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 .
DETAILED ACTION
1. The amendment filed 04/06/2026 has been received and considered. Claims 21-40 are presented for examination.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/process/file/efs/guidance/eTD-info-I.jsp.
2. Claims 21, 30, and 36 are rejected on the ground of nonstatutory double patenting as being unpatentable over claim 1, 9 and 16 of U.S. Patent No. US 11215363 B2. Although the claims at issue are not identical, they are not patentably distinct from each other because the Claims in the instant invention are anticipated by Claims of U.S. Patent No. US 11215363 B2 thus constitutes an obvious variation.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
3. Claims 21, 30 and 36 are rejected under 35 U.S.C. 103 as being unpatentable over, in view of Bickford (US 6917839 B2), and further in view of Delahoz et al. (“Survey on Fall Detection and Fall Prevention Using Wearable and External Sensors”).
As per Claim 21, 30 and 36, Bickford teaches a method/ apparatus comprising: at least one processing device configured/ non-transitory computer readable medium containing instructions that when executed cause at least one processing device (Fig. 1 & 9 and the description: computer along with its typically associated memory means) to:
processing data associated with operation of equipment to generate first indicators (Col. 7 lines 66-67, col. 8 lines 1-15 “referring to FIG. 1, the system 10 is generally comprised of a method and apparatus for performing high sensitivity surveillance of a wide variety of assets including industrial, utility, business, medical, transportation, financial, and biological processes and apparatuses wherein such process and/or apparatus asset preferably has at least two distinct modes or domains of operation (e.g., transient and steady state modes or domains). The system includes a training procedure 20 wherein a decision model 50 of an asset 12 (e.g., a process and/or apparatus) is derived from historical operating data using at least one of a plurality of computer-assisted techniques. Historical operating data includes a set of observations from normal operation of the asset 12 that is acquired and digitized by a data acquisition means 40 using any combination of electronic data acquisition hardware and signal processing software”);
classifying the first indicators into multiple classes, …, the … indicators indicative of the equipment operating under from at least one specified condition, the …indicators indicative of the equipment not operating under from the at least one specified condition (Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; Fig. 5-8,11, and 15, for classification of faults; col. 3 lines 40-51 “Moreover, one embodiment of the invention provides a surveillance system and method that provides an operating mode partitioning of the decision model which enables different parameter estimation methods, fault detection methods, and fault classification methods to be used for surveillance within each individual operating mode of an asset. This ability enables surveillance to be performed by the instant invention with lower false alarm rates and lower missed alarm rates than can be achieved by the known prior-art methods.” col. 3 lines 51- col 4 ln.50 “Furthermore, and in contrast to the known prior art, and in one embodiment of the invention, parameter estimation methods, fault detection methods, and fault classification methods may be individually tailored for each operating mode of the asset thereby providing additional capability to reduce decision error rates for the surveillance system.”; col. 11 lines 63-67, col. 12 lines 1-10 “Improving the accuracy of the fault classification procedure 76 accomplishes a reduction in the number of false alarms sent to a process operator or control system that can in turn result in an erroneous alarm or control action by the alarm or control action procedure 74. Further, improving the accuracy of the fault classification procedure 76 accomplishes a reduction in the number of missed alarms thereby accomplishing more timely alarm or control action by the alarm or control action procedure 74.”); and
generating by the one or more processors, a notification in response to one or more first indicators being classified into the class of… indicators (Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; Fig. 5-8,11, and 15, for classification of faults; col. 3 lines 40-51 “Moreover, one embodiment of the invention provides a surveillance system and method that provides an operating mode partitioning of the decision model which enables different parameter estimation methods, fault detection methods, and fault classification methods to be used for surveillance within each individual operating mode of an asset. This ability enables surveillance to be performed by the instant invention with lower false alarm rates and lower missed alarm rates than can be achieved by the known prior-art methods.” col. 3 lines 51- col 4 ln.50 “Furthermore, and in contrast to the known prior art, and in one embodiment of the invention, parameter estimation methods, fault detection methods, and fault classification methods may be individually tailored for each operating mode of the asset thereby providing additional capability to reduce decision error rates for the surveillance system.”; col 14 lines 1-5 “The results of the fault classification are thereafter communicated by a conventional communications link 80… to an operator console 82 or automated process control system 84 for possible alarm and/or control action.”);
automatically adjusting by the one or more processors, one or more parameters of the equipment based on the notification to control one or more operations of the equipment (col. 10 lines 48-59 “The surveillance procedure 60 further includes implementing an alarm or control action for the purpose of notifying an operator or taking a corrective action in response to a detected unacceptable status or condition of the asset 12.”, col. 11 lines 56-59 “The classified fault is then provided to the alarm or control action procedure 74 for the useful purpose of enabling an automated or operator directed corrective action or warning.”; col. 14 lines 1-6,” automated process control system 84 for possible alarm and/or control action.”; Claim 11 “including means for performing asset control as a function of the fault classification of determined fault indications.”).
Bickford fails to teach explicitly the multiple classes including true positive indicators and false positive indicators, and storing by the one or more processors, the true positive indicators for refining future classification of the at the at least one specified condition.
Delahoz et al. teaches the multiple classes including true positive indicators and false positive indicators (“confusion matrix” on Pg19816), and
storing by the one or more processors, the true positive indicators for refining future classification of the at the at least one specified condition (Fig. 4 -5“The results of a classifier are commonly stored in an array known as confusion matrix. It allows to visualize the learning algorithm's performance in a specific table. … Performance indicators used to evaluate the efficiency of learning algorithms… ROC (Receiver Operating Characteristic) curves are also used as a tool for diagnostic test evaluation. The performance of a binary classifier is plotted as the classifier’s discrimination threshold is varied. It is created by plotting the recall vs. the fall-out, at various threshold settings. The threshold is used to determine the class the current instance belongs to, when the output of the classifier is a real value (continuous output).” on Pg19816-19817).
Bickford and Delahoz et al. are analogous art because they are both related to a method for classification and fault detection with trainable models.
It would have obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to combine the teachings of cited references. Thus, one of ordinary skill in the art before the effective filling date of the claimed invention would have been motivated to incorporate Delahoz et al. into Bickford’s invention for purpose of a method for a fault detection and a fault classification to visualize the learning algorithm’s performance of a classifier for a particular dataset (Delahoz et al.: Pg19816).
4. Claims 22-29 and 37-40 are rejected under 35 U.S.C. 103 as being unpatentable over, in view of Bickford (US 6917839 B2), in view of Delahoz et al. (“Survey on Fall Detection and Fall Prevention Using Wearable and External Sensors”) and further in view of Marco (“Alarm flood reduction using multiple data sources”).
Bickford as modified by Delahoz et al. teaches most all the instant invention as applied to claims 21, 30 and 36 above.
As per Claim 22 and 37, Bickford as modified by Delahoz et al. teaches further comprising:
collecting the data associated with the operation of the equipment from the one or more sensors (Bickford: Col. 33 lines 65-67, Col. 27 lines 1-44, “all sensors and data observations”, “Time to Failure Detection (Failure Simulations Only)--This is a measure of the elapsed time between the first observation containing sensor failure data and the observation”),
filtering the data to identify: training data for training a detection algorithm during a training period (Delahoz et al.: section 2.3.4, Figure 3).
Bickford as modified by Delahoz et al. fails to teach explicitly evaluation data to evaluate whether the equipment is operating under from the at least one specified condition during an evaluation period; identifying the training period and the evaluation period based on one or more rules, wherein the one or more rules specify operational ranges for multiple process variables of the data to select the training data; and generating one or more models and threshold values, by the detection algorithm, based on the training data.
Marco teaches evaluation data to evaluate whether the equipment is operating under from the at least one specified condition during an evaluation period (Chapter 3 “Root cause analysis”, “root cause analysis is performed for the different clusters in step four” on Pg 28, section 5. 5 “In the alarm analysis stage, faults that start alarm floods are identified and characterized.”);
identifying the training period and the evaluation period based on one or more rules, wherein the one or more rules specify operational ranges for multiple process variables of the data to select the training data (section 5.3.3 “The clustering method used in this thesis is the agglomerative hierarchical clustering.” “filtering out these alarms is to consider just those alarms that have a high frequency of occurrence in the training set of alarm floods”. Section 5.5 “The first step for the topology based validation of the data-driven analysis is to identify the assets that correspond to the signals used in the data-driven analysis. These assets will be indicators and controllers. Additionally, the alarm template that is being analyzed there may contain alarms that are not associated to signals but to assets, e.g. alarms indicating a malfunctioning pump. These assets are also automatically included in the topology analysis”); and
generating one or more models and threshold values, by the detection algorithm, based on the training data (Section 5.4 - 5.5 “threshold”, “root-cause suggestions … with plant connectivity”, Figure 5.4, Figure 6.16).
Bickford, Delahoz et al., and Marco are analogous art because they are all related to a method for classification and fault detection with trainable models.
It would have obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to combine the teachings of cited references. Thus, one of ordinary skill in the art before the effective filling date of the claimed invention would have been motivated to incorporate Marco into Bickford as modified by Delahoz et al.’s invention for purpose of a method for a fault detection and a fault classification to visualize the learning algorithm’s performance of a classifier for a particular dataset (Delahoz et al.:Pg19816) and to reduce alarm flood periods in process plants so to increase the efficiency of plants (Marco: Abstract, Pg 13).
As per Claim 23 and 38, Bickford as modified by Delahoz et al. teaches wherein generating a notification comprises:
predicting, by the detection algorithm the true positive indicators indicative of the equipment operating under from the at least one specified condition (Bickford: Claim 7-10; Delahoz et al.: “confusion matrix” on Pg19186); and
transmitting the notification to one or more operator consoles based on an identity of the true positive indicators (Bickford: Col. 14 lines 1-15; Delahoz et al.: ““confusion matrix” on Pg19186).
As per Claim 24 and 39, Bickford as modified by Delahoz et al. teaches further comprising:
generating one or more additional models to classify the first indicators into the multiple classes (Bickford: Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults; col. 3, line 40-51 “Moreover, one embodiment of the invention provides a surveillance system and method that provides an operating mode partitioning of the decision model which enables different parameter estimation methods, fault detection methods, and fault classification methods to be used for surveillance within each individual operating mode of an asset. This ability enables surveillance to be performed by the instant invention with lower false alarm rates and lower missed alarm rates than can be achieved by the known prior-art methods.” See also, col. 3, line 51- col 4 ln.50 “Furthermore, and in contrast to the known prior art, and in one embodiment of the invention, parameter estimation methods, fault detection methods, and fault classification methods may be individually tailored for each operating mode of the asset thereby providing additional capability to reduce decision error rates for the surveillance system.”); and
training at least one additional model to classify the first indicators into the multiple classes (Bickford: Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults; col. 3, line 40-51 “Moreover, one embodiment of the invention provides a surveillance system and method that provides an operating mode partitioning of the decision model which enables different parameter estimation methods, fault detection methods, and fault classification methods to be used for surveillance within each individual operating mode of an asset. This ability enables surveillance to be performed by the instant invention with lower false alarm rates and lower missed alarm rates than can be achieved by the known prior-art methods.” See also, col. 3, line 51- col 4 ln.50 “Furthermore, and in contrast to the known prior art, and in one embodiment of the invention, parameter estimation methods, fault detection methods, and fault classification methods may be individually tailored for each operating mode of the asset thereby providing additional capability to reduce decision error rates for the surveillance system.”), wherein training at least one additional model comprises: combining (i) data associated with the multiple process variable signals, (ii) threshold values identified using the training data, and (iii) statistical values identified using the evaluation data into combined data, the process variable signals identifying values of process variables associated with the equipment (Bickford: FIG. 2, col. 8, ln. 33- col. 10, line 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;);
identifying first observations in the combined data that do not result in the at least one specified condition and assigning the first observations to the class of false positive indicators (Bickford: FIG. 2, col. 8, ln. 33- col. 10, line 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;); and
identifying second observations in the combined data that precede occurrences of the at least one specified condition and assigning the second observations to the class of true positive indicators (Bickford: FIG. 2, col. 8, ln. 33- col. 10, line 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;).
As per Claim 25, Bickford as modified by Delahoz et al. teaches wherein classifying the first indicators into the multiple classes comprises using a support vector machine, the support vector machine separating the multiple classes in a higher-dimensional feature space to which the first indicators are mapped (Bickford: Fig. 12-13, 16-17, training vectors Fig. 24-27, vectors used for training, col. 31, line 25-40, “During each training epoch, the error rate for the neural network is calculated. The error rate is defined to be the fraction of input vectors that are incorrectly classified by the neural network. An input vector is correctly classified if the weight vector that is closest to it connects to an output node of the same class as the input vector. As each input vector in the training set is passed through the LVQ neural network during a training epoch, the program notes if the input vector was correctly or incorrectly classified. The error rate is then given by the ratio of the number of incorrectly classified input vectors to the total number of input vectors in the training set. By keeping track of the error rate, the training algorithm can be halted as soon as the neural network stops learning.”).
As per Claim 26, Bickford as modified by Delahoz et al. teaches wherein the support vector machine uses a radial basis function as a kernel function to perform distance calculations in the higher-dimensional feature space (Bickford: col. 31, ln. 60 – col. 32, ln. 16, “cluster the input vectors that belong to the class into a number of clusters that equals the number of output nodes that belong to the class. For instance for class j, the K-means clustering algorithm is used to divide the input vectors into noutf clusters and to evaluate the centers of the clusters. The cluster centers for class j are used to initialize the weight vectors whose output nodes belong to the class. The K-means clustering algorithm evaluates cluster centers for the class by minimizing the Euclidean distances between each of the input vectors in the class and the cluster center nearest to each. Thus, each cluster center is the mean value of the group of input vectors in a cluster domain. The K-means clustering algorithm was found to improve the recall capabilities of the neural network over the random initialization scheme, at a minimal increase in the computational cost of the training calculations.”).
As per Claim 27 and 40, Bickford as modified by Delahoz et al. teaches transmitting the alerts to one or more operator consoles (Bickford: Col. 14 lines 1-15).
Bickford as modified by Delahoz et al. fails to teach explicitly further comprising: generating alerts corresponding to the first observations and the second observations; and displaying, on the one or more operator consoles, at least one of: a possible root cause of the alerts and a potential solution to the alerts.
Marco teaches generating alerts corresponding to the first observations and the second observations (Section 5.5 “root-cause suggestions … with plant connectivity”, Figure 6.16); and
displaying, on the one or more operator consoles, at least one of: a possible root cause of the alerts and a potential solution to the alerts (Section 5.5 “root-cause suggestions … with plant connectivity”, Figure 6.16).
As per Claim 28, Bickford as modified by Delahoz et al. teaches wherein: during the training period, (i) identifying the one or more models and one or more first statistical values using at least some of the training data and (ii) determining a threshold value using the one or more first statistical values, the one or more models representing the operation of the equipment (Bickford: FIG. 2, col. 8, ln. 33- col. 10, line 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;); and
during the evaluation period, (i) determining one or more second statistical values using at least some of the evaluation data and the one or more models, (ii) comparing the one or more second statistical values to the threshold value, and (iii) determining whether the equipment is operating under from the at least one specified condition based on the comparison (Bickford: FIG. 2, col. 8, ln. 33- col. 10, line 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;).
As per Claim 29, Bickford as modified by Delahoz et al. teaches wherein: the evaluation period occurs at a first interval; and the training period occurs at a second interval longer than the first interval and is performed to retrain the one or more models to changing conditions or operational modes of the equipment (Bickford: Fig. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also Fig. 5-8,11, and 15, for classification of faults;).
5. Claims 31-35 are rejected under 35 U.S.C. 103 as being unpatentable over, in view of Bickford (US 6917839 B2), in view of Delahoz et al. (“Survey on Fall Detection and Fall Prevention Using Wearable and External Sensors”) and Marco (“Alarm flood reduction using multiple data sources”), and further in view of Hofman et al. (US 20090240636 A1).
Bickford as modified by Delahoz et al. teaches most all the instant invention as applied to claims 21, 30 and 36 above.
As per Claim 31, Bickford as modified by Delahoz et al. teaches a process data acquisition system and a process data historian wherein the process data acquisition system and the process data historian comprise one or more processors, wherein the one or more processors are configured (Fig. 1 & 9 and the description: computer along with its typically associated memory means) to collect the data related to the equipment from one or more sensors (Bickford: Col. 33 lines 65-67, Col. 27 lines 1-44, “all sensors and data observations”, “Time to Failure Detection (Failure Simulations Only)--This is a measure of the elapsed time between the first observation containing sensor failure data and the observation”);
a data cleansing function to filter the data associated with the operation of the equipment (Delahoz et al.: section 2.3.4, Figure 3) to identify: training data for training a detection algorithm during a training period (Delahoz et al.: section 2.3.4, Figure 3);
…
predict the true positive indicators indicative of the equipment operating under from the at least one specified condition based on the evaluation data (Bickford: Claim 7-10; “confusion matrix” on Pg19186); and
an event detection function configured to provide an identity of the true positive indicators to at least one of: an event notification function (Bickford: Col. 14 lines 1-15; Delahoz et al.: “confusion matrix” on Pg19186) and an event visualization function to generate event notification messages to be transmitted to one or more operator consoles.
Bickford as modified by Delahoz et al. fails to teach explicitly wherein the data comprises measurements of combustible levels, draft levels, oxygen (02) levels, pressure levels, and temperature levels associated with the equipment; and
evaluation data to evaluate whether the equipment is operating under from the at least one specified condition during an evaluation period;
the detection algorithm configured to:
generate one or more models and threshold values based on the training data.
Marco teaches evaluation data to evaluate whether the equipment is operating under from the at least one specified condition during an evaluation period (Chapter 3 “Root cause analysis”, “root cause analysis is performed for the different clusters in step four” on Pg 28, section 5. 5 “In the alarm analysis stage, faults that start alarm floods are identified and characterized.”);
the detection algorithm configured to: generate one or more models and threshold values based on the training data (section 5.3.3 “The clustering method used in this thesis is the agglomerative hierarchical clustering.” “filtering out these alarms is to consider just those alarms that have a high frequency of occurrence in the training set of alarm floods”. Section 5.5 “The first step for the topology based validation of the data-driven analysis is to identify the assets that correspond to the signals used in the data-driven analysis. These assets will be indicators and controllers. Additionally, the alarm template that is being analyzed there may contain alarms that are not associated to signals but to assets, e.g. alarms indicating a malfunctioning pump. These assets are also automatically included in the topology analysis”; Section 5.4, Figure 5.4, “threshold”; Section 5.5 “root-cause suggestions … with plant connectivity”, Figure 6.16).
Bickford, Delahoz et al., and Marco are analogous art because they are all related to a method for classification and fault detection with trainable models.
It would have obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to combine the teachings of cited references. Thus, one of ordinary skill in the art before the effective filling date of the claimed invention would have been motivated to incorporate Marco into Bickford as modified by Delahoz et al.’s invention for purpose of a method for a fault detection and a fault classification to provide a real-time anomaly detection in multi-threaded processes (e.g., controller, sensor processing, and sensor fusion implementations) trained classifier with efficacy (Delahoz et al.: Abstract and right column of Pg 666) and to reduce alarm flood periods in process plants so to increase the efficiency of plants (Marco: Abstract, Pg 13)
Further Hofman et al. teaches wherein the data comprises measurements of combustible levels, draft levels, oxygen (02) levels, pressure levels, and temperature levels associated with the equipment (paras. [0056]-[0082]; details how trainable models, including neural networks [0071]-077 are trained based on variables describing the combustion process “With the statistical trainable model, such as a neural network, the model parameters can be weights of the trainable statistical models. During training of such a model the weights are then adapted by using training data. In a development, the variables comprise at least a few of the following variables and/or variables derived therefrom: external air pressure, gas pressure, pressure after compression, pressure difference at an air filter, pilot gas, pilot gas stream, load, air temperature, gas temperature, temperature after compression, pressure reduction in a combustion chamber, a 'blade' position (Schaufel), load, gas turbine output (GtLstg), air temperature inflow (SaugT), air pressure (UmgPr), pressure difference over an air filter (PrDFi), pressure after compression stage (VerPr), temperature after compression stage (VerdT), inlet guide blade position, pressure difference in combustion chamber (DrVBr), gas pressure (GasDr), gas temperature (GasT), rotational frequency, rotational speed (Drehzahl), exhaust gas temperature (AbgasT), first humming pressure amplitude (WD0l ), second humming pressure amplitude (WD02). Furthermore, the influencing variables can comprise further variables for describing (further): fuel flows, fuel pressures, fuel temperatures or fuel flow ratios.”).
Bickford as modified by Delahoz et al. and Marco and Hofman et al. are analogous art because they are all related to a method for classification and fault detection with trainable models for equipment.
Thus, one of ordinary skill in the art before the effective filling date of the claimed invention would have been motivated to incorporate Hofman et al. into Bickford as modified by Delahoz et al. and Marco’s invention for purpose of a method for a fault detection and a fault classification to provide a more reliable and more accurate analysis of equipment including influencing a combustion process in the combustion chamber (Hofman et al.: Abstract, [0058]).
As per Claim 32, Bickford as modified by Delahoz et al. teaches further comprises:
…
wherein the training period comprises: (i) identifying the one or more models and one or more first statistical values using at least some of the training data and (ii) determining a threshold value using the one or more first statistical values, the one or more models representing the operation of the equipment (Bickford: FIG. 2, cols. 8, ln. 33- col. 10, lns. 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults;), and
wherein the evaluation period comprises: (i) determining one or more second statistical values using at least some of the evaluation data and the one or more models, (ii) comparing the one or more second statistical values to the threshold value, and (iii) determining whether the equipment is operating under from the at least one specified condition based on the comparison (Bickford: FIG. 2, cols. 8, ln. 33- col. 10, lns. 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults;).
Bickford as modified by Delahoz et al. fails to teach explicitly one or more rules to identify the training period and the evaluation period, wherein the one or more rules specify operational ranges for multiple process variables of the data to select the training data.
Marco teaches one or more rules to identify the training period and the evaluation period, wherein the one or more rules specify operational ranges for multiple process variables of the data to select the training data (section 5.3.3 “The clustering method used in this thesis is the agglomerative hierarchical clustering.” “filtering out these alarms is to consider just those alarms that have a high frequency of occurrence in the training set of alarm floods”. Section 5.5 “The first step for the topology based validation of the data-driven analysis is to identify the assets that correspond to the signals used in the data-driven analysis. These assets will be indicators and controllers. Additionally, the alarm template that is being analyzed there may contain alarms that are not associated to signals but to assets, e.g. alarms indicating a malfunctioning pump. These assets are also automatically included in the topology analysis”).
As per Claim 33, Bickford as modified by Delahoz et al. teaches further comprises:
one or more additional models to classify the first indicators into the multiple classes (Bickford: FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults; cols. 3, lns. 40-51 “Moreover, one embodiment of the invention provides a surveillance system and method that provides an operating mode partitioning of the decision model which enables different parameter estimation methods, fault detection methods, and fault classification methods to be used for surveillance within each individual operating mode of an asset. This ability enables surveillance to be performed by the instant invention with lower false alarm rates and lower missed alarm rates than can be achieved by the known prior-art methods.” See also, col. 3, lns. 51- col 4 ln.50 “Furthermore, and in contrast to the known prior art, and in one embodiment of the invention, parameter estimation methods, fault detection methods, and fault classification methods may be individually tailored for each operating mode of the asset thereby providing additional capability to reduce decision error rates for the surveillance system.”), wherein the at least one additional model is trained based on: combining (i) data associated with the multiple process variable signals, (ii) the threshold values identified using the training data, and (iii) statistical values identified using the evaluation data into combined data, the process variable signals identifying values of process variables associated with the equipment (Bickford: FIG. 2, cols. 8, ln. 33- col. 10, lns. 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults;);
identifying first observations in the combined data that do not result in the at least one specified condition and assigning the first observations to the class of false positive indicators (Bickford: FIG. 2, cols. 8, ln. 33- col. 10, lns. 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults;); and
identifying second observations in the combined data that precede occurrences of the at least one specified condition and assigning the second observations to the class of true positive indicators (Bickford: FIG. 2, cols. 8, ln. 33- col. 10, lns. 46, Description of the Training Procedure section, “Again referring to FIG. 2, the unique method for the 65 training procedure 20 also includes at least one of a parameter estimation submodel creation procedure 29, a fault detection submodel creation procedure 30, or a fault classification submodel creation procedure 31 for creating at least one decision submodel for inclusion in the decision model 50 using at least one training data subset 28. In practice, the designer first selects the operating modes that will be included in the decision model 50 by means of an operating mode enable procedure 32. The method thereafter is comprised of a training loop wherein each possible operating mode of the asset 12 is assessed for inclusion in the decision model 50. The training loop is in general controlled by two decision procedures. The mode enabled decision procedure 34 determines whether the designer intends a specific operating mode to be included in the decision model 50. If the operating mode is not to be included, no further processing is required and the training loop proceeds to the next possible operating mode as controlled by the more modes decision procedure 36. If the operating mode is to be included, the training data subset 28 associated with the currently selected operating mode is selected from the training data set 24. Depending on the preference of the 20 designer implementing the training loop, the operating mode determination and training data subset extraction procedures may be, in general, performed as needed or in advance of the submodel creation loop.” FIGs. 1-3, element 20, includes training data, based on operational data, and classification for fault detection for training, and then element 60 is fault detection; see also FIGs. 5-8,11, and 15, for classification of faults;).
As per Claim 34, Bickford as modified by Delahoz et al. teaches wherein classifying the first indicators into the multiple classes comprises using a support vector machine, the support vector machine separating the multiple classes in a higher-dimensional feature space to which the first indicators are mapped (Bickford: FIGs. 12-13, 16-17, training vectors FIGs. 24-27, vectors used for training, col. 31, lns. 25-40, “During each training epoch, the error rate for the neural network is calculated. The error rate is defined to be the fraction of input vectors that are incorrectly classified by the neural network. An input vector is correctly classified if the weight vector that is closest to it connects to an output node of the same class as the input vector. As each input vector in the training set is passed through the LVQ neural network during a training epoch, the program notes if the input vector was correctly or incorrectly classified. The error rate is then given by the ratio of the number of incorrectly classified input vectors to the total number of input vectors in the training set. By keeping track of the error rate, the training algorithm can be halted as soon as the neural network stops learning.”), and wherein the support vector machine uses a radial basis function as a kernel function to perform distance calculations in the higher-dimensional feature space (Bickford: col. 31, ln. 60 – col. 32, ln. 16, “cluster the input vectors that belong to the class into a number of clusters that equals the number of output nodes that belong to the class. For instance for class j, the K-means clustering algorithm is used to divide the input vectors into noutf clusters and to evaluate the centers of the clusters. The cluster centers for class j are used to initialize the weight vectors whose output nodes belong to the class. The K-means clustering algorithm evaluates cluster centers for the class by minimizing the Euclidean distances between each of the input vectors in the class and the cluster center nearest to each. Thus, each cluster center is the mean value of the group of input vectors in a cluster domain. The K-means clustering algorithm was found to improve the recall capabilities of the neural network over the random initialization scheme, at a minimal increase in the computational cost of the training calculations.”).
As per Claim 35, Bickford as modified by Delahoz et al. fails to teach explicitly wherein the one or more operator consoles is configured to display: one or more alerts generated corresponding to the first observations and the second observations and a possible root cause of the alerts and a potential solution to the alerts.
Marco teaches wherein the one or more operator consoles is configured to display: one or more alerts generated corresponding to the first observations and the second observations (Section 5.5 “root-cause suggestions … with plant connectivity”, Figure 6.16); and a possible root cause of the alerts and a potential solution to the alerts (Section 5.5 “root-cause suggestions … with plant connectivity”, Figure 6.16).
Response to Arguments
6. Applicant's arguments filed 04/06/2026 have been fully considered but they are not persuasive.
Examiner respectfully withdraws Objection to Specification in view of the amendment and/or applicant’s arguments.
Examiner respectfully withdraws Claim Objections in view of the amendment and/or applicant’s arguments.
As per Double Patenting, TD was filed on 12/15/2025, but it has been disapproved. Therefore, Double Patenting Rejection maintains.
Examiner respectfully withdraws Claim Interpretation in view of the amendment and/or applicant’s arguments.
Examiner respectfully withdraws Claim Rejections - 35 USC § 112 in view of the amendment and/or applicant’s arguments.
Examiner respectfully withdraws Claim Rejections - 35 USC § 101 in view of the amendment and/or applicant’s arguments.
As per 103 rejection, applicants have argued that:
The Applicant submits that the combination of Bickford and Delahoz do not teach, suggest, or render obvious one or more features of amended independent claim 21. For instance, Bickford and Delahoz either alone or in combination fails to describe, for example, the feature(s) of "generating by the one or more processors, a notification in response to one or more first indicators being classified into the class of true positive indicators; automatically adjusting by the one or more processors, one or more parameters of the equipment based on the notification to control one or more operations of the equipment; and storing by the one or more processors, the true positive indicators for refining future classification of the at the at least one specified condition" as recited in amended independent claim 1. [1]
…
However, Bickford does not disclose or imply any automatic control response that modifies operational parameters of the equipment in real time. [2] The control logic in Bickford terminates at the alert stage and requires manual operator action to correct a problem.[3] In contrast, the claimed method implements a processor- driven automatic adjustment of equipment parameters directly in response to true- positive classifications, thereby transforming a diagnostic tool into a self-correcting control apparatus.
As per [1], Examiner disagrees. As rejected above, it is Examiner’s position that the prior arts teach the claimed limitation. It is noted that Examiner relies on the teaching in Bickford to teach the limitation of “generating by the one or more processors, a notification in response to one or more first indicators being classified into the class of… indicators; automatically adjusting by the one or more processors, one or more parameters of the equipment based on the notification to control one or more operations of the equipment;” while Delahoz is relied upon for a teaching of “the multiple classes including true positive indicators and false positive indicators; storing by the one or more processors, the true positive indicators for refining future classification of the at the at least one specified condition".
As per [2], Examiner disagrees as Bickford teaches automated process control system 84 for possible alarm and/or control action along with operator-console 82 in parallel (Fig. 9, col. 14 lines 4-6). Moreover, Bickford discloses the control action control procedure 74 in Fig.11 which enabling an automated corrective automated or operator directed corrective action or warning. Also, real-time operation is expressly demonstrated in Bickford. Please see col. 36 lines 5-26 where it states:
Processing speed results demonstrated the real-time monitoring capability of the operating mode partitioned decision model. Single observation processing times of 5-msec (200 samples/second) were demonstrated with the seventeen (17) sensor SSME sensor validation module running on a 300-MHz Pentium II processor. It is reasonable to allocate between 2 and 50-msec per data cycle for sensor validation processing in SSME real-time control applications. The results of this testing show these goals are only attainable with operating mode partitioning of the MSET model in accordance with the instant invention. The unobvious benefits of the instant invention are therefore demonstrated by this reduction to practice.
As per [3], Examiner disagrees as Bickford’s control logic has two termini: operator-alert terminus (classification -> link 80 -> operator console 82 in Fig. 9) and automated-control terminus (classification -> link 80-> automated process control system 84 in Fig. 9). The automated process control system 84 executes corrective action automatically (col. 14 lines 4-6, col. 36 lines 5-26). Thus 103 rejection maintains.
Conclusion
7. THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
8. Any inquiry concerning this communication or earlier communications from the examiner should be directed to EUNHEE KIM whose telephone number is (571)272-2164. The examiner can normally be reached Monday-Friday 9am-5pm ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Ryan Pitaro can be reached at (571)272-4071. 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.
EUNHEE KIM
Primary Examiner
Art Unit 2188
/EUNHEE KIM/ Primary Examiner, Art Unit 2188
Prosecution Insights