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 .
Response to Arguments
Applicant's arguments filed August 25th, 2020 have been fully considered but they are not persuasive.
The Bang reference is overcome in light of the amendments; however a new reference, Wu, is used instead. Wu teaches extracting fault surfaces from seismic attribute images using reference windows across multiple directions and depth points (Wu, Fig. 4b, Fig. 5a). In combination with Sun, this yields predictable results: enhanced automation and robustness in fault surface extraction and earth modeling. Accordingly, the claims remain unpatentable over Sun in view of Wu.
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.
Claims 1, 3, 5-6, 8, 10-13, 15, 17, and 19-20 are rejected under 35 U.S.C. 103 as being unpatentable over US Patent 11,428,078 (Sun et al; Sun) in view of “Automatic fault interpretation with optimal surface voting,” Geophysics, Vol. 83:5, Wu et al; Wu.
Regarding claim 1, and analogous claims 8 and 15:
Sun teaches:
1. A method, comprising: receiving attributes related to a well;
(Sun, col. 2:38-42)
“First, the system may retrieve or obtain production rates and well operation constraints from one or more sources such as one or more databases. The system also may retrieve or obtain other temporal and spatial measurement data associated with the one or more wells.”
2. providing the attributes as inputs to a machine learning model,
(Sun, col. 2:49-54)
“Next, the system may use the training data subset, the validation data subset, and the test data subset to build and train a model using LSTM or GRU and/or CNN. The system may perform history matching and model training and perform model evaluation based on prediction accuracy of the test data subset.”
3. wherein: the machine learning model has been trained through a supervised learning process based on training data that comprises: wellbore attributes relating to a plurality of depth points;
(Sun, col. 6:30-36)
“In a first step, temporal and spatial data is collected [i.e. wherein: the machine learning model has been trained through a supervised learning process based on receiving training data,]. Production rates and well operation constraints and other optional temporal and spatial measurement data may be collected. The temporal and spatial data may be retrieved from available datasets from on-premise computing devices and/or cloud computing devices [i.e. comprising: wellbore attributes relating to a plurality of depth points;].”
Examiner notes that supervised learning is defined by its use of labeled datasets (temporal and spatial data, in this case) to “train algorithms that to classify data or predict outcomes accurately.”
4. and the machine learning model, as a result of being trained through the supervised learning process based on the training data, is configured to output values [that relate to the plurality of directions corresponding to the reference window of the particular size in horizontal space and the corresponding reference window of the respective size in vertical space with respect to a given depth point rather than only relating to the given depth point;]
(Sun, col. 8: 9-14)
“The model may use input features from one or more time stamps as an input sequence (i.e., six previous time stamps) in a bottom input layer, perform machine learning over a hidden middle layer, and generate an output of production responses across a same or fewer time stamps (i.e., three time stamps) in an output layer.”
5. receiving outputs from the machine learning model in response to the inputs, wherein the outputs indicate predicted wellbore attribute values relating to: at least one depth point;
(Sun, col. 8: 9-14)
“The model may use input features from one or more time stamps as an input sequence (i.e., six previous time stamps) in a bottom input layer, perform machine learning over a hidden middle layer, and generate an output of production responses across a same or fewer time stamps (i.e., three time stamps) in an output layer [i.e. receiving outputs from the machine learning model in response to the inputs,].”
(Sun, col. 12:24-32)
“Certain aspects of the model may be tuned by adjusting hyper parameters such as a number of time stamps for the input sequence, a number of time stamps for the output sequence, CNN-related parameters such as kernel size, dropout rate, and learning rate, among others. In this case, the model may be based on LSTM or GRU whereby the output responses are based on previous production rates, well operation constraints, and spatial static data or images, among other input [i.e. wherein the outputs indicate predicted wellbore attribute values relating to: at least one depth point;].”
6. and generating an earth model based on the outputs from the machine learning model.
(Sun, col. 9: 24-35)
“The model development engine 204 can be used to obtain data, process data, analyze data, and generate one or more models that can be used by the forecast engine 206 to forecast well production for one or more wells [i.e. and generating an earth model]. The model development engine 204 captures time-dependency of well production responses on previous production history for one or more wells and also captures localized features such as workover operations. The model generated by the model development engine 204 learns from previous production history and corresponding temporal and spatial input data to decide when, where, and how to capture details that impact future production forecasts for the one or more wells [i.e. based on the outputs from the machine learning model].”
Sun does not explicitly teach:
1. and the plurality of directions corresponding to a reference window of the particular size in horizontal XY space and the corresponding reference window of the respective size in vertical Z space with respect to the at least one depth point;
2. [and the machine learning model, as a result of being trained through the supervised learning process based on the training data, is configured to output values] that relate to the plurality of directions corresponding to the reference window of the particular size in horizontal XY space and the corresponding reference window of the respective size in vertical Z space with respect to a given depth point rather than only relating to the given depth point;
3. and adjacent waveform data comprising corresponding wellbore attribute values relating to a plurality of directions corresponding to a reference window of a particular size in horizontal XY space and a corresponding reference window of a respective size in vertical Z space with respect to each depth point of the plurality of depth points;
Wu teaches:
1. and the plurality of directions corresponding to a reference window of the particular size in horizontal XY space and the corresponding reference window of the respective size in vertical Z space with respect to the at least one depth point;
(Wu, pg. 6, Fig. 4b, Fig. 5a, ¶3)
“To pick the fault surface from the fault attribute image (Figure 4b), we first transpose the attribute image into a new space shown in (a), where the vertical axis represents the inline [i.e. and the corresponding reference window of the respective size in vertical Z space] and horizontal axes represent the depth and crossline [i.e. and the plurality of directions corresponding to a reference window of the particular size in horizontal XY space]. The fault surface to be picked in this new space is a slightly dipping surface that laterally extends throughout the whole attribute image. Picking such a fault surface from the transposed attribute image can be considered as a problem of searching for an optimal surface that passes through the control point (white circle) and follows globally maximum attribute values [i.e. with respect to the at least one depth point;].”
PNG
media_image1.png
334
1395
media_image1.png
Greyscale
2. [and the machine learning model, as a result of being trained through the supervised learning process based on the training data, is configured to output values] that relate to the plurality of directions corresponding to the reference window of the particular size in horizontal XY space and the corresponding reference window of the respective size in vertical Z space with respect to a given depth point rather than only relating to the given depth point;
(Wu, pg. 6, Fig. 4b, Fig. 5a, ¶3)
“To pick the fault surface from the fault attribute image (Figure 4b), we first transpose the attribute image into a new space shown in (a), where the vertical axis represents the inline [i.e. and the corresponding reference window of the respective size in vertical Z space] and horizontal axes represent the depth and crossline [i.e. that relate to the plurality of directions corresponding to the reference window of the particular size in horizontal XY space]. The fault surface to be picked in this new space is a slightly dipping surface that laterally extends throughout the whole attribute image. Picking such a fault surface from the transposed attribute image can be considered as a problem of searching for an optimal surface that passes through the control point (white circle) and follows globally maximum attribute values [i.e. with respect to a given depth point rather than only relating to the given depth point;].”
PNG
media_image1.png
334
1395
media_image1.png
Greyscale
3. and adjacent waveform data comprising corresponding wellbore attribute values relating to a plurality of directions corresponding to a reference window of a particular size in horizontal XY space and a corresponding reference window of a respective size in vertical Z space with respect to each depth point of the plurality of depth points;
(Wu, pg. 6, Fig. 4b, Fig. 5a, ¶3)
“To pick the fault surface from the fault attribute image (Figure 4b), we first transpose the attribute image into a new space shown in (a), where the vertical axis represents the inline [i.e. and adjacent waveform data comprising corresponding wellbore attribute values relating to a plurality of directions and the corresponding reference window of the respective size in vertical Z space] and horizontal axes represent the depth and crossline [i.e. and the plurality of directions corresponding to a reference window of the particular size in horizontal XY space]. The fault surface to be picked in this new space is a slightly dipping surface that laterally extends throughout the whole attribute image. Picking such a fault surface from the transposed attribute image can be considered as a problem of searching for an optimal surface that passes through the control point (white circle) and follows globally maximum attribute values [i.e. with respect to each depth point of the plurality of depth points;].
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Sun with Wu. The motivation is to integrate Sun’s machine learning framework with Wu’s fault imaging process in order to enhance the automation, accuracy, and generalizability of fault interpretation. Specifically, Wu provides fault attribute images and extracted surfaces that are inherently noisy and discontinuous, and Sun provides machine learning techniques that capture patterns and generate predictive outputs. The combination would yield predictable results:
Regarding claim 3 and analogous claims 10 and 17:
The combination of Sun and Wu teach the method of claim 1.
Sun teaches:
1. receiving updated attributes related to the well;
(Sun, col. 9:24-31)
“The model development engine 204 can be used to obtain data, process data, analyze data, and generate one or more models that can be used by the forecast engine 206 to forecast well production for one or more wells. The model development engine 204 captures time-dependency of well production responses on previous production history for one or more wells and also captures localized features such as workover operations [i.e. receiving updated attributes related to the well;].”
2. providing the updated attributes as inputs to the machine learning model;
(Sun, col. 9:24-31)
“The model development engine 204 can be used to obtain data, process data, analyze data, and generate one or more models that can be used by the forecast engine 206 to forecast well production for one or more wells. The model development engine 204 captures time-dependency of well production responses on previous production history for one or more wells and also captures localized features such as workover operations [i.e. providing the updated attributes as inputs to the machine learning model].”
3. and generating an updated earth model based on updated outputs from the machine learning model. (Sun, col. 9:31-35)
“The model generated by the model development engine 204 learns from previous production history and corresponding temporal and spatial input data to decide when, where, and how to capture details that impact future production forecasts for the one or more wells [i.e. and generating an updated earth model based on updated outputs from the machine learning model].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Sun and Che. The motivation is the same as claim 1.
Regarding claim 5 and analogous claims 12 and 19:
The combination of Sun and Che teach the method of claim 1.
Sun teaches:
1. wherein the wellbore attributes comprise one or more of: gamma, resistivity, neutron, density, compressional, shear, or elastic properties.
(Sun, col. 7:34-38)
“Additionally, the spatial features may include laboratory data including routine core or special core analysis, chemical additives associated with the well including surfactant, microproppants, friction reducer, clay control, and other information [i.e. wherein the wellbore attributes comprise one or more of: shear].”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Sun and Che. The motivation is the same as claim 1.
Regarding claim 6 and analogous claims 13 and 20:
The combination of Sun and Wu teach the method of claim 1.
Sun teaches:
1. wherein the machine learning model comprises a neural network.
(Sun, col. 8: 31-35)
“FIG. 2 illustrates a well productivity system 201 according to an example. The well productivity system 201 can be implemented for forecasting well productivity accurately and efficiently, in real-time, using deep learning neural networks as described herein.”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Sun and Wu. The motivation is the same as claim 1.
Claims 4, 11, and 18 are rejected under 35 U.S.C. 103 as being unpatentable over US Patent 11,428,078 (Sun et al; Sun) in view of US Pre-Grant 2018/0023385 (Bang et al; Bang), further in view of the machine translation of Chinese Patent 110109180 A (Che et al; Che).
Regarding claim 4 and analogous claims 11 and 18:
The combination of Sun and Wu teach the method of claim 1.
Sun and Wu do not explicitly teach:
1. wherein the plurality of directions comprises one or more directions selected from a list comprising: up, down, left, right, forward, and backward.
Che teaches:
1. wherein the plurality of directions comprises one or more directions selected from a list comprising: up, down, left, right, forward, and backward.
(Che, pg. 6, ¶5)
“The amplitude logarithmic display method of the azimuth acoustic wave cementing quality logging of the present invention is used to calculate the relative value between the first wave amplitude value and the first wave amplitude maximum value of eight different waveforms of eight different directions at the same effective depth point. And multiplying the logarithmic value of the average value of the first wave amplitudes of the eight waveforms in eight different directions.”
Examiner notes that, given the physical existence of a depth point, one of those directions would at least have to be up or down, and only one of the list needs to be taught to teach the limitation.
One of ordinary skill in the art, at the time the invention was filed, would have been motivated to modify Sun and Wu with Che. The motivation is to improve the system by incorporating various different standardized directions to better understand the tortuosity of a given area in a well.
Claims 7 and 14 are rejected under 35 U.S.C. 103 as being unpatentable over US Patent 11,428,078 (Sun et al; Sun) in view of US Pre-Grant 2018/0023385 (Bang et al; Bang), still further in view of “Rock Property Data Volumes From Well Logs,” Search and Discovery #40268 (Denham et al; Denham).
Regarding claim 7 and analogous claim 14:
The combination of Sun and Wu teach the method of claim 1.
The combination of Sun and Wu do not explicitly teach:
1. wherein the adjacent waveform data corresponds to a given radius with respect to each depth point of the plurality of depth points.
(Denham, pg. 2, Abstract, ¶2)
“Wells within a specified radius were used, weighted inversely with distance, and well samples
over a limited depth range, weighted inversely with depth difference from the sample depth. The
computed three-dimensional array was written to disk in SEG Y format, and loaded into a seismic
interpretation system.”
One of ordinary skill in the art, at the time the invention was filed, would have been motivated
to include the teaching of Denham to modify the combination of Sun and Che. The motivation is to make “well data resemble seismic data, not just a single-trace, but a complete volume (Denham, pg. 2, Abstract, ¶1).”
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to PAUL JUSTIN BREENE whose telephone number is (571)272-6320. 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, Michael J Huntley can be reached on 303-297-4307. 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.
/P.J.B./ Examiner, Art Unit 2129
/MICHAEL J HUNTLEY/Supervisory Patent Examiner, Art Unit 2129