu, ¶26) "The data 132 includes reservoir fluid pressure-volume-temperature (PVT) data of the hydrocarbon reservoir 128. The PVT data refers to the fluid properties (such as density, viscosity, or a combination) as a function of the pressure, volume, and temperature. The data 132 includes a permeability and a porosity of the hydrocarbon reservoir 128. The permeability is the ability to preferentially flow a particular fluid through a rock when other immiscible fluids are present in the reservoir. The porosity refers to an interconnected pore volume in a rock that contributes to fluid flow in the hydrocarbon reservoir 128.")
obtaining, by the processor of the digital twin manager, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes: The computer system is connected to a hydrocarbon fluid flow model, wherein the fluid flow model provides predicted hydrocarbon production rate, and the model includes parameters by which can be leveraged to determine the mass flow rate ((Liu, ¶4) "The hydrocarbon fluid flow model provides the predicted hydrocarbon production rate as a function of time based on the parameters")((Liu, ¶37) "The computer system 100 predicts the hydrocarbon production rate for the hydrocarbon reservoir 128 using the hydrocarbon fluid flow model 104. The hydrocarbon fluid flow model 104 is communicably coupled to the artificial neural network 140 to provide the predicted hydrocarbon production rate as a function of time based on the parameters 136."); ((Liu, ¶36) "The parameters 136 of the hydrocarbon fluid flow model 104 include a total hydraulic fracture area, an average fracture spacing, a permeability and a porosity of the hydrocarbon reservoir 128, a hydrocarbon in place (HIP) metric, a pressure drawdown of the hydrocarbon reservoir 128, and PVT data for producing fluids in the hydrocarbon reservoir 128 as described in more detail later with reference to FIG. 4.")
[[a valve module that corresponds to and emulates a control mechanism of the well system;]]
[[an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;]]
[[a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;]]
[[a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; ]]
[[and a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;]]
coupling, by the processor of the digital twin manager, the well mass flow data of the well system and the modeled well mass flow data [[as a coupled input data set;]]
training, by the processor of the digital twin manager, a physics constrained machine learning model using one or more machine learning algorithms [[based on the coupled input data set;]] ((Liu, ¶38) "The training data 120 includes measured reservoir characterization data and measured hydraulic fracturing data of multiple hydrocarbon reservoirs. The training data 120 further includes measured hydrocarbon production rates 208 and measured flowback rates of the multiple hydrocarbon reservoirs. The computer system 100 trains the physics-constrained machine learning model 116 to generate predicted hydrocarbon production rates for multiple reservoirs and production wells as a function of time based on the training data 120. The training is performed using backpropagation. Backpropagation is a method used to efficiently train the artificial neural network 140 following a gradient descent approach that exploits the chain rule. A feature of backpropagation is iterative, recursive and efficient calculation of the weights updates to train the artificial neural network 140."); (See also Liu Claim 6)
obtaining, by the processor of the digital twin manager, real-time well mass flow data of the well system; Data is measured from the reservoir and obtained by the computer system (Liu, ¶23 “Data measuring or logging instruments at the hydrocarbon reservoir 128 are communicably coupled to the computer system 100”); Measured data includes historical data of multiple hydrocarbon reservoirs and production wells (Liu ¶31); A hydrocarbon reservoir is described as the well site (Liu, ¶53; Figure 1, 128)
generating, by the processor of the digital twin manager using the physics constrained machine learning model and [[the coupled input data set]], predicted well mass flow data using the real-time well mass flow data Predicted hydrocarbon production rate is generated, based on data from the hydrocarbon reservoir and output by the computer system (Liu, Fig 5. 508);
[[and adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on]] the predicted well mass flow data,
wherein the predicted well mass flow data has a time resolution [[that is greater than a time resolution of the real-time well mass flow data from the well system and captures non- linear dynamics behavior of the well system, and]] A time window is described in which to predict the production rates for a hydrocarbon reservoir as a function of time (Liu ¶21, “The implementations disclosed herein describe methods and systems to integrate the physics of fluid flow with machine learning, or physics-constrained machine learning, to predict production rates for an unconventional hydrocarbon reservoir as a function of time.”)
[[wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem. ]]
Liu does not explicitly disclose; however, Liu in view of Benish discloses a valve module that corresponds to and emulates a control mechanism of the well system; A well may be equipped with downhole valves ((Benish, ¶68) "FIGS. 7A-7C are illustrative diagrams of modeling with multiple stacked reservoirs according to certain embodiments of the present invention. In FIG. 7 A, a portion of a well 700 may intersect two reservoir intervals 702, 704 and may be equipped with downhole valves 706, 708"). The valves are described as being controllable by opening and closing ((Benish, ¶68) "In a calibration mode, the top valve 706 may be opened while the bottom valve 708 remains closed."). A choke flow model is employed (as a valve module) wherein the choke flow model reflects the choke’s physical properties such as the orifice size (as the emulation of the control mechanism) ((Benish, ¶62) "Another method for comparing and back allocating production rates for commingled zones is illustrated in the flowchart of FIG. 5. In this method, the pressure drop across a choke 501 may be utilized to predict the flow rate across the choke, and therefore, from the well. Treated as a flow meter, the choke may be located at the wellhead or downhole. A single-phase or multi-phase model for flow across an orifice may be employed depending on the fluids produced. The model may include other inputs such as choke properties 503 ( e.g. size of the orifice) and fluid properties 505 ( e.g. viscosity, GOR and water cut). The model may be calibrated using a typical production well test or another suitable approach. After inputting the necessary values and exercising the choke flow model in step 502, the predicted rates for an individual well may be calculated at step 504.")
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module; A completion flow model is described, whereby the flow is modeled between the reservoir and a downhole valve. ((Benish, ¶64) "Accordingly, a series of models may be employed to represent the pressure drop across and the flow rates of different portions of the well completion. A well completion can be any configuration of hardware including, but not limited to, screens, valves, casing, tubing, gravel, nipples, or fixed chokes."); See ¶64-73 for complete discussion of “An Exemplary Completion Pressure Drop Modeling and Allocation Adjustment Method”.; See also Fig. 6, item 602; See also provided annotations of Figure 7a.
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir; A tubing size of the well is known, indicating a constant volume segment of the well system ((Benish, ¶59) "The downhole pressure 401 and a downstream pressure 403 ( e.g. wellhead pressure) may be measured for a given well whose properties 405, such as tubing size and roughness, may be known."). A downhole valve is present between the constant volume midstream chamber and the reservoir (Benish, Figure 7a, Item 708). A choke pressure drop model can be employed for modeling well mass flow ((Benish, ¶62) " In this method, the pressure drop across a choke 501 may be utilized to predict the flow rate across the choke, and therefore, from the well. Treated as a flow meter, the choke may be located at the wellhead or downhole. A single-phase or multi-phase model for flow across an orifice may be employed depending on the fluids produced. The model may include other inputs such as choke properties 503 ( e.g. size of the orifice) and fluid properties 505 ( e.g. viscosity, GOR and water cut)."). See ¶62-63 for complete discussion of “An Exemplary Choke Pressure Drop Modeling and Allocation Adjustment Method”.; See also provided annotations of Figure 7a.
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; A tubing size of the well is known, indicating a constant volume segment of the well system ((Benish, ¶59) "The downhole pressure 401 and a downstream pressure 403 ( e.g. wellhead pressure) may be measured for a given well whose properties 405, such as tubing size and roughness, may be known."). Mass flow can be modeled in the wellbore between two points ((Benish, ¶60) "The wellbore flow model may predict the rates for each well (1-n) by modeling the relationship between fluid flow rates and the pressure drop between two points. Such a model may incorporate basic physics including hydrostatic losses, frictional losses, and form losses, such as diverging or converging flow though tubulars of changing diameters."). The constant volume tube is connected to the well head via a valve ((Benish, ¶62) "Treated as a flow meter, the choke may be located at the wellhead or downhole"); See ¶59-61 for complete discussion of “An Exemplary Wellbore Modeling and Allocation Adjustment Method”.; See also figure 7, item 706; See also provided annotations of Figure 7a.
and a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head; A tubing size of the well is known, indicating a constant volume segment of the well system ((Benish, ¶59) "The downhole pressure 401 and a downstream pressure 403 ( e.g. wellhead pressure) may be measured for a given well whose properties 405, such as tubing size and roughness, may be known."). A wellhead valve is present between the constant volume midstream chamber and the well head (Benish, Figure 7a, Item 706). A choke pressure drop model can be employed for modeling well mass flow ((Benish, ¶62) "In this method, the pressure drop across a choke 501 may be utilized to predict the flow rate across the choke, and therefore, from the well. Treated as a flow meter, the choke may be located at the wellhead or downhole. A single-phase or multi-phase model for flow across an orifice may be employed depending on the fluids produced. The model may include other inputs such as choke properties 503 ( e.g. size of the orifice) and fluid properties 505 ( e.g. viscosity, GOR and water cut)."). See ¶62-63 for complete discussion of “An Exemplary Choke Pressure Drop Modeling and Allocation Adjustment Method”.; See also provided annotations of Figure 7a.
…and the physics-based model includes a set of initial, second, connection, and terminal stages…. The submodels as taught by Benish previously can be components of a global model. ((Benish, ¶74) " In a further exemplary embodiment of the present invention, combinations of models including those described above may be employed to back allocate commingled zone rates as illustrated in FIG. 8"). For further clarity, Examiner has annotated Figure 7 of Benish to emphasize where the sub-models described previously would apply in the zones of the well system.
PNG
media_image1.png
721
445
media_image1.png
Greyscale
PNG
media_image2.png
729
751
media_image2.png
Greyscale
Benish is analogous art to the claimed invention because it is related to the same field of endeavor of optimizations of well system operations. When read in light of the specification, a flow restriction module is defined as “module emulating a general pressure drop between the reservoir and the well subsystem” (¶49, ¶50, ¶52). Benish teaches the usage of models for calculating a pressure drop over different zones in a global model of the well system. The well system as taught by Benish contains the reservoir, a constant volume chamber, and a well head whereby to incorporate sub-models that describe changes in pressure for various zones. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have combined the prior art references, because substituting one prior art element for another would yield predictable results. Liu discloses a physics-based hydrocarbon fluid flow model representative of the hydrocarbon fluid flow from the reservoir to a hydraulic fracture but provides few details as to the specifics of the infrastructure ((Liu, ¶37) "The hydrocarbon fluid flow model 104 represents a physics-based one-dimensional hydrocarbon fluid flow from the hydrocarbon reservoir 128 to one or more hydraulic fractures"). Benish describes zones of the well system as areas of rock which are connected to fractures ((Benish, ¶37) "As used herein, the term "zone" generally refers to an interval or unit of rock differentiated from surrounding rocks on the basis of its fossil content or other features, such as faults or fractures. The term zone may also refer to a discrete well within a commingled group, a section of a well with stacked intervals, or a portion of a reservoir.") and subsequently provides a comingled zone model for pressure drop as a particular implementation ((Benish, ¶62) "Another method for comparing and back allocating production rates for commingled zones is illustrated in the flowchart of FIG. 5. In this method, the pressure drop across a choke 501 may be utilized to predict the flow rate across the choke, and therefore, from the well."). By applying the specific model described by Benish as the hydrocarbon production model as in disclosed by Liu, one would arrive at the claimed invention. One would be compelled to make this particular substitution because commingling zones of a well system is a standard practice in the art and understanding the production rates and volume information within each zone yields predictable results for reservoir management and reservoir simulation applications. ((Benish, ¶4-5) "A common oil and gas industry practice is to commingle production, either downhole or at the surface. Where multiple zones are present in a reservoir, an economic development plan would involve commingling the production of at least some of the zones. Commingling production is also cost effective in multiple well scenarios, such as on a platform. These practices lead to uncertainty in how much fluid is produced from a given zone, such as a well or reservoir, because a single flow rate measurement is commonly taken downstream of the commingling. Both the production rate and volume information allocated to each zone are useful for many reservoir surveillance and management tasks. For example, many techniques for estimating the remaining producible oil or gas in a reservoir depend on accurately knowing the amount of oil and/or gas produced from the reservoir in combination with the downhole pressure. History-matching oil, gas and water production data using a reservoir simulator is a common practice in workflows used for making investment decisions, such as whether or not to drill more wells or perform remedial operations on the well.").
The proposed combination does not disclose; however, the proposed combination in view of Karra discloses combining measured and synthetic data as a coupled input data set to a machine learning model. Simulated data is simultaneously integrated with observational data as input to ML workflows ((Karra, Page 2, ¶4) " Our computational approach (see Fig. (1)(b)), AdjointNet, allows domain scientists to incorporate simulators into scalable ML workflows without any loss of physics (up to the accuracy of a simulator) and simultaneously integrate observational data.")
Karra further discloses training is based on the coupled input data set; ((Karra, Page 2, ¶4) " Our computational approach (see Fig. (1)(b)), AdjointNet, allows domain scientists to incorporate simulators into scalable ML workflows without any loss of physics (up to the accuracy of a simulator) and simultaneously integrate observational data. Moreover, our method can directly assimilate real-time data streams in the training step.")
Furthermore, Karra discloses generating predictions using a machine learning model that leverages the coupled input data set. ((Karra, Page 2, ¶4) " Our computational approach (see Fig. (1)(b)), AdjointNet, allows domain scientists to incorporate simulators into scalable ML workflows without any loss of physics (up to the accuracy of a simulator) and simultaneously integrate observational data. Moreover, our method can directly assimilate real-time data streams in the training step. The trained ML models can then be used to forecast system behavior or inform system optimization. ")
Karra further describes a methodology which is able to make predictions captures non- linear dynamics behavior of the well system, and. The framework is described as including the solving of nonlinear partial different equations characterized by the dynamics of multiple physical properties of a system with fluid flow components ((Karra, Page 4, ¶)2 " For instance, in the case of porous media, the parallel code PFLOTRAN solves the discrete form of the non-linear diffusion PDE [22]: [[eq 2.3]] (2.3) where u is the fluid pressure, k is the permeability parameter, - is porosity, is fluid viscosity, and is the fluid density. PDE discretization in PFLOTRAN is performed using a two-point flux finite volume method with backward Euler for time-stepping. The resulting nonlinear algebraic equations are solved using a Newton-Krylov solver. ")
Karra is analogous to the claimed invention because it is reasonably pertinent to the problem faced by the inventor- that is using predictive computing technologies to characterize media flow. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have incorporated the simultaneous integration of observational data and simulated data as input to the training of a ML model because combining prior art elements according to known methods would have led one having ordinary skill to do so. Liu discloses the coupling of a hydrocarbon fluid flow model to an artificial neural network to constrain the outputs of the neural network, thereby indicating that the data is coupled ((Liu, ¶22) " The hydrocarbon fluid flow model is communicably coupled to the artificial neural network in a manner to constrain the outputs of the artificial neural network. The artificial neural network is thus constrained to provide meaningful outputs in accordance with the physics-based fluid flow of the hydrocarbon reservoir"). Liu does not appear to contemplate the coupling of the models so as to form a couple dataset as an input for training, and Liu relies on purely measured data for the training of the physics-constrained machine learning model ((Liu, ¶31) "The computer system 100 extracts the training data 120 for the physics-constrained machine learning model 116 from measured, historical data of multiple hydrocarbon reservoirs and multiple production wells. In some implementations, the training data 120 is stored on the storage device 112. The training data 120 includes measured reservoir characterization data, measured hydraulic fracturing data, measured hydrocarbon production rates, and measured flowback rates of the multiple hydrocarbon reservoirs and multiple production wells. The computer system 100 aggregates the measured, historical hydrocarbon production rates as a function of time from the multiple hydrocarbon reservoirs and multiple production wells into the training data 120 to verify the hydrocarbon production prediction results and train the physics-constrained machine learning model 116."). Kara recognizes that most physics-informed machine learning approaches doe not include simulation data and experiment observations in the loop during the ML training process and provides proposed method to overcome this technical gap by incorporating physics-based modeling as part of the training procedure ((Karra, Page 2, ¶3) " Most PIML/KGML methods do not include simulation data and experimental observations in the loop during the ML training process. Training is usually a two-step process, where an ML model is first built from simulations. Then real-time data streams are integrated later to infer system decision parameters (as shown in Fig. (1)(a)). If new data is assimilated beyond the training range, one needs to generate more training data from simulations and then re-train the ML model. Our proposed method overcomes this serious technical gap by directly incorporating physics-based codes into the ML training procedure. We can use our approach to perform ML while ensuring that a given physics constraint (e.g., the balance of mass, the balance of momentum) is satisfied everywhere, as modeled by a physics-based code. Our innovation is the use of adjoint sensitivities, which can be obtained directly from physics calculations, to calculate the gradients of the loss function in the training process of a neural network. Since one needs to solve the forward problem using the physics code to calculate these adjoint sensitivities, the underlying physics will be ‘automatically’ satisfied and constrained. Due to the ‘correct’ constraining of physics in the entire domain, one can still get highly predictive ML models with limited, sparsely-sampled, and noisy data."). As noted above, the approach enables highly predictive ML models with limited, sparsely sampled, and noisy data, and therefore the results of leveraging this methodology would have been predictable in this way. Accordingly, the combination of references would have been obvious.
The proposed combination does not disclose; however the proposed combination in further view of Kashinath discloses a prediction with temporal resolution that is greater than a time resolution of the real-time well mass flow data from the well system and. Super resolution is achieved using a framework called MeshfreeFlowNet to generate continuous spatio-temporal resolution of complex systems from low resolution inputs ((Kashinath, Page 20, ¶5) " Here, we review MeshfreeFlowNet, a novel SR framework to generate continuous (grid-free) spatio-temporal solutions of complex systems from low-resolution inputs [104]. "); ((Kashinath, Page 22, ¶2) " MeshfreeFlowNet’s effectiveness is tested on the turbulent RBC (see [104] for details on the problem set-up, solvers and datasets). Figure 10 shows sample tuples of low-resolution input data, the high-resolution super-resolved data by MeshfreeFlowNet, and the ground truth high resolution data for the four physical parameters of the RBC system, i.e. T, p, u,w, respectively, as the temperature, pressure, and the x and z components of the velocity. The super-resolved data is essentially indistinguishable from the true high-resolution data. ")
Kashinath is analogous art to the claimed invention because the approach is reasonably pertinent to the problem faced by the inventor- that is making high fidelity predictions based on sparse data inputs for dynamic and complex systems. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have incorporated the super-resolution framework disclosed by Kashinath because known work in one field of endeavor may prompt variations of it for use in a different one based on design incentives and the variant would have been predictable to one having skill in the art. While Kashinath does not disclose the super-resolution of well mass flow data and particularly pertains to weather and climate data, the methodology is applicable to integrate into the proposed combination because Kashinath describes using physics-informed machine learning approaches for making predictions of complex systems for coupled processes. Kashinath suggests that the framework has applicability to applications in realistic three-dimensional turbulent flows ((Kashinath, Page 22, ¶5) " MeshfreeFlowNet SR framework presented above has many powerful features: spatio-temporal coherence; PDE-constrained loss; improved performance on physically motivated metrics; the ability to super-resolve at arbitrary spatial and temporal locations (grid-free) on arbitrarily large domains; generalizability; and high scalability. Thus it is well-poised for applications in realistic three-dimensional turbulent flows in the atmosphere and ocean. "). Kashinath further suggests that high-resolution data is necessary for understanding scientific phenomenon better for planning purposes ((Kashinath, Page 16, ¶4) "Accurate and reliable high-resolution weather and climate data are essential for understanding scientific phenomena better and for a wide range of climate impact studies, planning and policy making under climate change. This is especially important in the event of highly localized phenomena such as weather and climate extremes, in urban areas, and in regions with high topographic complexity and sharp gradients like mountains or coastal regions. However, fully resolving these complex systems in conventional numerical weather and climate models is intractable and most observational datasets do not contain reliable information at the fine scales. Therefore, there is a pressing need for efficient and accurate methods to enhance the resolution of weather and climate data. "). One having skill in the art would recognize applicability of such methods to complex systems such as well systems with coupled processes. The proposed combination (particularly with respect to Liu) provides a physics-constrained machine learning model for predicting well data and Kashinath provides super resolution and downscaling techniques for physics-informed machine learning models that could reasonably be applied towards a different field of use. Accordingly, by applying super resolution techniques to the physics-constrained machine learning model, one would arrive at the claimed invention and the combination would yield the predictable results of having high-fidelity data predictions by which to leverage for planning and phenomenological understanding purposes.
The proposed combination does not disclose; however, the proposed combination in view of Rashid discloses adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on a predicted value. A solution is generated (as a predicted value) for optimizing production of wells (as the maintenance of flow performance to the ideal value) and is then used to adjust an operating parameter of the well to generate specified wellhead pressure ((Rashid, Col 2, Lines 5-22) "The offline model generates, for each well, an intermediate solution to optimize the production of each well using the data collected by the sensors, a mixed-integer nonlinear program solver, and production curves based on a choke state and a given offline wellhead pressure. The network model calculates, using each intermediate solution of each well, a current online wellhead pressure for each well. The current online wellhead pressure for each well is dependent upon each other well. The optimization engine sets the intermediate solution as a final solution based on determining that a difference between the current online wellhead pressure of each well and a prior online wellhead pressure of the each well is less than a tolerance amount. The final solution of each well identifies a value for the at least one operating parameter. The optimization engine further adjusts, using final solution of each well, at least one operating parameter of the wells to generate an actual wellhead pressure of each well."); ((Rashid, Col 5, Lines 46-50) "In particular, the state of various production equipment, such as choke position and amount of lift gas, may be dictated by the final solution(s) generated by the iterative online-offline procedure, thereby adjusting the production of hydrocarbons in the oilfield."). The operating parameters of the well include the state of a choke (as a control mechanism of the well) ((Rashid, Col. 7, Lines 12-19) "Examples of data that the user (250) may receive from the production computer system (202) include, but are not limited to, operating parameters (including a choke state of one or more chokes), production data, production profiles, and lift gas flow. The user (250) may be an engineer, a production supervisor, an accountant, a risk manager, or some similar individual or software program involved in the production of the network of wells."). The choke is described as being opened, closed or at other positions in between ((Rashid, Col 5, Lines 13-15) "The choke (136) may be adjusted to a fully open position, a fully closed position, or any position between fully open and fully closed.")
Examiner notes that the inclusion of this reference more explicitly maps to the recitation of digital twin throughout the claim, wherein the digital representation of the system is coupled for real time monitoring, prediction, and control of the well system. Accordingly, by incorporating such functionality, the computer system of the proposed combination is more clearly indicative of a digital twin manager.
Rashid as analogous art because it is related to the same field of endeavor of optimizations for wells using computing technologies. It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the adjustment of the control mechanism of the well based on predicted values into the system of the proposed combination because some teaching, suggestion, or motivation in the prior art references would have led one having ordinary skill in the art to do so in order to arrive at the claimed invention. Liu discloses the use of a physics constrained machine learning model to produce an estimated production rate for a hydrocarbon reservoir which is displayed to a user for determining if a production well should be drilled at a particular location, which is a different intended use of the physics constrained machine learning model. Rashid discloses the use of an oilfield production network model to obtain well data, optimize well data, and subsequently control a physical well based on an online model, which is a digital representation (or twin) of the actual system during current operation ((Rashid, Col 9, Lines 31-36) "Specifically, the network model (208) performs one or more simulations while considering all wells in an integrated network. The network model (208) may be referred to as an online model. The network model (208) is a representation of the actual gathering system at prevailing time."). Rashid suggests that such configuration enables improved operations ((Rashid, Col 1, Lines 20-32) " At times, production engineers use computer models of the network for production operation purposes. Such models simulate multi-phase flow behavior through the network and are used for investigation and prediction purposes, such as, but not limited to, production monitoring, facility design and sizing, scenario analysis, multi-phase flow assurance, and field pressure management to ensure hydrocarbon flow to a delivery sink.) From an operational point of view, one objective is to maximize production (or profit) from the produced, and saleable, oil and gas components, while minimizing production costs and meeting all existing operating constraints."). By integrating the physics constrained machine learning model disclosed by Liu as part of the optimization application of the live system disclosed by Rashid, one would arrive at the claimed invention comprising the monitoring, optimization, and control of the well system. One having skill would be particularly compelled to do so since Rashid suggests a use case where production predictions may alternatively be beneficial for operations purposes as opposed to purely for planning purposes (as the intended use of Liu). Accordingly, the combination would have been obvious.
The proposed combination does not disclose; however, the proposed combination in view of Lakshmanan discloses wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem… and …for the sales header subsystem A compressor station within a well system is depicted in (Lakshmanan, Figure A3) and a sales point component of the well system is depicted in (Lakshmanan, Figure A4).
Lakshmanan is analogous to the claimed invention because it is related to the same field of endeavor of hydrocarbon well production systems. It would have been obvious to one of ordinary skill in the art to which said subject matter pertains at the time in which the invention was filed to have implemented the inclusion of sales header subsystems and compressor subsystems taught by Lakshmanan into the proposed combination because combining prior art elements according to known methods would yield predictable results. Lakshmanan discloses components of well systems to include compressor stations and sales point outlets wherein such components are utilized in a digital twin application for managing operations. In applying the proposed combination to the use case of control and production optimizations of well systems, it would be reasonable to include the sales points as part of the system because revenue generation ultimately drives the need for optimal production.
Regarding claim 3, the proposed combination discloses The method of claim 1, as stated previously.
The proposed combination in further view of Liu discloses (except the limitations surrounded by brackets ([[..]])) wherein the physics-based model emulates components of well mass flow behavior of the well system [[that are below a predetermined frequency, and]] A physics based hydrocarbon fluid flow model for predicting production rates is described ((Liu, ¶17) "FIG. 4 illustrates a physics-based hydrocarbon fluid flow model for predicting hydrocarbon production rates for an unconventional hydrocarbon reservoir, in accordance with one or more implementations."). A simulation is leveraged for one dimensional simulated fluid flow as components of flow behavior (Liu ¶ 46 “The one-dimensional fluid flow is simulated either using numerical methods or analytical methods.”); ((Liu, ¶30) " The hydrocarbon fluid flow model 104 provides the predicted hydrocarbon production rate as a function of time based on the parameters 136").
wherein the physics constrained machine learning model is trained to predict components of the well mass flow behavior of the well system that are above, below, and include the predetermined frequency. ((Liu, ¶21) “The physics - constrained machine learning model includes an artificial neural network to generate parameters that correspond to the physical mechanisms for fluid flow in the unconventional hydrocarbon reservoir. The physics - constrained machine learning model includes a reservoir flow model to predict hydrocarbon production rates for the hydrocarbon reservoir as a function of time based on the parameters). A threshold is imposed for a daily production as a predetermined frequency, wherein the threshold is evaluated for not being satisfied by the predicted data (Liu ¶34 “The computer system determines whether to drill the production well at the particular geographical location based on the predicted hydrocarbon production rate for the hydrocarbon reservoir 128. For example, if the predicted hydrocarbon production rate at a particular time is greater than a threshold (such as 150,000 bbl./day), the computer system 100 can determine that the geographical location is a sweet spot for drilling a production well”)
The proposed combination in further view of Liu does not particularly disclose; however the proposed combination in further view of Kashinath discloses using coarse data generated by a numerical simulation as that are below a predetermined frequency, and which are subsequently processed for super resolution. ((Kashinath, Page 17, ¶6) "SR is performed on 15 years of wind velocity fields from a numerical simulation of the Weather Research and Forecasting (WRF) model over southern California. The spatial resolution of the data is 1.5km and temporal resolution is hourly. SR enhances the spatial resolution by 4× in each dimension (see [103] for more details on the dataset and the training procedure). The proposed physics-based SR method, PSD-Net, is compared against three baselines: (i) the standard ESRGAN; (ii) SR-CNN, a CNN architecture used by Vandal et al. [118]; and (iii) upsampling using bicubic interpolation. Figure 6a shows a schematic of a GAN for SR.")
It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have generated a coarse dataset as the physics-based simulated dataset because combining prior art references to include this would yield predictable results. Generating datasets with coarse values would yield the predictable results of accurately reflecting sparse datasets representative of that which may be obtained. Additionally, generating coarse datasets would reduce computational complexity of the simulation. Kashinath demonstrates that leveraging super-resolution techniques mitigates the needs for high-resolution simulated data since using the approach yields detailed results ((Kashinath, Page 28, ¶7) "In this article,we review progress in PIML towards addressing some critical challenges in weather and climate modelling, namely: (i) building better emulators for complex multi-scale physical processes; (ii) downscaling (super-resolving) coarse data to produce high-fidelity high-resolution data; and (iii) forecasting the spatio-temporal dynamics of the atmosphere and ocean. Using the 10 approaches described in §2b, the case studies characterized in table 1 illustrate"). Further, Kashinath suggests training models on sparse and incomplete data in order to develop robust and reliable physics informed machine learning models ((Kashinath, Page 30, item 4) "Train models with data characteristics, such as noise, sparsity, and incompleteness, that are representative of the downstream application."). Accordingly, it would have been obvious to further modify the proposed combination.
Regarding claim 4, the proposed combination discloses The method of claim 1, as stated previously.
The proposed combination in further view of Liu discloses (except the limitations surrounded by brackets ([[..]])) wherein the physics constrained machine learning model is trained based on at least six months of the well mass flow data, and wherein the well mass flow data includes data for both [[normal operational conditions and shut- down conditions. ]] The physics constrained machine learning model is trained on historical data ((Liu, ¶31) " The computer system 100 extracts the training data 120 for the physics-constrained machine learning model 116 from measured, historical data of multiple hydrocarbon reservoirs and multiple production wells. In some implementations, the training data 120 is stored on the storage device 112. The training data 120 includes measured reservoir characterization data, measured hydraulic fracturing data, measured hydrocarbon production rates, and measured flowback rates of the multiple hydrocarbon reservoirs and multiple production wells. The computer system 100 aggregates the measured, historical hydrocarbon production rates as a function of time from the multiple hydrocarbon reservoirs and multiple production wells into the training data 120 to verify the hydrocarbon production prediction results and train the physics-constrained machine learning model 116")
For purposes of this examination, the claim limitation based on at least six months is understood to include non-functional descriptive material (MPEP 2111.05(III)). One of ordinary skill in the art would recognize the specific age of the historical data as a matter of routine optimization of the claimed device (MPEP 2144(II)(A). It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the proposed combination to optimally select historical data (Liu ¶31) including at least six months of history to ensure that an appropriate amount of data has been obtained in-order to adequately train the model.
The proposed combination in further view of Liu does not disclose; however, in further view of Lakshmanan discloses normal operational conditions and shut- down conditions. ((Lakshmanan, Page 3 ¶5 & 4 ¶1) "Source of data. Data sets utilized for developing digital twin solution comprises of real-time operations data & stationary information. The sources of these data sets are as follows: Sensor captured live operations data is connected to a server in the cloud through gateways. From this server, the data is transmitted to the digital twin platform in real-time.O ffline stationary/static information like compressor curves, water-gas ratios, etc. is read by the digital twin platform in the form of spreadsheets. ")
It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have included shut down conditions taught by Lakshmanan as part of the well system operational data captured by the proposed combination to yield the predictable results of having a complete and comprehensive dataset of a well system for all operating conditions to implement a realistic digital twin solution (“Data sets utilized for developing digital twin solution comprises of real-time operations data & stationary information”, Lakshmanan, p. 3, Source of Data).
Regarding claim 5, the proposed combination discloses The method of claim 1: as stated previously.
The proposed combination in further view of Liu discloses wherein the physics constrained machine learning model is obtained using a machine learning algorithm selected from a group consisting of a Levenberg-Marquardt algorithm, a Gauss-Newton algorithm, a steepest descent algorithm, and an artificial neural network. ((Liu, ¶30) " The physics-constrained machine learning model 116 includes a combination of a physics-based reservoir flow model (the hydrocarbon fluid flow model 104) and an artificial neural network 140 (a machine learning algorithm).")
Regarding claim 6, the proposed combination discloses The method of claim 1: as stated previously.
The proposed combination in view of Karra discloses (except the limitations surrounded by brackets ([[..]])) wherein the physics constrained machine learning model uses a misfit function which includes a well mass flow prediction error, and A loss function characterizing the fit of the data to the original data is disclosed for pressure data ((Karra, Page 6, ¶1) "The loss function is formulated based on the mean-squared error between the true and predicted pressure values at these ten points.").
[[wherein the well mass flow prediction error is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME. ]]
It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the utilization of a prediction error into the proposed combination for the value being predicted because combining prior art methods according to known methods would yield predictable results. In order to ensure that the predictive model achieves desired accuracy, one having skill in the art would obviously be compelled to quantify their predicted value compared to true measured values so as to ensure the model is tuned correctly. Accordingly, the combination would have been obvious.
The proposed combination in further view of Karra does not disclose; however the proposed combination in further view of Lakshmanan discloses using a prediction error that is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME. ((Lakshmanan, Page 21, ¶3) "Tuning the type of loss function (between mean absolute error & mean squared error. ")
It would have been obvious to one of ordinary skill to which said subject matter pertains at the time the invention was filed to have implemented the mean error as the loss function because simple substitution of one known element for another would yield predictable results. The proposed combination in view of Karra discloses the utilization of a loss function characterized by mean squared error. Lakshmanan discloses the adjustment of the loss function for hyperparameter tuning of a predictable model. By substituting the particular loss function for another, one having skill in the art would have predictable results of having a modified metric by which to evaluate the performance of a predictive model. Accordingly, the combination would have been obvious.
Regarding claim 7, the limitations are substantially similar to that recited in claim 1. The limitations not previously discussed with regard to claim are discloses by Liu: A well system, comprising:
a well site; as the well including the hydrocarbon reservoir (Liu, ¶25; ¶26; Figure 1, 128);
a physics-based modeling server as the hydrocarbon fluid flow model that “is communicably coupled to the artificial neural network” (Liu ¶22 Figure 2, 104) that outputs modeled well mass flow data for the well site based on a physics-based model; and as meaningful outputs in accordance with the physics-based reservoir flow model (“The hydrocarbon fluid flow model 104 provides the predicted hydrocarbon production rate as a function of time based on the parameters” Liu ¶30)
a digital twin manager, coupled to the physics-based modeling server and the well site, that includes a processor, A computer system comprises a physics constrained machine learning model which comprises the physics based flow model ((Liu, ¶30) " The physics-constrained machine learning model 116 includes a combination of a physics-based reservoir flow model (the hydrocarbon fluid flow model 104) and an artificial neural network 140 (a machine learning algorithm)."). The computer system received well log data, thereby indicating connection to a well site ((Liu, ¶4) " In some implementations, the methods include using a computer system to receive data from a hydrocarbon reservoir. The data includes reservoir characterization data, well log data, and hydraulic fracturing data of the hydrocarbon reservoir. The computer system includes a physics-constrained"). The computer system contains or more computer processors (Liu ¶29, Figure 6, 108),
wherein the processor of the digital twin manager: The computer system contains or more computer processors (Liu ¶29, Figure 6, 108),
The remaining limitations are rejected as stated previously in the rejection of claim 1:
obtains, based on a predetermined monitoring criterion, well mass flow data of the well site;
obtains modeled well mass flow data for the well site using the physics-based model, wherein the physics-based model includes:
a valve module that corresponds to and emulates a control mechanism of the well site;
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; and
a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;
couples the well mass flow data of the well system and the modeled well mass flow data as a coupled input data set;
trains a physics constrained machine learning model using one or more machine learning algorithms based on the coupled input data set;
obtains real-time well mass flow data of the well site; and
generates, by using the physics constrained machine learning model and the coupled input data set, predicted well mass flow data using the real-time well mass flow data; and
wherein the control mechanism of the well site is adjusted to maintain well mass flow performance of the well site based on the predicted well mass flow data,
wherein the predicted well mass flow data has a time resolution that is greater than a time resolution of the real-time well mass flow data from the well site and captures non-linear dynamics behavior of the well site, and
wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem.
Regarding claim 9, the proposed combination discloses The well system of claim 7, as stated previously.
The remaining limitations: wherein the physics-based model emulates components of well mass flow behavior of the well site that are below a predetermined frequency, and
wherein the physics constrained machine learning model is trained to predict components of the well mass flow behavior of the well site that are above, below, and include the predetermined frequency. are substantially similar to that recited in claim 3 and are therefore rejected under the same rationale.
Regarding claim 10, the proposed combination discloses The well system of claim 7 as stated previously.
The remaining limitations: wherein the physics constrained machine learning model is trained based on at least six months of the well mass flow data, and
wherein the well mass flow data includes data for both normal operational conditions and shut- down conditions. are substantially similar to that recited in claim 4 and are therefore rejected under the same rationale.
Regarding claim 11, the proposed combination discloses The well system of claim 7: as stated previously.
The remaining limitations: wherein the physics constrained machine learning model is obtained using a machine learning algorithm selected from a group consisting of a Levenberg-Marquardt algorithm, a Gauss-Newton algorithm, a steepest descent algorithm, and an artificial neural network. are substantially similar to that recited in claim 5 and are therefore rejected under the same rationale.
Regarding claim 12, the proposed combination discloses The well system of claim 7: as stated previously. The remaining limitations: wherein the physics constrained machine learning model uses a misfit function which includes a well mass flow prediction error, and
wherein the well mass flow prediction error is selected from a group consisting of integral square error (ISE), mean error (ME), normalized ISE, and normalized ME. are substantially similar to that recited in claim 6 and are therefore rejected under the same rationale.
Regarding claim 13 the limitations are substantially similar to that recited in claim 1. The limitations not previously discussed with regard to claim are discloses by Liu: A non-transitory computer readable medium storing instructions executable by a computer processor, the instructions comprising functionality for: (Liu, See claim 9, “Anon-transitory computer-readable storage medium storing instructions executable by a computer system, the instructions when executed by the computer system cause the computer system to:”)
The remaining limitations are rejected as stated previously in the rejection of claim 1:
obtaining, by the computer processor, well mass flow data of a well system based on a predetermined monitoring criterion;
obtaining, by the computer processor, modeled well mass flow data for the well system using a physics-based model, wherein the physics-based model includes:
a valve module that corresponds to and emulates a control mechanism of the well system;
an initial stage that models well mass flow from a model reservoir including a reservoir flow restriction module;
a second stage that models well mass flow from a constant volume midstream chamber, including a midstream flow restriction module, connected to the model reservoir;
a connection branch stage that models well mass flow from the constant volume midstream chamber connected to a model well head; and
a terminal stage that models well mass flow from a constant volume chamber including a terminal flow restriction model in the model well head;
coupling, by the computer processor, the well mass flow data of the well system and the modeled well mass flow data as a coupled input data set;
training, by the computer processor, a physics constrained machine learning model using one or more machine learning algorithms based on the coupled input data set;
obtaining, by the computer processor, real-time well mass flow data of the well system;
generating, by the computer processor using the trained physics constrained machine learning model and the coupled input data set, predicted well mass flow data using the real-time well mass flow data ; and
adjusting the control mechanism of the well system to maintain well mass flow performance of the well system based on the predicted well mass flow data,
wherein the predicted well mass flow data has a time resolution that is greater than a time resolution of the real-time well mass flow data from the well system and captures non- linear dynamics behavior of the well system, and
wherein the well system comprises interconnected subsystems that include a compressor subsystem and a sales header subsystem, and the physics-based model includes a set of initial, second, connection, and terminal stages for the sales header subsystem.
Regarding claim 15, the proposed combination discloses The non-transitory computer readable medium of claim 13, as stated previously.
The remaining limitations: wherein the physics-based model emulates components of well mass flow behavior of the well system that are below a predetermined frequency, and
wherein the physics constrained machine learning model is trained to predict components of the well mass flow behavior of the well system that are above, below, and include the predetermined frequency. are substantially similar to that recited in claim 3 and are thus rejected under the same rationale.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to EMILY GORMAN LEATHERS whose telephone number is (571)272-1880. The examiner can normally be reached Monday-Friday, 9:00 am-5:00 pm ET.
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 (571) 272-3652. 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.
/E.G.L./Examiner, Art Unit 2187
/ANDRE PIERRE LOUIS/Primary Patent Examiner, Art Unit 2187 December 1, 2025
Prosecution Insights