DETAILED ACTION
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 .
Election/Restrictions
Applicant’s election with traverse of group II in the reply filed on 05/14/2026 is acknowledged. Upon reconsideration, the restriction requirement mailed on 09/04/2025 is improper, because it was based on linking claims rather than on a restriction of species, which doesn’t apply in this instance. Accordingly, the restriction requirement is hereby withdrawn. Claim 2-4, 15-17, and 28-30, previously withdrawn from consideration, are rejoined.
Claim Status
Claims 1-39 are pending and are presented for examination.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 2023/10/09. The submission is in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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 1–10, 12–23, 25–36, 38, and 39 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (an abstract idea) without reciting significantly more.
Regarding independent claims 1, 14, and 27
Step 1 -- whether the claim falls within any statutory category. See MPEP 2106.03
Independent claim 1 is drawn to a method. Independent claim 14 is drawn to an apparatus comprising at least one processing device. Independent claim 27 is drawn to a non-transitory machine readable medium, which the specification limits to media excluding transitory signals (Spec. ¶ [00158]). Therefore, claims 1, 14, and 27 each fall under at least one of the four categories of statutory subject matter: claim 1 as a process, claim 14 as a machine, and claim 27 as a manufacture.
Step 2A Prong 1 – whether the claim recites a judicial exception. See MPEP 2106.04, subsection II.
Regarding independent claim 1, the claim is directed to a method of analyzing information, generating recommendations, and generating control instructions.
The limitations of "analyzing information to be processed, wherein analyzing the information comprises classifying invalid data contained in the information and substituting replacement data in place of at least some of the invalid data in the information" recite an abstract idea because they describe reviewing data, judging which values are invalid, and replacing those values with other values. A person can practically inspect a set of sensor readings, label readings that are missing or out of range as invalid, and replace them with, for example, an average or most-frequent value (see Spec. ¶ [0053]) using pen and paper. These limitations fall within the mental processes grouping of abstract ideas, i.e., concepts performed in the human mind, including observation, evaluation, judgment, and opinion. See MPEP § 2106.04(a)(2), subsection III. To the extent the replacement values are computed (e.g., averaging), the limitations also fall within the mathematical concepts grouping. See MPEP § 2106.04(a)(2), subsection I.
The limitation of "generate one or more recommendations based on the analyzed information," as recited in "the at least one trained machine learning model used to generate one or more recommendations based on the analyzed information," recites an abstract idea because it describes evaluating information and forming a recommendation, which is a judgment. This limitation falls within the mental processes grouping of abstract ideas. See MPEP § 2106.04(a)(2), subsection III.
The limitation of "translating each of the one or more recommendations into one or more actions" recites an abstract idea because it describes converting a recommendation into actions to take, e.g., a plant operator determining which setpoint changes accomplish a recommended operating condition. This is an evaluation and judgment that can practically be performed in the human mind. See MPEP § 2106.04(a)(2), subsection III.
The limitation of "generating one or more control instructions based on the one or more actions for at least one of the one or more recommendations" recites an abstract idea because it describes formulating instructions from determined actions, which a person can practically perform by writing down the instructions with pen and paper. The claim does not require that the instructions be transmitted to or executed by any device. See MPEP § 2106.04(a)(2), subsection III.
Accordingly, independent claim 1 recites an abstract idea under Step 2A Prong 1 because the claim includes limitations falling within the mental processes and mathematical concepts groupings of abstract ideas. Therefore, the analysis should proceed to Step 2A Prong 2.
Independent claim 14 is directed to an apparatus comprising at least one processing device configured to perform operations, and it recites limitations that are the same as those of independent claim 1. As discussed above with respect to independent claim 1, the limitations of classifying invalid data and substituting replacement data, generating recommendations, translating recommendations into actions, and generating control instructions recite abstract ideas, including mental processes and mathematical concepts. See MPEP § 2106.04(a)(2), subsections I and III. Therefore, independent claim 14 recites an abstract idea for the same reasons as independent claim 1.
Independent claim 27 is directed to a non-transitory machine readable medium containing instructions that cause at least one processor to perform operations, and it recites limitations that are the same as those of independent claim 1. As discussed above, these limitations recite abstract ideas, including mental processes and mathematical concepts. See MPEP § 2106.04(a)(2), subsections I and III. Therefore, independent claim 27 recites an abstract idea for the same reasons as independent claim 1.
Step 2A Prong 2 -- whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is "directed to" the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Regarding independent claim 1, this claim recites additional elements of:
"training at least one machine learning model based on some of the analyzed information,"
"providing other of the analyzed information to the at least one trained machine learning model," and
"the at least one trained machine learning model used to generate one or more recommendations."
The limitation "training at least one machine learning model" is recited at a high level of generality. It does not specify any particular model architecture, training algorithm, or manner of training, and amounts to no more than mere instructions to apply the abstract idea using a generic machine learning model as a tool. See MPEP § 2106.05(f). The use of a generically recited trained machine learning model to generate the recommendations likewise amounts to no more than mere instructions to apply the abstract idea on a computer. Applying established machine learning methods to a new data environment, without disclosing an improvement to the machine learning technique itself, does not integrate the exception. See MPEP § 2106.05(f); Recentive Analytics, Inc. v. Fox Corp., No. 2023-2437 (Fed. Cir. Apr. 18, 2025).
The limitation "providing other of the analyzed information to the at least one trained machine learning model" amounts to no more than mere data gathering and inputting, which is insignificant extra-solution activity. See MPEP § 2106.05(g).
The claim does not recite:
an improvement to the functioning of a computer or to machine learning technology (MPEP § 2106.05(a)),
a particular machine that is integral to the claim (MPEP § 2106.05(b)),
a transformation of a particular article (MPEP § 2106.05(c)), or
any other meaningful limitation beyond generally linking the abstract idea to a computer environment (MPEP § 2106.05(e)).
The generated control instructions are not applied to control any machine or process. The claim merely produces information, which is analogous to collecting information, analyzing it, and presenting the results. See Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350 (Fed. Cir. 2016).
Accordingly, the additional elements, individually and in combination, do not integrate the recited judicial exception into a practical application. Therefore, independent claim 1 is directed to the abstract idea under Step 2A Prong 2.
Regarding independent claim 14, this claim is drawn to an apparatus reciting similar limitations to claim 1 and is rejected under the same rationale. Claim 14 also recites the additional element of:
"at least one processing device configured to."
This limitation amounts to no more than generally linking the use of the judicial exception to a particular technological environment, namely implementation on a computer. See MPEP § 2106.05(h). The recited processing device is a generic computer component that merely acts as a tool on which the abstract idea is performed. See MPEP § 2106.05(f). Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application. Therefore, independent claim 14 is directed to the abstract idea under Step 2A Prong 2.
Regarding independent claim 27, this claim is drawn to a non-transitory machine readable medium reciting similar limitations to claim 1 and is rejected under the same rationale. Claim 27 also recites the additional elements of:
"A non-transitory machine readable medium containing instructions that when executed cause at least one processor to."
These limitations amount to no more than generally linking the use of the judicial exception to a particular technological environment, namely implementation on a computer. See MPEP § 2106.05(h). The recited medium and processor are generic computer components that merely act as tools on which the abstract idea is performed. See MPEP § 2106.05(f). Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application. Therefore, independent claim 27 is directed to the abstract idea under Step 2A Prong 2.
Step 2B -- whether the claim amounts to significantly more than the judicial exception. See MPEP § 2106.05.
Regarding independent claim 1, the claim recites additional elements of:
training at least one machine learning model,
providing information to the at least one trained machine learning model, and
using the at least one trained machine learning model to generate recommendations.
These additional elements, individually and in combination, do not amount to significantly more than the judicial exception.
The generically recited training and use of a machine learning model amount to no more than mere instructions to apply the abstract idea using a computer, which cannot provide an inventive concept. See MPEP § 2106.05(f).
Providing data to the model is receiving and transmitting data, which is well-understood, routine, and conventional activity. See MPEP § 2106.05(d), subsection II (citing OIP Techs. and buySAFE, receiving or transmitting data over a network).
Accordingly, the claim does not include any additional elements that amount to significantly more than the judicial exception.
Regarding independent claim 14, the claim recites the additional element of:
at least one processing device.
This element is a generic computer component performing generic computer functions, which amounts to no more than mere instructions to apply the abstract idea using a computer. See MPEP § 2106.05(f). The remaining additional elements are the same as those of claim 1 and do not amount to significantly more for the same reasons. Accordingly, the claim does not include any additional elements that amount to significantly more than the judicial exception.
Regarding independent claim 27, the claim recites the additional elements of:
a non-transitory machine readable medium containing instructions, and
at least one processor.
These elements are generic computer components performing generic computer functions, which amount to no more than mere instructions to apply the abstract idea using a computer. See MPEP § 2106.05(f). The remaining additional elements are the same as those of claim 1 and do not amount to significantly more for the same reasons. Accordingly, the claim does not include any additional elements that amount to significantly more than the judicial exception.
Regarding dependent claims 2–10, 12, 13, 15–23, 25, 26, 28–36, 38, and 39
Dependent claims 2–10, 12, 13, 15–23, 25, 26, 28–36, 38, and 39 merely narrow the previously identified abstract idea limitations recited in independent claims 1, 14, and 27. For the reasons described above with respect to the independent claims, the judicial exceptions recited in these dependent claims are not meaningfully integrated into a practical application, nor do they amount to significantly more than the abstract ideas. The additional limitations include:
determining types of invalid data and generating type-based replacement values,
imposing bounds, imputing values, and using proxy variables,
identifying averages or most-frequent values,
performing an optimization,
training on historical data,
comparing model errors to a threshold and selecting a more accurate model,
limiting the field to thermodynamic power generation,
selecting model inputs,
honoring constraints determined from historical data, and
determining unmetered characteristics.
These limitations merely further define the abstract idea using mathematical relationships and evaluative processes, or limit it to a particular field of use. Accordingly, these dependent claims also recite abstract ideas that do not integrate into a practical application and do not amount to significantly more than the judicial exception, and are rejected under 35 U.S.C. § 101.
Step 1 -- whether the claim falls within any statutory category. See MPEP § 2106.03.
Dependent claims 2–10, 12, and 13 are drawn to a method according to claim 1 (a process). Claims 15–23, 25, and 26 are drawn to an apparatus according to claim 14 (a machine). Claims 28–36, 38, and 39 are drawn to a non-transitory machine readable medium according to claim 27 (a manufacture). Therefore, each of these claims falls under at least one of the four categories of statutory subject matter.
Step 2A Prong 1 – whether the claim recites a judicial exception. See MPEP 2106.04, subsection II.
Regarding claims 2, 15, and 28, these claims recite the limitations of "determining at least one type of invalid data contained in the information" and "generating the replacement data based on the at least one type of invalid data contained in the information."
These limitations are directed toward the abstract idea of a mental process, because recognizing whether an invalid value is, e.g., missing, an outlier, or out of range, and choosing an appropriate replacement for that type of error, are observations, evaluations, and judgments that can practically be performed in the human mind or with pen and paper. To the extent the replacement values are computed, these limitations also involve mathematical concepts. See MPEP § 2106.04(a)(2), subsections I and III.
Regarding claims 3, 16, and 29, these claims recite the limitations of imposing "bounding conditions on each variable associated with the invalid data," imputing "at least one value for each variable associated with the invalid data to generate at least one first replacement value," identifying "at least one proxy variable for each variable associated with the invalid data," and deploying "the at least one proxy variable for use in generating at least one second replacement value," "wherein the replacement data comprises the first and second replacement values."
These limitations are directed toward the abstract ideas of mental processes, because setting minimum and maximum limits on a variable and identifying a related variable that tracks it are evaluations and judgments, and of mathematical concepts, because imputing values and computing a replacement value from a correlated proxy variable are mathematical calculations and relationships. See MPEP § 2106.04(a)(2), subsections I and III.
Regarding claims 4, 17, and 30, these claims recite the limitation of "identifying an average or a most-frequent value for each variable associated with the invalid data."
This limitation is directed toward the abstract idea of mathematical concepts, because computing an average or a most-frequent value (mode) is a mathematical calculation. See MPEP § 2106.04(a)(2), subsection I.
Regarding claims 5, 18, and 31, these claims recite the limitation of "performing optimization using outputs generated by the at least one trained machine learning model, the one or more recommendations based on the optimization."
This limitation is directed toward the abstract idea of mathematical concepts, because optimization is a mathematical calculation that selects values maximizing or minimizing an objective subject to constraints (see Spec. ¶¶ [00100]–[00103], Eq. (1)). See MPEP § 2106.04(a)(2), subsection I.
Regarding claims 6, 19, and 32, these claims recite the limitation of "wherein the analyzed information used to train the at least one machine learning model comprises historical data included in the analyzed information."
This limitation merely specifies the type of data used, which narrows the abstract idea by identifying the information evaluated. See MPEP § 2106.04(a)(2), subsection III.
Regarding claims 7, 20, and 33, these claims recite the limitations of "determining that a first of the multiple trained machine learning models generates predictions having an error exceeding a threshold" and "determining whether a second of the multiple trained machine learning models generates more-accurate predictions."
These limitations are directed toward the abstract ideas of mathematical concepts, because computing a prediction error and comparing it to a threshold are mathematical calculations and relationships, and of mental processes, because comparing the accuracy of two models and choosing the better one is an evaluation and judgment. See MPEP § 2106.04(a)(2), subsections I and III.
Regarding claims 8, 21, and 34, these claims recite the limitations that "the one or more recommendations are associated with a thermodynamic power generation system, the thermodynamic power generation system comprising multiple thermodynamic power generators" and that "the one or more actions associated with each of the one or more recommendations comprises one or more actions that modify operation of the thermodynamic power generation system or one or more of the thermodynamic power generators."
These limitations merely specify the subject of the recommendations and actions that are determined mentally, and therefore narrow the mental processes recited in claim 1. See MPEP § 2106.04(a)(2), subsection III.
Regarding claims 9, 22, and 35, these claims recite the limitation of "wherein the at least one trained machine learning model is configured to use as inputs only one or both of: (i) input flows of material to the thermodynamic power generators and (ii) output flows of material from the thermodynamic power generators."
This limitation merely selects the type of information evaluated, which is an evaluation and judgment regarding which data to consider. See MPEP § 2106.04(a)(2), subsection III.
Regarding claims 10, 23, and 36, these claims recite the limitations that "the multiple thermodynamic power generators comprise multiple turbogenerators configured to generate electrical energy based on flows of steam in the thermodynamic power generation system" and that "the one or more actions associated with each of the one or more recommendations comprises one or more actions that control the flows of steam in the thermodynamic power generation system."
These limitations merely specify the subject of the actions that are determined mentally, e.g., deciding to open or close a steam valve, and therefore narrow the mental processes recited in claim 1. See MPEP § 2106.04(a)(2), subsection III.
Regarding claims 12, 25, and 38, these claims recite the limitations of generating the recommendations "while honoring constraints on a system being controlled" and determining "one or more of the constraints" "based on historical data associated with the system."
These limitations are directed toward the abstract ideas of mathematical concepts, because constrained optimization and deriving bounds from historical data (e.g., percentiles; Spec. ¶ [00108]) are mathematical calculations, and of mental processes, because determining operating limits from past data is an evaluation. See MPEP § 2106.04(a)(2), subsections I and III.
Regarding claims 13, 26, and 39, these claims recite the limitations of generating the recommendations "based on one or more unmetered characteristics associated with a system being controlled" and each unmetered characteristic being "(i) determined based on the analyzed information or (ii) treated as an optimization variable that fluctuates based on at least one other variable."
These limitations are directed toward the abstract idea of mathematical concepts, because estimating an unmeasured value from other values (e.g., by mass balance; Spec. ¶ [00112]) and treating a quantity as an optimization variable are mathematical calculations and relationships. See MPEP § 2106.04(a)(2), subsection I.
Conclusion for Step 2A Prong 1
The dependent claims further narrow the abstract ideas recited in independent claims 1, 14, and 27. Accordingly, they recite judicial exceptions, namely mathematical concepts and mental processes, under MPEP § 2106.04(a)(2).
Step 2A Prong 1 result for claims 2–10, 12, 13, 15–23, 25, 26, 28–36, 38, and 39: YES -- the claims recite a judicial exception.
Step 2A Prong 2 -- whether the claim as a whole integrates the recited judicial exception into a practical application of the exception or whether the claim is "directed to" the judicial exception. This evaluation is performed by (1) identifying whether there are any additional elements recited in the claim beyond the judicial exception, and (2) evaluating those additional elements individually and in combination to determine whether the claim as a whole integrates the exception into a practical application. See MPEP 2106.04(d).
Regarding claims 2–4, 15–17, and 28–30, these claims recite no additional elements beyond those of the independent claims. The recited operations on invalid data are themselves part of the abstract idea.
Regarding claims 7, 20, and 33, these claims recite the additional elements of "multiple trained machine learning models" and "using the second of the multiple trained machine learning models to generate the one or more recommendations." These limitations amount to no more than mere instructions to apply the abstract idea using generic machine learning models as tools. See MPEP § 2106.05(f).
Regarding claims 6, 9, 19, 22, 32, and 35, the limitations specifying historical training data and flow-only model inputs amount to no more than selecting the type of data to be gathered and manipulated, which is insignificant extra-solution activity. See MPEP § 2106.05(g).
Regarding claims 8, 10, 21, 23, 34, and 36, the recited thermodynamic power generation system, thermodynamic power generators, turbogenerators, and flows of steam amount to no more than generally linking the use of the judicial exception to a particular technological environment or field of use. See MPEP § 2106.05(h). These claims do not require that any action or control instruction be applied to operate the generators or to control the steam flows.
Regarding claims 15–23, 25, and 26, these claims further recite "the at least one processing device," and claims 28–36, 38, and 39 further recite a non-transitory machine readable medium containing "instructions that when executed cause the at least one processor." These are generic computer components that merely act as tools on which the abstract idea is performed. See MPEP §§ 2106.05(f), 2106.05(h).
The claims do not recite an improvement to the functioning of a computer or other technology (MPEP § 2106.05(a)), a particular machine integral to the claim (MPEP § 2106.05(b)), a transformation of an article (MPEP § 2106.05(c)), or any other meaningful limitation (MPEP § 2106.05(e)).
Accordingly, the additional elements, individually and in combination, do not integrate the judicial exception into a practical application.
Step 2B -- whether the claim amounts to significantly more than the judicial exception. See MPEP § 2106.05.
Regarding dependent claims 2–10, 12, 13, 15–23, 25, 26, 28–36, 38, and 39, the additional elements identified in Step 2A Prong 2 do not amount to significantly more than the judicial exception, individually or in combination, for the following reasons:
The generically recited machine learning models, processing device, processor, and medium amount to no more than mere instructions to apply the abstract idea using a computer. See MPEP § 2106.05(f).
The selection of historical training data and flow-only inputs is data gathering. Receiving and processing data is well-understood, routine, and conventional activity. See MPEP § 2106.05(d), subsection II.
The limitation to thermodynamic power generation, turbogenerators, and steam flows is a field-of-use limitation that cannot provide an inventive concept. See MPEP § 2106.05(h).
Accordingly, these claims do not include any additional elements that amount to significantly more than the judicial exception, and claims 2–10, 12, 13, 15–23, 25, 26, 28–36, 38, and 39 are rejected under 35 U.S.C. § 101.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
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 text of those sections of Title 35, U.S. Code not included in this action can be found in a prior Office action.
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 1–6, 8, 10–19, 21, 23–32, 34, and 36–39 are rejected under 35 U.S.C. 103 as being unpatentable over Chan et al. (Chan), US 2020/0387818 A1, in view of Al-Owaidh et al. (Al-Owaidh), US 11,886,157 B2, and further in view of Agarwal et al. (Agarwal), WO 2020/255173 A2.
Regarding claim 1, Chan teaches:
"A method comprising:" (Chan, ¶ [0005]: "One example embodiment is a method of building and deploying a model to optimize assets in an industrial process.")
"analyzing information to be processed," (Chan, ¶ [0094]: the method "screens the process variables' continuous measurements, and identifies measurement data (partial and whole) that is of bad quality (invalid) for modeling and predicting one or more process properties associated with the subject plant process." Chan screens the loaded process measurements, i.e., analyzes information to be processed.)
"and substituting replacement data in place of at least some of the invalid data in the information;" (Chan, ¶ [0096]: "In cases of outliers, gaps, and other bad data segments in the measurement data for a candidate process variable, the method 100, at step 120-2, may apply model based data repair to replace these bad data segments with internal model-produced measurement estimation values." The model-produced estimation values are replacement data substituted for the invalid data.)
"training at least one machine learning model based on some of the analyzed information;" (Chan, ¶ [0155]: "The model training module 430, using the reduced dataset or determined latent variables, performs functions for training the PSE models for process online optimization"; ¶ [0012]: "the process model can be trained as, for example, a clustering model, a classification model, a dimension-reduction model, or a deep learning neural network model." The model is trained on the cleansed and reduced historical dataset, i.e., some of the analyzed information.)
"providing other of the analyzed information to the at least one trained machine learning model," (Chan, ¶ [0121]: the method "receives real-time input measurements from plant sensors, online analyzer readings, and lab sample analysis results, and the like, ... and, together, are fed into the process model deployed online"; ¶ [0072]: "(c) Apply data imputation or sensor re-construction algorithms to repair missing and invalid data when necessary." The repaired real-time measurements are analyzed information other than the historical training data, and they are fed to the deployed trained model.)
"the at least one trained machine learning model used to generate one or more recommendations based on the analyzed information;" (Chan, ¶ [0040]: "a process optimizer can compare various conditions and calculate a set of optimal operation setpoints to, for example, maximize profits and/or minimize costs of the asset"; ¶ [0072]: "(e) Execute model predictions and solve optimization problems online ... (f) Export model prediction and solution results for decision making or real-time process control and optimization implementation." The optimal operation setpoints, generated using the deployed model, are recommendations.)
"and generating one or more control instructions based on the one or more actions for at least one of the one or more recommendations." (Chan, ¶ [0157]: "The model execution module 440 may also automatically provide input (adjust parameters/variables/constraints) to the DCS 404 ... The Instrumentation, Control, Operation Computer 405, based on the input, may then automatically adjust or program (via network 408) physical valves, actuators, heaters, and the like 409A-409I ... to execute the calculated PSE solution in the plant process." The input provided to the DCS is the control instructions for executing the calculated solution.)
Chan teaches identifying measurement data of bad quality (missing values, frozen signals, outliers, out-of-limit values, and noisy values) and marking it for possible exclusion (Chan, ¶ [0094]). However, Chan does not teach "wherein analyzing the information comprises classifying invalid data contained in the information."
In the same field of endeavor, Al-Owaidh teaches "wherein analyzing the information comprises classifying invalid data contained in the information" (Al-Owaidh, col. 7, ll. 52–67: "The computer system 104 can implement step 418 in the second level to check if the received steam flow value is within a threshold range (i.e., maximum and minimum limitations)," and a correlation provides "a more accurate steam consumption value for a steam flow rate that has a faulty reading"; col. 8, ll. 51–55: "The computer system 104 outputs the steam value with the classification (class A— trusted value that does not require an adjustment; class B— good value that requires minimum adjustment; class C— default value that requires some adjustment)."). Al-Owaidh classifies each out-of-range, faulty (invalid) reading into Class B or Class C according to the replacement value used (col. 8, ll. 12–36).
Chan and Al-Owaidh are analogous to the claimed invention as all are from the same field of endeavor of validating industrial process data for model-based optimization of plant operation. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the invalid-data identification and repair of Chan with the classification of invalid measurements of Al-Owaidh. The motivation to combine is to address "the accuracy of real-time data due to missing instrumentation or uncalibrated measurements" (Al-Owaidh, col. 3, ll. 60–63) by distinguishing trusted measurements from replaced ones, so that each value receives the appropriate level of adjustment (Al-Owaidh, col. 8, ll. 51–55).
The combination of Chan and Al-Owaidh, however, does not teach "translating each of the one or more recommendations into one or more actions;"
In the same field of endeavor, Agarwal teaches "translating each of the one or more recommendations into one or more actions;" (Agarwal, ¶ [041], p. 11: "The recommendations from the real-time process optimization system include optimal settings of MVs such as percentage opening of fuel control valves, percentage opening or angle of inlet guide vanes, turbine cooling water flow rate, proportion of mixing of various fuels and percentage opening of steam control valves"; ¶ [055], p. 16: the recommendation module 420 is configured "to recommend final value of MV which should be passed from the real time process optimization module 110 to the CCGT plant 102."). Agarwal expresses each recommendation as manipulated-variable settings (valve-opening percentages, guide-vane angles, flow rates), i.e., actions, that are passed to the plant.
Chan, Al-Owaidh, and Agarwal are analogous to the claimed invention as all are from the same field of endeavor of data-driven optimization of industrial plant operation using validated sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the optimal operation setpoints of Chan, as modified by Al-Owaidh, with the translation of recommendations into manipulated-variable settings of Agarwal. The motivation to combine is to implement the optimizer's recommendations as concrete settings of plant equipment "in order to improve the key performance parameters" of the plant (Agarwal, ¶ [041], p. 11).
Regarding claim 2, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"wherein: classifying the invalid data contained in the information comprises determining at least one type of invalid data contained in the information; and" (Chan, ¶ [0094]: "The identified bad quality measurement data for a candidate process variable includes, but are not limited to, missing values (gaps), frozen signals (constant values crossing over the whole history), short-term outliers, and values are out of process in high/low process limits or highly noisy in the continuous measurements of the candidate process variable"; ¶ [0009]: "The measurements that are invalid in quality can include any of: missing values, frozen signals, outlier values, values out of process in high and low limits, and extremely high noisy values." Chan determines whether each invalid measurement is a missing value, a frozen signal, an outlier, an out-of-limit value, or a noisy value, i.e., determines at least one type of invalid data contained in the information.)
"substituting the replacement data in place of at least some of the invalid data in the information comprises generating the replacement data based on the at least one type of invalid data contained in the information." (Chan, ¶ [0096]: "In cases of missing measurement values (gaps) for a candidate process variable, the method 100, at step 120-2, may fill-in the gaps in the continuous measurement data with interpolation. In cases of outliers, gaps, and other bad data segments in the measurement data for a candidate process variable, the method 100, at step 120-2, may apply model based data repair to replace these bad data segments with internal model-produced measurement estimation values"; ¶ [0010]: repairing includes "filling in missing values using interpolation, applying none-phase-shift filters to de-trend drifting and filter noisy values, replacing values with model-produced values." Chan generates interpolated values for gaps, filtered values for noisy data, and model-produced values for outliers and other bad segments, i.e., generates the replacement data based on the type of invalid data.)
Regarding claim 3, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"wherein substituting the replacement data in place of at least some of the invalid data in the information comprises:" (Chan, ¶ [0096]: "the method 100, at step 120-2, may repair some or all of the marked bad quality measurement data for the candidate process variables in the dataset." Chan's repair of the marked invalid data is the substituting of replacement data.)
"imputing at least one value for each variable associated with the invalid data to generate at least one first replacement value;" (Chan, ¶ [0072]: "(c) Apply data imputation or sensor re-construction algorithms to repair missing and invalid data when necessary"; ¶ [0096]: "In cases of missing measurement values (gaps) for a candidate process variable, the method 100, at step 120-2, may fill-in the gaps in the continuous measurement data with interpolation." The interpolated value imputed for each candidate process variable having invalid data is a first replacement value.)
Chan teaches identifying measurement values that are "out of process in high/low process limits" and reconstructing sensor values (Chan, ¶¶ [0094], [0096]). However, Chan does not teach the following limitations. In the same field of endeavor, Al-Owaidh teaches:
"imposing bounding conditions on each variable associated with the invalid data;" (Al-Owaidh, col. 7, ll. 52–55: "The computer system 104 can implement step 418 in the second level to check if the received steam flow value is within a threshold range (i.e., maximum and minimum limitations)"; col. 8, ll. 12–15: the computer system 104 "checks if the value is within a threshold range (i.e., maximum and minimum limitations)." Al-Owaidh imposes maximum and minimum limits on each measured variable, and values outside those limits are treated as faulty, i.e., invalid.)
"identifying at least one proxy variable for each variable associated with the invalid data; and" (Al-Owaidh, col. 7, ll. 55–59: "If the computer system 104 determines that the value is not within the threshold range (decision branch 'NO'), then the computer system 104 implements a user correlation to quantify steam flow versus as a function of process area feed rate as applicable"; col. 7, ll. 64–66: "The correlations are developed based on historical trends for key parameters that have direct impact on the steam consumption." For a steam flow variable with a faulty reading, Al-Owaidh identifies a correlated key parameter, e.g., the process area feed rate, i.e., a proxy variable for the variable associated with the invalid data.)
"deploying the at least one proxy variable for use in generating at least one second replacement value; and" (Al-Owaidh, col. 7, ll. 59–62: "The correlation equation is an alternative way to provide a more accurate steam consumption value for a steam flow rate that has a faulty reading"; col. 8, ll. 41–46: "If not, then the computer system 104 implements correlation operations to determine a closest approximation of what the sensor value would have been and uses that correlated value in the mass balance and energy balance determinations." The correlated value computed from the proxy variable is a second replacement value used in place of the faulty reading.)
"wherein the replacement data comprises the first and second replacement values." (Al-Owaidh, col. 8, ll. 26–30 and ll. 51–55: the correlated steam flow value is used and output "with the classification (class A— trusted value that does not require an adjustment; class B— good value that requires minimum adjustment; class C— default value that requires some adjustment)." In the combination, the replacement data comprises the imputed first replacement values of Chan and the proxy-correlated second replacement values of Al-Owaidh.)
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of validating industrial process data for model-based optimization of plant operation. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the data imputation of Chan with the bounding conditions and proxy-variable correlation of Al-Owaidh, such that invalid data is bounded, replaced by imputed values, and replaced by values generated from proxy variables. The motivation to combine is that "The correlation equation is an alternative way to provide a more accurate steam consumption value for a steam flow rate that has a faulty reading" (Al-Owaidh, col. 7, ll. 59–62), which supplements Chan's imputation where a correlated variable gives a better estimate than interpolation.
Regarding claim 4, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 3," and the limitations of claim 3 are rejected under the same rationale as claim 3.
Chan teaches imputing missing values by interpolation and computing calculated averages when resampling data (Chan, ¶¶ [0010], [0096]). However, Chan does not teach the following limitation. In the same field of endeavor, Al-Owaidh teaches:
"wherein imputing the at least one value for each variable associated with the invalid data comprises identifying an average or a most-frequent value for each variable associated with the invalid data." (Al-Owaidh, col. 8, ll. 15–19: if the value "is not within the threshold range (decision branch 'NO'), then, at step 430, the computer system 104 uses steam flow design value or a user defined default value"; col. 8, ll. 19–23: "The value in Class C, is a given value (i.e. default, suggested by an experience user for a common average value), where, design value is the value mentioned in the design document for this specific user." For a variable whose reading is invalid, Al-Owaidh identifies a common average value for that variable and uses it as the replacement value, i.e., identifies an average value for each variable associated with the invalid data.)
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of validating industrial process data for model-based optimization of plant operation. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the data imputation of Chan with the common-average default value of Al-Owaidh, such that imputing a value for each variable associated with the invalid data comprises identifying an average value for that variable. The motivation to combine is that Al-Owaidh's default average value supplies a replacement where neither the measured nor a correlated value is within limits, so that "there is always a reliable value representing the specific operational physical parameter" (Al-Owaidh, col. 9, ll. 51–54).
Regarding claim 5, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"further comprising: performing optimization using outputs generated by the at least one trained machine learning model," (Chan, ¶ [0156]: "Using the deployed PSE models, the model execution module 440 may perform process optimization online for a plant process"; ¶ [0072]: "(e) Execute model predictions and solve optimization problems online for the ultimate application solutions at a repeated cycle"; ¶ [0040]: "These online calculations are performed based on a process model and an online solver to solve an optimization problem." Chan solves the online optimization problem using the predictions produced by the deployed trained process model, i.e., outputs generated by the at least one trained machine learning model.)
"the one or more recommendations based on the optimization." (Chan, ¶ [0040]: "a process optimizer can compare various conditions and calculate a set of optimal operation setpoints to, for example, maximize profits and/or minimize costs of the asset"; ¶ [0072]: "(f) Export model prediction and solution results for decision making or real-time process control and optimization implementation." The optimal operation setpoints and exported solution results are recommendations produced by the optimization.)
Regarding claim 6, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"wherein the analyzed information used to train the at least one machine learning model comprises historical data included in the analyzed information" (Chan, ¶ [0081]: "various process data are obtained 120 by importing P & ID plant design data, and loading plant historical operating data. An improved dataset is generated by aggregation, data cleansing, and pre-processing 120"; ¶ [0151]: "The data importer/exporter generates a dataset from the loaded historical measurements of the selected process variable candidates"; ¶ [0155]: "The model training module 430, using the reduced dataset or determined latent variables, performs functions for training the PSE models for process online optimization." Chan trains the process model on the dataset generated from loaded historical plant measurements after data cleansing and pre-processing, i.e., on historical data included in the analyzed information.)
Regarding claim 8, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1.
Chan teaches generating optimal operation setpoints to optimize assets in an industrial process (Chan, ¶¶ [0005], [0040]). However, Chan does not teach "wherein: the one or more recommendations are associated with a thermodynamic power generation system, the thermodynamic power generation system comprising multiple thermodynamic power generators; and the one or more actions associated with each of the one or more recommendations comprises one or more actions that modify operation of the thermodynamic power generation system or one or more of the thermodynamic power generators."
In the same field of endeavor, Al-Owaidh teaches:
"wherein: the one or more recommendations are associated with a thermodynamic power generation system, the thermodynamic power generation system comprising multiple thermodynamic power generators; and" (Al-Owaidh, col. 4, ll. 43–47: "FIG. 1 is a schematic representation of an industrial steam power and utility system 100 (sometimes referred to as 'an industrial plant') that operates multiple equipment including, for example, a cogeneration unit 102a, a steam turbine 102b, boilers 102c and others"; col. 11, ll. 15–17: "FIG. 9 is a workflow showing data validation procedures 900 for mechanical drivers and power generators"; col. 12, ll. 60–63: "the computer system 104 implements an optimization layer that provides advisory recommendations to adjust loading of major equipment in the operating facilities." The industrial steam power and utility system is a thermodynamic power generation system, and the cogeneration unit and the steam turbine generator are multiple thermodynamic power generators, each generating electricity from thermal energy. The advisory recommendations are associated with that system.)
"the one or more actions associated with each of the one or more recommendations comprises one or more actions that modify operation of the thermodynamic power generation system or one or more of the thermodynamic power generators." (Al-Owaidh, col. 13, ll. 2–4: "The variables in the optimization include steam and power loading of major equipment in the steam systems"; col. 4, ll. 13–24: the results "are also used to adjust operational physical parameters of the equipment to optimize performance and operation of each equipment and that of the industrial power and utility system," and "can be used to automatically adjust the operational physical parameters (e.g., by transmitting computer instructions to the equipment) without human intervention." Adjusting the steam and power loading and the operational physical parameters of the cogeneration unit and steam turbine are actions that modify the operation of the power generation system and of its generators.)
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of data-driven optimization of industrial plant operation using validated sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the model-based generation of optimal operation setpoints of Chan with the thermodynamic power generation system of Al-Owaidh, such that the recommendations of Chan are associated with Al-Owaidh's multiple thermodynamic power generators and the associated actions modify the loading and operation of those generators. The motivation to combine is that "Optimum operation of such industrial steam and power utility systems helps to lower operating cost, increase operational efficiency and lower carbon dioxide emissions" (Al-Owaidh, col. 3, ll. 56–59).
Regarding claim 10, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 8," and the limitations of claim 8 are rejected under the same rationale as claim 8.
Chan teaches providing input to the distributed control system to adjust physical valves and actuators in the plant process (Chan, ¶ [0157]). However, Chan does not teach "wherein: the multiple thermodynamic power generators comprise multiple turbogenerators configured to generate electrical energy based on flows of steam in the thermodynamic power generation system; and the one or more actions associated with each of the one or more recommendations comprises one or more actions that control the flows of steam in the thermodynamic power generation system."
In the same field of endeavor, Agarwal teaches:
"wherein: the multiple thermodynamic power generators comprise multiple turbogenerators configured to generate electrical energy based on flows of steam in the thermodynamic power generation system; and" (Agarwal, ¶ [032], p. 8: "The CCGT plant can constitute of a plurality of gas turbines and steam turbines"; ¶ [033], p. 8: "Steam can exit the HRSG at different pressures and is used to run a series of steam turbines configured for high, medium and low pressures leading to generation of more electricity." The series of high-, medium-, and low-pressure steam turbines driven by the HRSG steam to generate electricity are multiple turbogenerators that generate electrical energy based on flows of steam in the plant.)
"the one or more actions associated with each of the one or more recommendations comprises one or more actions that control the flows of steam in the thermodynamic power generation system." (Agarwal, ¶ [034], p. 9: "The performance of the plant can be modulated by varying manipulated variables (MVs) such as ... steam flow rates used for cooling and heating (varied using steam control valves)"; ¶ [041], p. 11: "The recommendations from the real-time process optimization system include optimal settings of MVs such as ... percentage opening of steam control valves." Setting the percentage opening of the steam control valves is an action, associated with a recommendation, that controls the flows of steam in the plant.)
Chan and Agarwal are analogous to the claimed invention as both are from the same field of endeavor of data-driven optimization of plant operation using sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the model-based recommendations and control of plant valves of Chan with the multiple steam turbogenerators and steam-control-valve actions of Agarwal, such that the actions associated with the recommendations control the flows of steam driving the turbogenerators. The motivation to combine is that "The performance of the plant can be modulated by varying manipulated variables (MVs)," including the steam flow rates varied by steam control valves (Agarwal, ¶ [034], p. 9), "in order to improve the key performance parameters of CCGT" (Agarwal, ¶ [041], p. 11).
Regarding claim 11, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"further comprising: implementing the one or more control instructions to modify operation of at least one component in a system." (Chan, ¶ [0157]: "The Instrumentation, Control, Operation Computer 405, based on the input, may then automatically adjust or program (via network 408) physical valves, actuators, heaters, and the like 409A-409I, or program any other plant or refinery control system or processing system coupled to the DCS system 404, to execute the calculated PSE solution in the plant process." The DCS executes the input received from the model execution module, i.e., implements the control instructions. In doing so it adjusts the physical valves, actuators, and heaters, i.e., modifies the operation of at least one component in the plant system.)
Regarding claim 12, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"wherein: the one or more recommendations are generated while honoring constraints on a system being controlled; and" (Chan, ¶ [0066]: the problem definition includes "(c) Various Constraint Conditions reflecting market and plant operation realities need to be satisfied"; ¶ [0040]: "a process optimizer can compare various conditions and calculate a set of optimal operation setpoints." The optimal operation setpoints, i.e., the recommendations, are calculated subject to constraint conditions reflecting the operation of the plant being controlled.)
Chan teaches constraint conditions reflecting plant operation realities (Chan, ¶ [0066]). However, Chan does not teach "one or more of the constraints are determined based on historical data associated with the system."
In the same field of endeavor, Al-Owaidh teaches "one or more of the constraints are determined based on historical data associated with the system." (Al-Owaidh, col. 13, ll. 5–15: "The optimization is subject to several constraints representing the operational and equipment limitations. Examples include: ... Equipment maximum and minimum output limitations"; col. 12, ll. 55–58: "Equipment steam consumption will be adjusted between minimum and maximum limits, according to specific flow chart for each equipment/user"; col. 10, ll. 6–9: "The power and steam limits where the cogeneration system is considered to be online is typically based on design data or historical data or a combination of both"; col. 12, ll. 3–5: "The power limits where the gas turbine unit is considered to be online is typically based on design data or historical data or a combination of both.") The equipment operating limits that bound Al-Owaidh's optimization are set from historical data of the equipment, i.e., one or more of the constraints are determined based on historical data associated with the system.
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of data-driven optimization of industrial plant operation using validated sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the constraint-bounded optimization of Chan with the historical-data-based equipment limits of Al-Owaidh, such that one or more of the constraints honored by the recommendations are determined from historical data of the controlled system. The motivation to combine is that limits derived from historical operating data provide a "more accurate representation of the equipment and its performance" (Al-Owaidh, col. 7, l. 67 – col. 8, l. 2), so that the recommended setpoints remain within limits the equipment has actually achieved in operation.
Regarding claim 13, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1.
Chan teaches computing "one or more unmeasurable batch properties marked as Simulation Data X' as computed variables" by first-principle simulation of the process (Chan, ¶ [0133]). However, Chan does not teach "wherein: the one or more recommendations are generated based on one or more unmetered characteristics associated with a system being controlled; and each unmetered characteristic is (i) determined based on the analyzed information or (ii) treated as an optimization variable that fluctuates based on at least one other variable."
In the same field of endeavor, Al-Owaidh teaches:
"wherein: the one or more recommendations are generated based on one or more unmetered characteristics associated with a system being controlled; and" (Al-Owaidh, col. 3, ll. 60–67: "One challenge when implementing such advisory systems is the accuracy of real-time data due to missing instrumentation or uncalibrated measurements. This disclosure describes techniques that overcome the challenges by implementing the advisory systems despite the absence of instrumentation"; col. 4, ll. 29–31: "Implementing the techniques provides plant operators with operational set points even if the plant has limitations on key measurements"; col. 12, ll. 60–63: "the computer system 104 implements an optimization layer that provides advisory recommendations to adjust loading of major equipment in the operating facilities." Al-Owaidh generates its advisory recommendations and operational set points using values of characteristics that no instrument measures, i.e., unmetered characteristics of the controlled plant.)
"each unmetered characteristic is (i) determined based on the analyzed information or (ii) treated as an optimization variable that fluctuates based on at least one other variable." (Al-Owaidh, col. 11, ll. 34–38: "If there is no steam reading or the reading is outside the minimum and maximum ranges, then a correlation is used to quantify the expected steam consumption based on process parameters related to the steam turbines or a design value is used"; col. 11, ll. 51–56: "If the motor power consumption PI tag is unavailable, the motor power consumption is determined based on other process parameters for pumps (e.g., pressure differential, flow, temperature differential, specific density and the like)." Each unmetered steam consumption or motor power value is computed from other validated process parameters, i.e., determined based on the analyzed information, satisfying alternative (i).)
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of data-driven optimization of industrial plant operation using validated sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the computation of unmeasurable properties and optimization of Chan with the determination of unmetered equipment characteristics from other process parameters of Al-Owaidh, such that the recommendations are generated based on unmetered characteristics determined from the analyzed information. The motivation to combine is that Al-Owaidh's techniques "negate the need to have a complete set of measurements to close steam's heat and mass balance" (Al-Owaidh, col. 4, ll. 31–33), allowing recommendations to be generated for a plant lacking instrumentation on key parameters.
Claim 14 is rejected under 35 U.S.C. 103. Chan teaches "An apparatus comprising: at least one processing device configured to" (Chan, ¶ [0006]: "The system includes a processor operatively coupled to a data storage system. The processor is configured to implement a data preparation module, a model development module, and an execution module."). The remaining limitations of claim 14 correspond to claim 1 and are rejected based on the same rationale as claim 1.
Claim 15 corresponds to claim 2 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 2.
Claim 16 corresponds to claim 3 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 3.
Claim 17 corresponds to claim 4 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 4.
Claim 18 corresponds to claim 5 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 5.
Claim 19 corresponds to claim 6 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 6.
Claim 21 corresponds to claim 8 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 8.
Claim 23 corresponds to claim 10 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 10.
Claim 24 corresponds to claim 11 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 11.
Claim 25 corresponds to claim 12 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 12.
Claim 26 corresponds to claim 13 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 13.
Regarding claim 27, Chan teaches:
"A non-transitory machine readable medium containing instructions that when executed cause at least one processor to:" (Chan, ¶ [0007]: "Another example embodiment is a non-transitory computer-readable data storage medium comprising instructions causing a computer to (i) generate a dataset by loading a set of process variables of a subject industrial process"; ¶ [0006]: "The system includes a processor operatively coupled to a data storage system." Chan's non-transitory computer-readable data storage medium stores instructions that, when executed, cause the computer's processor to perform the recited operations.)
"analyze information to be processed, wherein the instructions that when executed cause the at least one processor to analyze the information comprise:" (Chan, ¶ [0007]: the instructions cause the computer to "(ii) identify and remove measurements that are invalid in quality for modeling the subject industrial process"; ¶ [0094]: the method "screens the process variables' continuous measurements, and identifies measurement data (partial and whole) that is of bad quality (invalid)." The stored instructions cause the processor to screen the loaded process measurements, i.e., to analyze information to be processed.)
"and substitute replacement data in place of at least some of the invalid data in the information;" (Chan, ¶ [0096]: "In cases of outliers, gaps, and other bad data segments in the measurement data for a candidate process variable, the method 100, at step 120-2, may apply model based data repair to replace these bad data segments with internal model-produced measurement estimation values." The model-produced estimation values are replacement data substituted for the invalid data.)
Chan teaches instructions that identify measurement data of bad quality and mark it for possible exclusion (Chan, ¶¶ [0007], [0094]). However, Chan does not teach "instructions that when executed cause the at least one processor to classify invalid data contained in the information."
In the same field of endeavor, Al-Owaidh teaches "instructions that when executed cause the at least one processor to classify invalid data contained in the information" (Al-Owaidh, col. 7, ll. 52–67: "The computer system 104 can implement step 418 in the second level to check if the received steam flow value is within a threshold range (i.e., maximum and minimum limitations)," and a correlation provides "a more accurate steam consumption value for a steam flow rate that has a faulty reading"; col. 8, ll. 51–55: "The computer system 104 outputs the steam value with the classification (class A— trusted value that does not require an adjustment; class B— good value that requires minimum adjustment; class C— default value that requires some adjustment)"; col. 3, ll. 5–9: "a computer-readable storage medium (e.g., non-transitory computer-readable storage medium) storing computer instructions executable by the one or more computers or computer systems to perform the operations described here."). Al-Owaidh's stored instructions cause the processor to classify each out-of-range, faulty (invalid) reading into Class B or Class C.
Chan and Al-Owaidh are analogous to the claimed invention as both are from the same field of endeavor of validating industrial process data for model-based optimization of plant operation. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the invalid-data identification and repair instructions of Chan with the classification of invalid measurements of Al-Owaidh. The motivation to combine is to address "the accuracy of real-time data due to missing instrumentation or uncalibrated measurements" (Al-Owaidh, col. 3, ll. 60–63) by distinguishing trusted measurements from replaced ones, so that each value receives the appropriate level of adjustment (Al-Owaidh, col. 8, ll. 51–55).
The remaining limitations of claim 27, "train at least one machine learning model based on some of the analyzed information; provide other of the analyzed information to the at least one trained machine learning model and generate one or more recommendations based on the analyzed information; translate each of the one or more recommendations into one or more actions; and generate one or more control instructions based on the one or more actions for at least one of the one or more recommendations," correspond to the limitations of claims 1 and 14. They are rejected under 35 U.S.C. 103 based on the same rationale as claims 1 and 14.
Claim 28 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitations "the instructions that when executed cause the at least one processor to classify the invalid data contained in the information comprise: instructions that when executed cause the at least one processor to" and "the instructions that when executed cause the at least one processor to substitute the replacement data in place of at least some of the invalid data in the information comprise instructions that when executed cause the at least one processor to" are rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 2 and 15 and are rejected based on the same rationale as claims 2 and 15.
Claim 29 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitation "the instructions that when executed cause the at least one processor to substitute the replacement data in place of at least some of the invalid data in the information comprise: instructions that when executed cause the at least one processor to" is rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 3 and 16 and are rejected based on the same rationale as claims 3 and 16.
Claim 30 depends on claim 29 and is rejected under 35 U.S.C. 103. The limitation "the instructions that when executed cause the at least one processor to impute the at least one value for each variable associated with the invalid data comprise: instructions that when executed cause the at least one processor to" is rejected based on the same rationale as claims 27 and 29, which address a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 4 and 17 and are rejected based on the same rationale as claims 4 and 17.
Claim 31 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitation "further containing instructions that when executed cause the at least one processor to" is rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 5 and 18 and are rejected based on the same rationale as claims 5 and 18.
Claim 32 depends on claim 27 and is rejected under 35 U.S.C. 103. The non-transitory machine readable medium is rejected based on the same rationale as claim 27. The remaining limitations correspond to claims 6 and 19 and are rejected based on the same rationale as claims 6 and 19.
Claim 34 depends on claim 27 and is rejected under 35 U.S.C. 103. The non-transitory machine readable medium is rejected based on the same rationale as claim 27. The remaining limitations correspond to claims 8 and 21 and are rejected based on the same rationale as claims 8 and 21.
Claim 36 depends on claim 34 and is rejected under 35 U.S.C. 103. The non-transitory machine readable medium is rejected based on the same rationale as claims 27 and 34. The remaining limitations correspond to claims 10 and 23 and are rejected based on the same rationale as claims 10 and 23.
Claim 37 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitation "further containing instructions that when executed cause the at least one processor to" is rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 11 and 24 and are rejected based on the same rationale as claims 11 and 24.
Claim 38 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitations "the instructions when executed cause the at least one processor to generate" and "the instructions when executed cause the at least one processor to determine" are rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 12 and 25 and are rejected based on the same rationale as claims 12 and 25.
Claim 39 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitations "the instructions when executed cause the at least one processor to generate" and "the instructions when executed cause the at least one processor, for each unmetered characteristic, to" are rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 13 and 26 and are rejected based on the same rationale as claims 13 and 26.
Claims 7, 20, and 33 are rejected under 35 U.S.C. 103 as being unpatentable over Chan et al. (Chan), US 2020/0387818 A1, in view of Al-Owaidh et al. (Al-Owaidh), US 11,886,157 B2, further in view of Agarwal et al. (Agarwal), WO 2020/255173 A2, and further in view of Nistala et al. (Nistala), US 2022/0147672 A1.
Regarding claim 7, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 1," and the limitations of claim 1 are rejected under the same rationale as claim 1. Chan further teaches:
"wherein: the at least one trained machine learning model comprises multiple trained machine learning models; and" (Chan, ¶ [0074]: "the systems and methods disclosed herein may contain multiple models and solutions, and they can be developed and deployed in one or more computer servers and run simultaneously." Chan deploys multiple trained process models concurrently.)
Chan teaches monitoring model performance against a baseline of a model-quality KPI and adapting a model once the KPI becomes poor and a criterion threshold is reached (Chan, ¶¶ [0073], [0075]). However, Chan does not teach "the method further comprises: determining that a first of the multiple trained machine learning models generates predictions having an error exceeding a threshold; determining whether a second of the multiple trained machine learning models generates more-accurate predictions; and in response to determining that the second of the multiple trained machine learning models generates more-accurate predictions, using the second of the multiple trained machine learning models to generate the one or more recommendations."
The combination of Chan, Al-Owaidh, and Agarwal, however, does not teach these limitations.
In the same field of endeavor, Nistala teaches:
"the method further comprises: determining that a first of the multiple trained machine learning models generates predictions having an error exceeding a threshold;" (Nistala, ¶ [0042]: "The predictions from the models are compared with an actual value or the ground truth to compute model quality index (MQI)"; ¶¶ [0043]–[0044]: the MQI comprises performance metrics including "Error metrics (absolute error, absolute percentage error, root mean square error, etc.)"; ¶ [0053]: "If the MQI of the models is below the lower threshold (Th_MQI_lower), the performance of the models is unacceptable." Because a lower MQI reflects a larger prediction error, an MQI of the active model below the threshold corresponds to that model generating predictions whose error exceeds a threshold.)
"determining whether a second of the multiple trained machine learning models generates more-accurate predictions; and" (Nistala, ¶ [0057]: "'k' relevant prediction, detection, classification, diagnosis or prognosis models are selected from the models database"; ¶ [0058]: "For each of the 'k' models, the MQI is computed by comparing the predictions and the actual values or the ground truth of response variables. Each of the 'k' MQI values from the models are compared against the MQI thresholds." Nistala evaluates alternative trained models to determine whether any produces more accurate predictions than the drifted active model.)
"in response to determining that the second of the multiple trained machine learning models generates more-accurate predictions, using the second of the multiple trained machine learning models to generate the one or more recommendations." (Nistala, ¶ [0058]: "For any of the 'k' models, if the MQI is above the thresholds, the at least one model (typically the model with the highest MQI) among those is recommended for activation for subsequent predictions. The previously active model is recommended for deactivation"; ¶ [0004]: "optimization is carried out on models that mimic the behavior of industrial processes to arrive at optimum set points that can be suggested to the operators/plant engineers." The more accurate model replaces the drifted model and is used for subsequent predictions, from which the optimum set points, i.e., the recommendations, are generated.)
Chan, Al-Owaidh, Agarwal, and Nistala are analogous to the claimed invention as all are from the same field of endeavor of data-driven optimization of industrial plant operation using validated sensor data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the multiple deployed models of Chan, as modified by Al-Owaidh and Agarwal, with the model-quality-index threshold monitoring and replacement of a drifted model by a more accurate model of Nistala. The motivation to combine is that "the model predictions may drift leading to drop in performance accuracy of the model," in which case "the models would not be effective for carrying out online monitoring, diagnostics and optimization" (Nistala, ¶ [0005]); substituting a more accurate model maintains the reliability of the recommendations generated by the optimization.
Claim 20 corresponds to claim 7 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 7.
Claim 33 depends on claim 27 and is rejected under 35 U.S.C. 103. The limitation "the non-transitory machine readable medium further contains instructions that when executed cause the at least one processor to" is rejected based on the same rationale as claim 27, which addresses a non-transitory machine readable medium containing instructions that are executed by at least one processor. The remaining limitations correspond to claims 7 and 20 and are rejected based on the same rationale as claims 7 and 20.
Claims 9, 22, and 35 are rejected under 35 U.S.C. 103 as being unpatentable over Chan et al. (Chan), US 2020/0387818 A1, in view of Al-Owaidh et al. (Al-Owaidh), US 11,886,157 B2, further in view of Agarwal et al. (Agarwal), WO 2020/255173 A2, and further in view of Adelman et al. (Adelman), US 6,591,225 B1.
Regarding claim 9, the combination of Chan, Al-Owaidh, and Agarwal teaches "The method of Claim 8," and the limitations of claim 8 are rejected under the same rationale as claim 8.
Chan teaches selecting the inputs of the trained model, in that "the system selects a set of predictors from process variables and physically meaningful features by performing pre-modeling and iterative feature selections" (Chan, ¶ [0055]). However, Chan does not teach "wherein the at least one trained machine learning model is configured to use as inputs only one or both of: (i) input flows of material to the thermodynamic power generators and (ii) output flows of material from the thermodynamic power generators."
In the same field of endeavor, Adelman teaches "wherein the at least one trained machine learning model is configured to use as inputs only one or both of: (i) input flows of material to the thermodynamic power generators and (ii) output flows of material from the thermodynamic power generators." (Adelman, col. 5, ll. 62–67: "The optimization system models the steam turbine power as a piece-wise linear function of inlet throttle steam flow that is linearly dependent upon admission flows, extraction flows, and condenser pressure"; col. 6, ll. 1–37: the steam turbine power equation, with inputs including "HPHEAD_ST_M(I) throttle flow," "IPHEAD_ST_M(I) HP admission flow," "LPHEAD_ST_M(I) LP admission flow," "ST_HPPROC_M(I) HP extraction flow," and "ST_IPHEAD_M(I) IP extraction flow.") Adelman models the power of each steam turbine from the input flows of steam to the turbine (throttle and admission flows) and the output flows of steam from the turbine (extraction flows). Adelman's pressure terms are expressed as deviations from reference values, e.g., "(PCOND(I) − PCOND_REF(I)*STON(I))" (col. 6, ll. 13–14), and these deviations are zero at reference pressure conditions. At those conditions the modeled power depends only on the steam input flows and output flows. Restricting the inputs of Chan's trained model to those flows is a selection of predictors of the type Chan performs (Chan, ¶ [0055]).
Chan and Adelman are analogous to the claimed invention as both are from the same field of endeavor of model-based optimization of plant operation. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to combine the trained model and predictor selection of Chan with the flow-based steam turbine power model of Adelman, such that the trained model uses as inputs the input flows of material to and the output flows of material from the thermodynamic power generators. The motivation to combine is that Adelman's flow-based plant model is used to predict "the operational control set points (i.e., variables) of a combined-cycle power plant to achieve a profitable operating condition" (Adelman, Abstract), and Chan's feature selection retains only the predictors that best explain the modeled output (Chan, ¶ [0055]).
Claim 22 corresponds to claim 9 and is rejected under 35 U.S.C. 103 based on the same rationale as claim 9.
Claim 35 depends on claim 34 and is rejected under 35 U.S.C. 103. The non-transitory machine readable medium is rejected based on the same rationale as claims 27 and 34. The remaining limitations correspond to claims 9 and 22 and are rejected based on the same rationale as claims 9 and 22.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to HUNG VAN LE whose telephone number is (571)270-0164. The examiner can normally be reached 8 a.m. - 5 p.m..
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, Cesar Paula can be reached at (571) 272-4128. 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.
/HUNG VAN LE/Examiner, Art Unit 2145
/CESAR B PAULA/Supervisory Patent Examiner, Art Unit 2145
