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 . This action is made non-final.
This action is in response to the application and claims filed December 1st, 2022. Claims 1-9 are pending in the case and have been examined. Claims 1-9 are rejected.
Priority
Acknowledgment is made of applicant's claim for foreign priority based on an application filed in China on July, 26th 2022. The examiner acknowledges that a certified copy of Chinese Patent Application No. 202210884861.5 has been retrieved (on May 2nd, 2023, in Chinese), as required by 37 CFR 1.55. The examiner notes that a translation of Chinese Patent Application No. 202211188739.0 does not appear to have been furnished to-date.
Although a certified copy of the foreign priority application was retrieved, a translation of said application has not yet been made of record in accordance with 37 CFR 1.55. See MPEP §§ 215 and 216. Applicant is reminded of requirements set forth in 37 CFR 1.55(g)(3)-(4) Claim for foreign priority:
“(3) An English language translation of a non-English language foreign application is not required except:
(i) When the application is involved in an interference (see § 41.202 of this chapter) or derivation (see part 42 of this chapter) proceeding;
(ii) When necessary to overcome the date of a reference relied upon by the examiner; or
(iii) When specifically required by the examiner.
(4) If an English language translation of a non-English language foreign application is required, it must be filed together with a statement that the translation of the certified copy is accurate” (emphasis added).
Since an English language translation of Application No. 202210884861.5 has not been made of record to-date, the Examiner notes that prior art references with a filing date or a publication date prior to the instant Application’s filing date of December 1st, 2022 are considered applicable prior art references.
Information Disclosure Statement
Acknowledgment is made of the information disclosure statement filed December 1st, 2022, which complies with 37 CFR 1.97. As such, the information disclosure statement has been placed in the application file and the information referred to therein has been considered by the examiner.
Claim Objections
Claim 1 is objected to because of the following informalities: The claim introduced “a stacked time series network” in the preamble/lines 1-2 of the claim. As such, the 2nd recitation of “a stacked time series network” in step (4) of the claim should read “the [[a]] stacked time series network”. Appropriate correction is required.
Claims 2-9 are objected to because of the following informalities: Dependent directly or indirectly from claim 1, and are objected to based on their respective dependencies from claim 1. Appropriate correction is required.
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 1 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Claim 1 recites “collecting gas-in-oil information of each substation” in step (1). This recitation is unclear and indefinite. No plurality, group, or set of substations were introduced prior to the recitation of “each substation”. In particular, it is unclear what constitutes 'each substation' for purposes of collecting 'gas-in-oil information', and thus it is unclear whether the gas-in-oil information is collected for each electrical substation in a system, for each substation that includes a transformer, or only for substations associated with the transformer being diagnosed. For the purposes of determining patent eligibility and comparison with the prior art, “collecting gas-in-oil information of each substation” has been interpreted as collecting gas-in-oil information of each of any group, plurality or set of substations. Appropriate correction is required.
Also, claims 2-9 which depend directly or indirectly from claim 1, are rejected under 35 U.S.C. 112(b) as being indefinite under the same rationale as claim 1.
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.
To determine if a claim is directed to patent ineligible subject matter, the Court has guided the Office to apply the Alice/Mayo test, which requires:
Step 1: Determining if the claim falls within a statutory category.
Step 2A: Determining if the claim is directed to a patent ineligible judicial exception consisting of a law of nature, a natural phenomenon, or abstract idea; and Step 2A is a two prong inquiry. MPEP 2106.04(II)(A). Under the first prong, examiners evaluate whether a law of nature, natural phenomenon, or abstract idea is set forth or described in the claim. Abstract ideas include mathematical concepts, certain methods of organizing human activity, and mental processes. MPEP 2104.04(a)(2). The second prong is an inquiry into whether the claim integrates a judicial exception into a practical application. MPEP 2106.04(d).
Step 2B: If the claim is directed to a judicial exception, determining if the claim recites limitations or elements that amount to significantly more than the judicial exception. (See MPEP 2106).
Claims 1-9 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-9 are directed to a method (corresponding to a process), and therefore, are directed to the statutory category of a process.
Regarding claim 1
Step 2A Prong 1
Claim 1 recites the following mathematical concepts, that in each case under the broadest reasonable interpretation (BRI), covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“normalizing the collected gas-in-oil information to obtain a normalized matrix” (e.g., scaling/transforming values into matrix representations)
“dividing the normalized matrix into a training set and a test set in proportion to train network parameters” (e.g., organizing/partitioning numerical data into subsets according to a proportion for model training)
“constructing a stacked time series network based on Xgboost and a bidirectional gated neural network” (e.g., defining/configuring mathematical models boosted decision trees and neural network computations for processing time-series data)
“inputting the training set and the test set to train the stacked time series network, and learning a feature of gas-in-oil data of a transformer” (e.g., optimization/parameter fitting on training data to generate feature representations/weights from input data)
“performing fault diagnosis based on real-time gas-in-oil data during operation, and fine tuning a weight of the stacked time series network to enable the stacked time series network to continuously learn a new feature” (e.g., classification/prediction based on model outputs, statistical inference)1
Claim 1 also recites the following mental processes, that in each case under the BRI, covers performance of the limitation in the mind (including observation, evaluation, judgement, opinion) or with the aid of pencil and paper but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(III)].
“performing fault diagnosis based on real-time gas-in-oil data during operation” (e.g., a human can observe real time statuses of a system and determine if faults occur in real time)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of “collecting gas-in-oil information of each substation2, wherein the gas-in-oil information comprises monitoring information of an oil test and contents of dissolved gas and furan in oil in each substation” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of obtaining data to input for a model, i.e., pre-solution activity of data gathering (e.g., obtaining information for processing in a computer system (see MPEP 2106.05(g)).
Regarding the “stacked time series network” and “bidirectional gated neural network”, no details of the networks are recited, and the networks are recited at a high level of generality and the networks can be constructed by hand with pen and paper. Aside from merely repeating the claim language in paragraphs [0010], and [0016], and providing general examples in paragraphs [0042-0044] and [0049-0053] in stating “ the stacked time series network is constructed in the following manner: first, constructing an Xgboost model, as shown in FIG. 2 ; then constructing the bidirectional gated neural network, as shown in FIG. 3 ; and finally, using a meta learner to learn results of Xgboost and the bidirectional gated neural network” and “A bidirectional gate recurrent unit (GRU) structure can process input data xi=[x1, . . . , xn]T in both forward and backward directions, and then splice two obtained feature vectors together as another expression of an input vector”, applicant’s specification does not further define the “stacked time series network” and “bidirectional gated neural network”. Thus, the claimed “stacked time series network” and “bidirectional gated neural network”, under the BRI, in light of the specification, could be constructed and modified by hand with pen and paper based on observed data (i.e., the “gas-in-oil information” and “the training set and the test set”).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional element of
“collecting gas-in-oil information of each substation, wherein the gas-in-oil information comprises monitoring information of an oil test and contents of dissolved gas and furan in oil in each substation” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of obtaining data to input for a model, i.e., pre-solution activity of data gathering. The courts have found limitations directed to obtaining information electronically, recited at a high-level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B, the additional element individually or in combination does not amount to significantly more than the judicial exception.
Regarding claim 2
Step 2A Prong 1
Claim 2 does not introduce any new abstract ideas, but recites the abstract idea identified in the parent claim(s).
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of “wherein the gas-in-oil information comprises data of the transformer during operation and data recorded by an electric power company, and each group of data comprises gas-in-oil data and a fault state of the corresponding transformer, wherein the gas-in-oil data comprises contents of nine key states: a breakdown voltage (BDV), water, acidity, hydrogen, methane, ethane, ethylene, acetylene, and furan” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of defining what data is to be used in the system, i.e., pre-solution activity of selecting a particular data source of type of data to be manipulated (e.g., obtaining particular types of information for processing in a computer system (see MPEP 2106.05(g)).
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional element of
“wherein the gas-in-oil information comprises data of the transformer during operation and data recorded by an electric power company, and each group of data comprises gas-in-oil data and a fault state of the corresponding transformer, wherein the gas-in-oil data comprises contents of nine key states: a breakdown voltage (BDV), water, acidity, hydrogen, methane, ethane, ethylene, acetylene, and furan” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of obtaining specific types of data to input for a model, i.e., pre-solution activity of selecting a particular data source of type of data to be manipulated. The courts have found limitations directed to obtaining information electronically, recited at a high-level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B, the additional element individually or in combination does not amount to significantly more than the judicial exception.
Regarding claim 3
Step 2A Prong 1
Claim 3 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”,) [see MPEP 2106.04(a)(2)(I)].
“performing z-score normalization on the gas-in-oil information to obtain the normalized matrix” (e.g., standardizing numerical values by subtracting a mean and dividing by a standard deviation to produced normalized data/matrix values)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
In accordance with Step 2A, Prong 2, the claim does not include any additional elements and the judicial exception is not integrated into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding claim 4
Step 2A Prong 1
Claim 4 does not introduce any new abstract ideas, but recites the abstract idea identified in the parent claim(s).
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of “wherein the data in the gas-in-oil information is divided into two parts in step (3), wherein data of a certain proportion is used as the training set to train the stacked time series network, and a data of a remaining proportion is used as the test set to test a fault diagnosis effect of the stacked time series network for the transformer” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of defining what data is to be used in the system, i.e., pre-solution activity of selecting a particular data source of type of data to be manipulated (e.g., obtaining particular types of information for processing in a computer system (see MPEP 2106.05(g)). The examiner notes that this could also be interpreted under Step 2A Prong 1 under 2106.04(a)(2)(I) as a mathematical concept, as it is still partitioning a dataset by proportion into separate training and testing subsets to evaluate a model using the test subset.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional element of
“wherein the data in the gas-in-oil information is divided into two parts in step (3), wherein data of a certain proportion is used as the training set to train the stacked time series network, and a data of a remaining proportion is used as the test set to test a fault diagnosis effect of the stacked time series network for the transformer” limitation, the additional element is recited at a high-level of generality and amounts to extra-solution activity of obtaining specific types of data to input for a model, i.e., pre-solution activity of selecting a particular data source of type of data to be manipulated. The courts have found limitations directed to obtaining information electronically, recited at a high-level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Accordingly, at Step 2B, the additional element individually or in combination does not amount to significantly more than the judicial exception.
Regarding claim 5
Step 2A Prong 1
Claim 5 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“constructing the stacked time series network based on Xgboost and the bidirectional gated neural network to perform feature extraction and prediction on the gas-in-oil information, wherein construction of Xgboost comprises establishment of an integrated model, selection of an objective function, and solving of a loss function and the bidirectional gated neural network comprises a forward calculation layer, a backward calculation layer, an update gate, and a reset gate” (e.g., defining and training mathematical machine-learning models by selecting an objective function and optimizing/solving a loss function to compute model parameters for feature extraction and prediction)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
The judicial exception is not integrated into a practical application. In particular, the claim recites the additional elements of “predicting gas in the oil by using Xgboost and the bidirectional gated neural network, and outputting a prediction result of the gas-in-oil information”, and limitations, these additional elements are recited at a high-level of generality and amounts to extra-solution activity of inputting data into a model, to gather an output for additional training of a model, i.e., post-solution activity of outputting data (e.g., obtaining particular types of information for processing in a computer system to then train a model and output data(see MPEP 2106.05(g)).
Regarding the “training prediction results of Xgboost and the bidirectional gated neural network by using a meta learner, to output the prediction result of the gas-in-oil information, performing fault diagnosis on the stacked time series network by using a Softmax layer, and outputting the fault state of the transformer” limitation, which is recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using generic computer components (See MPEP 2106.05(f)). In particular, they are merely inputting data into different models to identify desired data.
Accordingly, at Step 2A, prong two, the additional elements individually or in combination do not integrate the judicial exception into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more that the judicial exception. As discussed above, the additional element of
“predicting gas in the oil by using Xgboost and the bidirectional gated neural network, and outputting a prediction result of the gas-in-oil information”, and limitations, these additional elements are recited at a high-level of generality and amounts to extra-solution activity of outputting results, i.e., post-solution activity of data outputting. The courts have found limitations directed to obtaining information electronically, recited at a high-level of generality, to be well-understood, routine, and conventional (see MPEP 2106.05(d)(II), “receiving or transmitting data over a network”, "electronic record keeping," and "storing and retrieving information in memory").
Regarding the “training prediction results of Xgboost and the bidirectional gated neural network by using a meta learner, to output the prediction result of the gas-in-oil information, performing fault diagnosis on the stacked time series network by using a Softmax layer, and outputting the fault state of the transformer” limitation, which is recited at a high-level of generality such that it amounts to no more than mere instructions to apply the exception using generic computer components (See MPEP 2106.05(f)). In particular, they are merely inputting data into different models to identify desired data.
Accordingly, at Step 2B, the additional element individually or in combination does not amount to significantly more than the judicial exception.
Regarding claim 6
Step 2A Prong 1
Claim 6 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“wherein the construction of Xgboost comprises the establishment of the integrated model, the selection of the objective function, and the solving of the loss function, wherein the establishment of the integrated model is to recursively construct a binary decision tree, and in input space of the training set, each region is recursively divided into two sub-regions based on a minimum squared-error criterion, and an output value of each sub-region is determined” (e.g., recursively partitioning an input space using a squared-error criterion to construct a decision tree and compute region output values)
“the selection of the objective function is to measure an error between a predicted value and a real value of a target, and the objective function is approximated through second-order Taylor expansion” (e.g., defining an objective/error function and using second-order Taylor expansion to approximate if for optimization)
“the solving of the loss function is to partition a sub-tree by using a greedy algorithm, enumerate feasible partitioning points, in other words, add a new partition to an existing leaf each time, and calculate a corresponding maximum gain” (e.g., optimizing a loss function by greedy search over candidate split points and selecting splits that maximize gain)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
In accordance with Step 2A, Prong 2, the claim does not include any additional elements and the judicial exception is not integrated into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding claim 7
Step 2A Prong 1
Claim 7 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“wherein the bidirectional gated neural network comprises the forward calculation layer, the backward calculation layer, the update gate, and the reset gate, wherein the reset gate helps to capture a short-term dependency in a time series, the update gate helps to capture a long-term dependency in the time series, and the forward calculation layer and the backward calculation layer process the input series in turn” (e.g., specifying a bidirectional gated neural-network architecture that uses gating functions to mathematically control state updates for short and long term time-series dependencies.)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
In accordance with Step 2A, Prong 2, the claim does not include any additional elements and the judicial exception is not integrated into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding claim 8
Step 2A Prong 1
Claim 8 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“wherein the meta learner trains and predicts the results of Xgboost and the bidirectional gated neural network, and the meta learner is constructed as a linear regression model to learn and predict the results of Xgboost and the bidirectional gated neural network” (e.g., applying a linear regression model to mathematically combine and predict outputs of multiple predictive models)
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
In accordance with Step 2A, Prong 2, the claim does not include any additional elements and the judicial exception is not integrated into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Regarding claim 9
Step 2A Prong 1
Claim 9 recites the following mathematical concepts, that in each case under the BRI, covers performance of mathematical relationships, mathematical formulas or equations, and mathematical calculations but for recitation of generic computer components (e.g., “stacked time series network” and “bidirectional gated neural network”) [see MPEP 2106.04(a)(2)(I)].
“performing z-score normalization on real-time collected gas-in-oil data, and then dividing normalized data into the training set and the test set to train the stacked time series network for fault diagnosis, wherein if a new data type or a relevant influencing factor needs to be added, the original stacked time series network is taken as a pre-training model to activate all layers for training” (e.g., mathematically standardizing numerical values by subtracting a mean and dividing by a standard deviation, partitioning numerical data into subsets for training and evaluation, optimizing model parameters using numerical training data to generate predictions/classifications and using a pre-trained model and updating/optimizing its parameters via additional; training (transfer learning/fine-tuning))
Accordingly, at Step 2A, prong one, the claim recites an abstract idea.
Step 2A Prong 2
In accordance with Step 2A, Prong 2, the claim does not include any additional elements and the judicial exception is not integrated into a practical application.
Step 2B
In accordance with Step 2B, the claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
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.
Claims 1, and 3-9 are rejected under 35 U.S.C. 103 as being unpatentable over He et al. (US 20210278478 A1, referred to as He), in view of Cheim et al. (“Furan analysis for liquid power transformers”, referred to as Cheim), in view of Luo et al. (Convolutional bi-directional long short term memory network based dynamic fault diagnosis for transformer DGA”, referred to as Luo), in view of Chen et al. (XGBoost: A Scalable Tree Boosting System”, referred to as Chen), in view of Siebel et al. (US 11954112 B2, referred to as Siebel).
The applied He reference has a common assignee with the instant application. Based upon the publication date of September 9, 2021 of the He reference, which is prior to the earliest effective filing date of the instant application, i.e., July, 26th 2022, He constitutes prior art under 35 U.S.C. 102(a)(1). The examiner additionally notes that the He reference claims priority to Chinese application No. 202010134616.3, filed on March 2nd 2020 which is more than one year prior to the earliest effective filing date of the instant application, i.e., July, 26th 2022.
Regarding claim 1, He teaches A power transformer fault diagnosis method based on a stacked time series network, comprising:
(1) collecting gas-in-oil information of each substation3, wherein the gas-in-oil information comprises monitoring information of an oil test and contents of dissolved gas and furan in oil in each substation ( He [0003-0008]: Describes transformer fault diagnosis using online monitoring data gathering from dissolved gas in transformer oil. It obtains multiple groups of monitoring data of dissolved gas and associates monitoring data with correspond fault labels. This collects information comprising gas-in-oil data obtained from transformers online from multiple locations. TI uses CNNs and recurrent neural networks in a LSTM fir transformer DGA fault diagnosis. ;[0042-0045]: Describes obtaining multiple groups of monitoring data of dissolved gas from a transformer as input for fault processing of fault diagnosis.);
Although He teaches collecting gas-in-oil information of each substation, wherein the gas-in-oil information comprises monitoring information of an oil test and contents of dissolved gas, He does not explicitly teach furan in oil in each substation.
In the same field, analogous art Cheim teaches furan in oil (Page 9 History of Furan Analysis: “The procedure for sampling the oil is the same as that for dissolved-gas analysis.” Corresponds to monitoring information of an oil test and tying it to dissolved gas analysis.; Page 10-11, Furan Analysis Techniques : Describes that furanic compounds are determined form transformer oil via routine oil sampling/testing, and links oil sampling for furan testing to dissolved gas analysis sampling.)
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s gas-in-oil information collection with Cheim’s furan in oil analysis. Doing so would have enabled the system to dynamically address multiple components in samples to further diagnose issues in remote sites allowing for a system that identifies faults in a timely manner, as suggested by Cheim (see, Page 1, Introduction).
(2) normalizing the collected gas-in-oil information to obtain a normalized matrix (He [0044-0051]: Describes normalization on the collected dissolved gas monitoring data. Instep S1 the data is collected and then a normalization process is performed on the data to create a output of normalized information for further processing. );
(3) dividing the normalized matrix into a training set and a test set in proportion to train network parameters (He [0053-0054]: Describes that the dissolved gas data is used to build feature parameters and “the normalized new data set is divided into a training set and a verification set according to the proportion as the input of the LSTM diagnosis model”.);
Although He in view of Cheim teaches collecting gas-in-oil information of each substation, wherein the gas-in-oil information comprises monitoring information of an oil test and contents of dissolved gas and furan in oil in each substation, normalizing the collected gas-in-oil information to obtain a normalized matrix, and dividing the normalized matrix into a training set and a test set in proportion to train network parameters, He in view of Cheim does not explicitly teach constructing a stacked time series network based on Xgboost and a bidirectional gated neural network, inputting the training set and the test set to train the stacked time series network, and learning a feature of gas-in-oil data of a transformer.
In the same field, analogous art Luo teaches, constructing a stacked time series network …and a bidirectional gated neural network, inputting the training set and the test set to train the stacked time series network, and learning a feature of gas-in-oil data of a transformer (Luo Page 1-2 Introduction: Describes a model as an end-to-end trainable model by stacking the multiple layers. It uses time-series as multi-time characteristic gas data as input, to extract time dimension features via a gated neural recurrent network The method is an online/dynamic transformer DGA fault diagnosis and its purpose as an online transformer diagnosis method for online modeling.);
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s in view of Cheim’s gas-in-oil information collection with Luo’s bidirectional gated recurrent neural network. Doing so would have enabled the system to improve temporal feature extraction to exploit forward and backward dependencies in the data to generate refined diagnosis faults, as suggested by Luo (see, Pages 1-2, Introduction).
Although He in view of Cheim and further in view of Luo teaches constructing a stacked time series network … and a bidirectional gated neural network, inputting the training set and the test set to train the stacked time series network, and learning a feature of gas-in-oil data of a transformer, He in view of Cheim and further in view of Luo does not explicitly teach constructing a stacked time series network based on Xgboost.
In the same field, analogous art Chen teaches based on Xgboost (Chen Section 2: Describes a XGBoost as a tree boosting (tree ensemble) model in which the final prediction is obtained by summing the outputs of multiple regression trees and training the ensemble by minimizing a regularized objective.)
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s in view of Cheim’s in view of Luo’s bidirectional gated recurrent neural network with Chen’s Xgboost architecture. Doing so would have enabled the system to improve classification performance on feature data to reliably assist in algorithm fault diagnosis, as suggested by Chen (see, Page 1, Introduction).
(5) performing fault diagnosis based on real-time gas-in-oil data during operation (Luo Pages, 1, 2, and 5, Introduction, Convolutional neural network, and Conclusion: Describes an online system for fault diagnosis, which monitors transformer DGA and uses multi-time characteristics gas data as input to the diagnostic model.),
Although He in view of Cheim in view of Luo in view of Chen teaches performing fault diagnosis based on real-time gas-in-oil data during operation, He in view of Cheim in view of Luo in view of Chen does not explicitly teach fine tuning a weight of the stacked time series network to enable the stacked time series network to continuously learn a new feature.
In the same field, analogous art Siebel teaches, fine tuning a weight of the stacked time series network to enable the stacked time series network to continuously learn a new feature (Col. 119 Lines 1-59: Describes updating a deployed machine-learning model based on operational feedback, including learning form investigation outcomes and using that information to “learn and update automatically”, “train and retrain” the model, and “adjust learned detection parameters”. These correspond to retraining the model with the adjusted learned detection parameters with updating model parameters/weights to fine-tune. These updates occur as new operational information becomes available to continuously learn from the new data. ).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He in view of Cheim in view of Luo, in view of Chen, with Siebel’s continuously fine tuning. Doing so would have enabled the system to proactively update parameters, to find faults as they occur, allowing for real time information to be available and fix potential issues as they occur, as suggested by Siebel (see, col. 74, line 65-col. 75, line 12 and col. 119, line 60-col. 120, line 3).
Regarding claim 3, as discussed above, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 1.
He further teaches wherein step (2) comprises: performing z-score normalization on the gas-in-oil information to obtain the normalized matrix (He [0044-0051]: Describe performing z-score normalization on gas-in-=oil information by applying a normalization formula using the mean E(X) and variance D(X) to obtain regularized value zi and expressing the resulting normalized dataset as datai-’={bi,1, bi,2...bi,N, si} where each bi,j is a normalized gas parameter value (entries of the normalized matrix).).
Regarding claim 4, as discussed above, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 1.
He further teaches wherein the data in the gas-in-oil information is divided into two parts in step (3), wherein data of a certain proportion is used as the training set to train the stacked time series network, and a data of a remaining proportion is used as the test set to test a fault diagnosis effect of the stacked time series network for the transformer (He [0044-0052]: Describes dividing the normalized matrix (normalized gas-in-oil information) into two parts in a specified proportion, using one portion as a training set to train the stacked time series network/model parameters and using the remaining portion as a test set to test the transformer fault diagnosis effect.).
Regarding claim 5, as discussed above, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 4.
He in view of Cheim and further in view of Luo does not explicitly teach, but Chen teaches wherein step (4) comprises:
(4.1) constructing the stacked time series network based on Xgboost and the bidirectional gated neural network to perform feature extraction and prediction on the gas-in-oil information, wherein construction of Xgboost comprises establishment of an integrated model, selection of an objective function, and solving of a loss function (Chen Section 2.1, and 2.2: Describes constructing an XGBoost model by establishing an integrated tree ensemble model comprising multiple additive decision trees, selecting a regularized objective function measuring error between predicted and target values, and iteratively solving the loss function during training through gradient-based optimization.);
and the bidirectional gated neural network comprises a forward calculation layer, a backward calculation layer, an update gate, and a reset gate (Luo pages 2-4, Sections 2, 3 and 4, and Equations 3-8: Describes a bidirectional long short-term memory (Bi-LSTM) network for transformer gas-in-oil time-series data, including a forward calculation layer, a backward calculation layer to process time series in opposite directions. The recurrent unit includes gating mechanisms (input/forget/output gates) that control how much prior state information is discarded and how much prior state is retained and updated with new input, which provides the same short-term and long-term dependency control as rest and update gates. Feature extraction and prediction on gas-in-oil time-series data is constructed using bidirectional gated architecture.);
(4.2) predicting gas in the oil by using Xgboost and the bidirectional gated neural network, and outputting a prediction result of the gas-in-oil information(Luo pages 3-5, Section 4: Describes inputting multi-time gas-in-oil data into the Bi-LSTM network and outputting prediction results corresponding to transformer operation/fault conditions.); and
(4.3) training prediction results of Xgboost and the bidirectional gated neural network by using a meta learner, to output the prediction result of the gas-in-oil information, performing fault diagnosis on the stacked time series network by using a Softmax layer, and outputting the fault state of the transformer (He [0020-0024]: Describes combining the prediction outputs of multiple models by obtaining diagnosis support degrees output from the SoftMax layers of different diagnosis models (LSTM and CNN) for the same input data, then forming a support degree matrix, and performing DS evidence fusion to obtain confidence values for fault labels and outputting the fault label having maximum confidence as the final diagnosis result. These fusion steps function as a second stage decision mechanism that takes model prediction results as inputs and outputs the final fault state. ; Luo pages 4-5, Sections 4, and 5: Describes performing fault classification on transformer gas-in-oil data using a SoftMax output layer, where fault states are output as classification results.).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s in view of Cheim’s in view of Luo’s bidirectional gated recurrent neural network with Chen’s Xgboost architecture. Doing so would have enabled the system to enable reliable training and operation to facilitate effective feature extraction and prediction, as suggested by Chen (see, Page 2, Section 2.1).
Regarding claim 6, as discussed above, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 5.
He in view of Cheim and further in view of Luo does not explicitly teach, but Chen teaches wherein the construction of Xgboost comprises the establishment of the integrated model, the selection of the objective function, and the solving of the loss function, wherein the establishment of the integrated model is to recursively construct a binary decision tree, and in input space of the training set, each region is recursively divided into two sub-regions based on a minimum squared-error criterion, and an output value of each sub-region is determined (Chen Section 2.1-2.2, Equations 4, 5, 7, and Algorithm 1: Describes constructing an integrated model in the form of a tree ensemble, where each base learner is a regression tree (CART). The tree is built by recursively partitioning the input space into left and right sub-regions via greedy splitting, and that each resulting leaf (sub-region) is assigned a continuous output value. It splits/partitions are selected to improve the objective by maximizing loss reduction (gain), which is selecting partitions that reduce prediction error as measured by the chosen loss function.);
the selection of the objective function is to measure an error between a predicted value and a real value of a target, and the objective function is approximated through second-order Taylor expansion (Chen Section 2.1-2.2 and equation 3: Describes selecting an objective function that measures the error between predicted values and target values via a loss function. It approximates the objective function using a second-order Taylor expansion, computing first-order and second-order derivatives of the loss to facilitate efficient optimization.); and
the solving of the loss function is to partition a sub-tree by using a greedy algorithm, enumerate feasible partitioning points, in other words, add a new partition to an existing leaf each time, and calculate a corresponding maximum gain (Chen Section 2.2, Equation 7 and Algorithm1: Describe solving the loss function using a greedy tree-construction algorithm, wherein the process begins from a single leaf and iteratively adds new partitions. It enumerates feasible partitioning points for candidate splits and calculating a corresponding gain (loss reduction) for each split, selecting the split that maximizes gain.).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s in view of Cheim’s in view of Luo’s bidirectional gated recurrent neural network with Chen’s Xgboost architecture. Doing so would have enabled the system to train and deploy the XGBoost with more efficiency and at scale, improving runtime performance while maintaining predictive accuracy, as suggested by Chen (see, Pages 2-3, Sections 2.1, 2.2 and 3.1).
Regarding claim 7, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 6.
He in view of Cheim does not explicitly teach, but Luo teaches, wherein the bidirectional gated neural network comprises the forward calculation layer, the backward calculation layer, the update gate, and the reset gate, wherein the reset gate helps to capture a short-term dependency in a time series, the update gate helps to capture a long-term dependency in the time series, and the forward calculation layer and the backward calculation layer process the input series in turn (Luo pages 2-4 Sections 2 and 3: Describes a bidirectional gated recurrent neural network comprising a forward recurrent layer and a backward recurrent layer that process an input time series in opposite temporal directions. Its gating mechanisms within each recurrent unit that selectively discard prior state information and selectively retain and update prior start information with new inputs, thereby capturing short-term and long-term dependencies in the time series. These gating mechanisms work as reset and update gates.).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He’s in view of Cheim’s gas-in-oil information collection with Luo’s bidirectional gated recurrent neural network. Doing so would have enabled the system to model temporal dependencies in the transformer DGA time-series data, as suggested by Luo (see, Section 3).
Regarding claim 8, as discussed above, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 7.
He in view of Cheim and Luo and Chen does not explicitly teach, but Siebel teaches wherein the meta learner trains and predicts the results of Xgboost and the bidirectional gated neural network (Chen Section 2: Describes that XGBoost is commonly combined with neural nets in ensembles, where solutions use nth an XGBoost model and a neural network model together as a n ensemble predictive approach.), and the meta learner is constructed as a linear regression model to learn and predict the results of Xgboost and the bidirectional gated neural network (Siebel Col 118, lines 50-67: Describes using linear regression as a machine learning model for prediction (a linear model used for regression), where a linear regression model can be constructed and used to learn/predict results.).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He in view of Cheim in view of Luo, in view of Chen, with Siebel’s linear regression. Doing so would have enabled the system to combine and refine prediction outputs from multiple base models to use linear regression efficiently, as suggested by Siebel (see, col. 118, lines 50-67).
Regarding claim 9, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 8.
He further teaches wherein step (5) comprises:
performing z-score normalization on real-time collected gas-in-oil data, and then dividing normalized data into the training set and the test set to train the stacked time series network for fault diagnosis (He[0044-0051]: Describe performing z-score normalization on gas-in-oil information by applying a normalization formula using the mean E(X) and variance D(X) to obtain regularized value zi and expressing the resulting normalized dataset as datai-’={bi,1, bi,2...bi,N, si} where each bi,j is a normalized gas parameter value (entries of the normalized matrix).
He in view of Cheim and Luo and Chen does not explicitly teach, but Siebel teaches wherein if a new data type or a relevant influencing factor needs to be added, the original stacked time series network is taken as a pre-training model to activate all layers for training (Siebel col 119 lines 1-59: Describes that a machine learning system operating on real-time/operational data can :learn and update automatically”, receive new information/feedback from investigations/operations, and use that information to train and retrain the machine learning model, where the model is updated as additional training data is collected over time (continued training based on new data/features).
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He in view of Cheim in view of Luo, in view of Chen, with Siebel’s continuously feeding of new data. Doing so would have enabled the system to deploy a trained model in an operational environment to automatically update/retrain while new data is feed in, improving robustness as conditions and input factors change, as suggested by Siebel (see, col. 119 lines 1-59).
Claim 2 is rejected under 35 U.S.C. 103 as being unpatentable over He et al. (US 20210278478 A1, referred to as He), in view of Cheim et al. (“Furan analysis for liquid power transformers”, referred to as Cheim), in view of Luo et al. (Convolutional bi-directional long short term memory network based dynamic fault diagnosis for transformer DGA”, referred to as Luo), in view of Chen et al. (XGBoost: A Scalable Tree Boosting System”, referred to as Chen), in view of Siebel et al. (US 11954112 B2, referred to as Siebel), and further in view of Worbel et al. (US 20190204289 A1, referred to as Worbel).
Regarding claim 2, He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches, the method according to claim 1, wherein the gas-in-oil information comprises data of the transformer during operation and data recorded by an electric power company, and each group of data comprises gas-in-oil data and a fault state of the corresponding transformer, wherein the gas-in-oil data comprises contents of nine key states: a breakdown voltage (BDV), water, acidity, hydrogen, methane, ethane, ethylene, acetylene, and furan (He [0003], and [0044-0045]: Describes that data may come from “relevant documents over the years” and “actual test data of power companies” (recorded by an electrical power company), in the context of monitoring/online monitoring data for transformer oil dissolved gas analysis.; [0046-0052]: Describes forming groups of monitoring data (dissolved gas contents) and associating each group with a transformer state/fault label corresponding to gas-in-oil data of transformer states, which can comprise hydrogen, methane, ethane, ethylene and acetylene as the contents in each data group.;
Cheim, Page 9 History of Furan Analysis: “The procedure for sampling the oil is the same as that for dissolved-gas analysis.” Corresponds to monitoring information of an oil test and tying it to dissolved gas analysis. ;Page 10-11, Furan Analysis Techniques : Describes that furanic compounds are determined form transformer oil via routine oil sampling/testing, and links oil sampling for furan testing to dissolved gas analysis sampling. It also discusses moisture as a major factor in paper degradation and includes test results for moisture in oil as part of recommended data collected alongside furans, corresponding to water and furan collection and analysis.).
Although He in view of Cheim, in view of Luo, in view of Chen, in view of Seibel teaches,
wherein the gas-in-oil information comprises data of the transformer during operation and data recorded by an electric power company, and each group of data comprises gas-in-oil data and a fault state of the corresponding transformer, wherein the gas-in-oil data comprises contents of nine key states: …, hydrogen, methane, ethane, ethylene, acetylene, and furan, Cheim, in view of Luo, in view of Chen, in view of Seibel does not explicitly teach that the contents contain a breakdown voltage (BDV), water, and acidity.
In the same field, analogous art Worbel teaches a breakdown voltage (BDV), water, and acidity ([0003-0005], and [0042-0047]: Describes monitoring transformer oil condition using measured operational data, including moisture (water) content, acidity, and breakdown voltage (BDV). During transformer operation, decomposition of insulating materials produces gases and water in transformer oil, and that moisture and acids in the oil substantially affect transformer condition and insulation performance. It determines characteristic physical properties of the transformer oil to include relative and/or absolute amount of acid, moisture content, and breakdown voltage as quantified parameters used to assess the state and health of the transformer oil.)
It would have been obvious to one of ordinary skill in the art at the time of the claimed invention to have combined He in view of Cheim in view of Luo, in view of Chen, in view of Siebel, with the breakdown voltage (BDV), water, and acidity of Worbel. Doing so would have enabled the system to gather more detailed data from transformers for fault diagnosis. This would enhance the system’s ability in identifying abnormal behavior in a single model without additional models or system architectures, as suggested by Worbel (see, paragraphs [0042-0047]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. See attached PTO-892 for additional art including:
US 9135557 B2: gas-in-oil diagnostic preprocessing
US 20200292608 A1: substation monitoring
US 11580842 B2: industrial monitoring and fault detection
US 11886320 B2: XGBoost and supervised learning
Any inquiry concerning this communication or earlier communications from the examiner should be directed to DONALD T RODEN whose telephone number is (571)272-6441. The examiner can normally be reached Mon-Thur 8:00-5:00 EST.
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, Omar Fernandez Rivas can be reached at (571) 272-2589. 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.
/D.T.R./Examiner, Art Unit 2128
/RANDALL K. BALDWIN/Primary Examiner, Art Unit 2125
1 “to enable the stacked time series network to continuously learn a new feature” is intended use language with no patentable weight. No continuous learning or any “new feature” is recited elsewhere in the claim or in any of its dependent claims.
2 As indicated above in the section 112(b) rejection of this claim, “collecting gas-in-oil information of each substation” has been interpreted as collecting gas-in-oil information of each of any group, plurality or set of substations.
3 As indicated above in the section 112(b) rejection of this claim, “collecting gas-in-oil information of each substation” has been interpreted as collecting gas-in-oil information of each of any group, plurality or set of substations.