Ex Parte Tang et al

Patent Trial and Appeal Board

Jun 29, 2018

13624039 (P.T.A.B. Jun. 29, 2018)

UNITED STA TES p A TENT AND TRADEMARK OFFICE
APPLICATION NO. FILING DATE FIRST NAMED INVENTOR
13/624,039 09/21/2012 Liang Tang
54549 7590 07/03/2018
CARLSON, GASKEY & OLDS/PRATT & WHITNEY
400 West Maple Road
Suite 350
Birmingham, MI 48009
UNITED STATES DEPARTMENT OF COMMERCE
United States Patent and Trademark Office
Address: COMMISSIONER FOR PATENTS
P.O. Box 1450
Alexandria, Virginia 22313-1450
www .uspto.gov
ATTORNEY DOCKET NO. CONFIRMATION NO.
63017US01;
67097-2011PUS1
1055
EXAMINER
DUNN, DARRIN D
ART UNIT PAPER NUMBER
2126
NOTIFICATION DATE DELIVERY MODE
07 /03/2018 ELECTRONIC
Please find below and/or attached an Office communication concerning this application or proceeding.
The time period for reply, if any, is set in the attached communication.
Notice of the Office communication was sent electronically on above-indicated "Notification Date" to the
following e-mail address( es):
ptodocket@cgolaw.com
PTOL-90A (Rev. 04/07)
UNITED STATES PATENT AND TRADEMARK OFFICE
BEFORE THE PATENT TRIAL AND APPEAL BOARD
Ex parte LIANG TANG and ALLAN J. VOLPONI
Appeal2017-009709
Application 13/624,039
Technology Center 2100
Before JAMES R. HUGHES, DENISE M. POTHIER, and
STEVEN M. AMUNDSON, Administrative Patent Judges.
POTHIER, Administrative Patent Judge.
DECISION ON APPEAL
Appeal2017-009709
Application 13/624,039
STATEMENT OF THE CASE
Appellants 1' 2 appeal under 35 U.S.C. § 134(a) from the Examiner's
rejection of claims 1, 4, 6-9, and 13-25. Appeal Br. 2. Claims 2, 3, 5, and
10-12 have been canceled. See Appeal Br. 2; June 8, 2016 Response to
Non-Final Action 2-3. We have jurisdiction under 35 U.S.C. § 6(b). We
affirm.
Invention
Appellants' invention relates to the health of a turbomachine, such as
a gas turbine engine (e.g., item 20 in Figure 1 ). Spec. i-fi-f l-2, Fig. 1.
Diagnostic information for turbomachines is error prone. Spec. i1 3. The
invention aims to merge information within a turbomachine health
management system. Spec. i1 4. For example, information collected (e.g.,
blade tip timing) from one subsystem (e.g., 70d) can be used to adjust
information (e.g., fan speed (NI)) at another subsystem (e.g., 70a) when the
sensor measuring fan speed (e.g., fan speed sensor 68c) fails. Spec. i-fi-15,
42--43, 51-52, Figs. 2, 4.
Claim 1 is reproduced below with emphasis:
1. A method of fusing information within a turbomachine
health management system, comprising:
sending sensor information collected by a first subsystem
from the first subsystem to a qualitative model of a system level
reasoner;
1 Throughout this opinion, we refer to the Final Action (Final Act.) mailed
September 22, 2016, the Appeal Brief (Appeal Br.) filed February 1, 2017,
the Examiner's Answer (Ans.) mailed May 22, 2017, and the Reply Brief
(Reply Br.) filed June 30, 2017.
2 The real party in interest is listed as United Technologies Corporation.
Appeal Br. 1.
2
Appeal2017-009709
Application 13/624,039
sending sensor information collected by a different,
second subsystem from the second subsystem to the qualitative
model of the system level reasoner; and
sending diagnostic feedback to the second subsystem
from the qualitative model of the system level reasoner, the
diagnostic feedback based on the sensor information from the
first subsystem, wherein the diagnostic feedback substitutes
sensor information collected by the second subsystem with
sensor information from the first subsystem.
The Examiner relies on the following as evidence of unpatentability:
Bellows
Kikta
Wiseman3
No vis
us 5, 122,976
US 2002/0152298 Al
US 6,598, 195 B 1
US 2009/0112519 Al
The Rejections
June 16, 1992
Oct. 17, 2002
July 22, 2003
Apr. 30, 2009
Claims 1, 4, 6-9, 13, 15-18, and 20-25 4 are rejected under 35 U.S.C.
§ 103(a) (pre-AIA) as unpatentable over Novis, Bellows, and Wiseman.
Final Act. 7-18, 20-24.
Claims 14 and 19 are rejected under 35 U.S.C. § 103(a) (pre-AIA) as
unpatentable over Novis, Bellows, Wiseman, and Kikta. Final Act. 19-20.
3 Wiseman is the second named inventor; Adibhatla is the first named
inventor. For consistency with the Examiner's discussion, we refer to this
reference as Wiseman.
4 Notably, the rejection's body does not address claims 20-25 (Final Act.
7-18) until pages 20 through 24, which occurs after the discussion of
another rejection of claims 14 and 19 (Final Act. 19-20). Although placed
after the discussion of claims 14 and 19, the discussion of claims 20-25 does
not address Kikta. Final Act. 20-24. We therefore presume the Examiner
intended to reject claims 20-25 based on Novis, Bellows, and Wiseman, as
do Appellants. See Appeal Br. 4 (including claims 20-25 in the heading of
the rejection ofNovis, Bellows, and Wiseman).
3
Appeal2017-009709
Application 13/624,039
THE OBVIOUSNESS REJECTION OVER NOVIS, BELLOWS,
AND WISEMAN
Regarding representative claim 1, 5 the Examiner finds Novis teaches
many of its recitations, turning to Bellows and Wiseman in combination
with Novis to teach or suggest the claimed "sending diagnostic feedback"
step. Final Act. 8-11 (citing Bellows 4:8-25, Abstract, Figs. 1-3 and
Wiseman 1:31-50).
Appellants argue the recited "sending diagnostic feedbac!C' step is not
taught by "the modified version ofNovis." Appeal Br. 4. Specifically,
Appellants argue Bellows does not teach sending diagnostic feedback to a
second subsystem but rather teaches communicating one of two routines to a
reheat pressure algorithm. Appeal Br. 4--5. Appellants further contend
Bellows does not teach communicating from a system level reasoner to a
subsystem, such that one skilled in the art would have known to modify
Novis's O/M interface to communicate information back to Novis's Primary
FE (feature extraction) Module 12. Appeal Br. 4--5. In Appellants' view,
Bellows teaches generally adjusting an algorithm based on sensor
information, but this does not represent a diagnostic feedback. Appeal Br. 5.
Appellants further argue the Examiner's reasoning to modify Novis as
proposed is speculative and lacks a rational basis. Reply Br. 2.
Even further, Appellants assert any such feedback is not based on the
sensor information from a first subsystem. Appeal Br. 5. As for Wiseman,
Appellants argue there is no teaching that a first subsystem is adjusted based
on information from a second subsystem. Appeal Br. 6.
5 Claims 1, 4, 20, and 21 are argued collectively. App. Br. 4--6. We select
claim 1 as representative. See 37 C.F.R. § 41.37(c)(l)(iv).
4
Appeal2017-009709
Application 13/624,039
ISSUE
Under§ 103, has the Examiner erred in rejecting claim 1 by finding
that Novis, Bellows, and Wiseman collectively would have taught or
suggested "sending diagnostic feedback to the second subsystem from the
qualitative model of the system level reasoner, the diagnostic feedback based
on the sensor information from the first subsystem"?
ANALYSIS
Based on the record before us, we are not persuaded of error in the
Examiner's rejection of independent claim 1.
Claims 1, 4, 20, and 21
One cannot show nonobviousness by attacking references individually
where rejections are based on combinations of references. See In re Keller,
642 F.2d 413, 426 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 1097
(Fed. Cir. 1986). Thus, any arguments by Appellants attacking Novis,
Bellows, or Wiseman individually (see Appeal Br. 4---6) fail to consider the
collective teachings of the references.
As the Examiner finds and explains, Bellows suggests a subsystem
feeds algorithm information to use one of two routines based on a faulty
sensor. Final Act. 8-9 (citing Bellows, Abstract, 4:8-35, Figs. 1-2). The
Examiner proposes to modify each ofNovis's PE modules to integrate
Bellows's algorithm (1) for executing or processing routines (e.g., routine 1
or routine 2) and (2) for controlling how to process data based on whether a
sensor is faulty or not. Final Act. 9-10; Ans. 3--4, 9.
Specifically, Bellows teaches a data center computer (e.g., 10) that
collects raw data and processes the data for a processing plant. Bellows,
5
Appeal2017-009709
Application 13/624,039
Abstract, Fig. 1. Such raw data is sent to a remote expert system (e.g., 16)
for sensor diagnosis of faulty sensors. Bellows, Abstract, Fig. 1. According
to Bellows, this technique shifts the burden of sensor malfunction diagnosis
and more efficiently performs such diagnosis. Bellows 3:65-7. Faulty
sensor results (e.g., confidence levels) are sent or fed back to the data center
to control sensor processing routines (e.g., 50 in Fig. 1 ), which use either the
raw data or further processed data. Bellows, Abstract, 3:48---60, 4:8-35,
Figs. 1-2. For example, in Bellows, when an output (e.g., confidence level
from an expert system) indicates the hot reheat pressure sensor has failed,
the algorithm uses interpolated data (e.g., performs routine 2 that
interpolates between a cold reheat pressure and an extraction pressure
information). Bellows 4:8-35, Fig. 2. Bellows further suggests the
communicated diagnostic value can substitute for the collected, sensor data.
Bellows 4:28-33.
We thus disagree with Appellants (Appeal Br. 4) that Bellows does
not at least suggest sending diagnostic feedback to subsystems (e.g.,
processing routines 50 in Fig. 1 ).
Furthermore, when combined with Novis as proposed, Bellows's
processing routines would be implemented within Primary FE modules (e.g.,
12). Final Act. 8-10; Ans. 3--4, 10. As such, contrary to Appellants'
assertions (Appeal Br. 4--5), the collective teachings resulting in Novis's
modified system suggest "sending diagnostic feedback to" a subsystem (e.g.,
sending confidence level information or a substituted diagnostic value to
Primary FE module 12) from a remote system (e.g., Novis's system level
reasoner (e.g., element 17)).
6
Appeal2017-009709
Application 13/624,039
Also, Novis teaches PFE modules have microprocessors to process
sensor signals and to extract primary features (e.g., using "soft" or virtual
sensors) from the processed signals. Novis i-fi-128, 30. Novis explains the
PFE' s signal processing function is calibrated, such that the primary
extracted features accurately represent the engine parameters appropriate for
further analysis. Novis i-fi-128, 30. 6 Given Novis's desire to calibrate each
PFE with accurate information (Novis i-fi-128, 30) and given Bellows's
teaching that sensors are known to be faulty (see Bellows, Abstract, 4:8-35),
one skilled in the art would have recognized and applied the above described
feedback technique in Bellows in order to calibrate Novis's signal
processing function accurately and ensure the sensor information and the
primary extracted features accurately represent engine parameters used for
further analysis. That is, Bellows's technique would indicate the accuracy of
sensor information in N ovis' s modified system and substitute sensor
information with a more correct, diagnostic value when the PFE's sensor(s)
will not represent accurately engine parameters (e.g., when a PFE's sensor is
faulty). See Bellows, Abstract, 3:48---60, 4:8-35, Figs. 1-2; see also Ans. 4
(stating the combination implements corrective action by controlling how
each module uses sensor data responsive to detecting faulty sensor data for a
particular module).
Thus, despite Appellants arguing N ovis as modified does not teach
sending diagnostic feedback to the subsystem (Appeal Br. 4--5) and Novis
communicates information to the MA Module but "not the other way
6 N ovis also suggests information from calibrated modeling 3 6A,
discriminators 35A, and assessor 35B are sent back to a PFE (e.g., from
36A, 35A,B to 32 through various steps). Novis i-fi-178-80, 83-85, Fig. 3.
7
Appeal2017-009709
Application 13/624,039
around" (Reply Br. 2), we disagree for the above reasons. The Examiner
also provides additional reasons for N ovis' s system level reasoner to send
diagnostic feedback to a second subsystem (e.g., one of modified PFEs 12 in
Novis) based on Bellows's teachings. Ans. 9-12. We therefore disagree
with Appellants (Reply Br. 2-3) that the Examiner's reasoning to modify
Novis as proposed is speculative and lacks a rational underpinning.
Even further, Appellants assert any such feedback is not based on the
sensor information from a first subsystem. Appeal Br. 5. This argument is
unavailing. The Examiner turns to both Bellows and Wiseman in
combination with Novis to teach this disputed feature. Final Act. 10-11
(citing Bellows 4:25-33 and Wiseman 1:31-50); Ans. 10-12 (citing Bellows
4:8-24 and Wiseman 4: 13-25).
As noted above, Bellows teaches outputting various types of values
based on detecting a faulty sensor, including substituting a diagnostic value
for the sensor data. Bellows 4:8-33. Bellows also teaches the alternative
routines (e.g., 1 or 2) include routines that use other sensor data (e.g., cold
reheat pressure and extraction pressure sensor data) in place of the original
sensor data (e.g., hot reheat pressure sensor). See Bellows 4:8-33; see Ans.
11-12. Additionally and to the extent Bellows does not fully teach the
disputed feedback is based on information from a first subsystem, Wiseman
teaches or suggests the recited feature. That is, Wiseman teaches replacing a
faulty sensor value with an alternate model value arrived at by using inputs
from other subsystems (e.g., compressor subsystem having a faulty
compressor inlet temperature sensor is modeled on fan subsystem's sensor
inputs, including a sensed fan inlet temperature and fan speed). Wiseman
1:31-50, 4: 12-25. We therefore disagree with Appellants (Appeal Br. 6)
8
Appeal2017-009709
Application 13/624,039
that Wiseman, when combined with the Novis/Bellows system, as
previously discussed, does not teach or suggest the diagnostic feedback is
"based on sensor information from the first subsystem" and "the diagnostic
feedback substitutes sensor information collected by the second subsystem
with sensor information from the first subsystem" as recited.
For the foregoing reasons, Appellants have not persuaded us of error
in the rejection of independent claim 1 and dependent claims 4, 20, and 21
not separately argued.
Claims 6-9, 13, 15-18, 23, and 25
For independent claims 6 and 15, Appellants present similar
arguments to those for claim 1. Appeal Br. 6-7. We are not persuaded for
reasons previously discussed.
Additionally, Appellants assert Novis fails to teach any first
subsystem and separate second subsystem. Appeal Br. 6-7. We disagree.
Novis teaches several separate subsystems (e.g., Primary FE Module 1,
Primary FE Module 2, and Primary FE Module m1) in Figure 1. Novis i-f 28,
Fig. 1 (element 12); see also Final Act. 7-8. Appellants also argue Bellows
does not teach sensors associated with different subsystems. Appeal Br. 6.
This argument attacks Bellows individually without considering the
collective teachings ofNovis and Bellows. When modified as proposed by
the Examiner, the resulting turbomachine health management system sends
diagnostic information both to a first and a second subsystem. See Final
Act. 7-10. Moreover, the resulting "diagnostic feedback" is provided from
N ovis' s "system level" as recited in claim 6 and contrary to Appellants'
assertion (Appeal Br. 6).
9
Appeal2017-009709
Application 13/624,039
Accordingly, Appellants have not persuaded us of error in the
rejection of independent claims 6 and 15 and dependent claims 7-9, 13,
16-18, 23, and 25, which are not separately argued.
Claim 22
Claim 22 depends from claim 1 and further recites "the qualitative
model substitutes the sensor information collected by the second subsystem
with sensor information from the first subsystem." Appeal Br. 12 (Claims
App'x). Appellants argue Wiseman does not teach or suggest "pertinent
function of substituting sensor information collected by a second subsystem
with sensor information collected by a first subsystem" as asserted by the
Examiner. Appeal Br. 8. This argument is unavailing for reasons already
stated.
Appellants also argue Wiseman does not teach a first subsystem
module is adjusted based on information from a second subsystem module
but rather discusses modeling sensors and determining whether a sensor is
faulty based on a comparison with a threshold. Appeal Br. 8. We are not
persuaded for the reason previously discussed, including what Wiseman (see
Wiseman 1 :31-50, 4: 12-25) teaches or suggests to an ordinarily skilled
artisan when combined with Novis and Bellows. See also Ans. 14.
Claim 24
Claim 24 depends from claim 6 and further recites "the diagnostic
feedback replaces information collected by at least one second sensor
associated with the at least one second subsystem with information collected
by the at least one first sensor at the system level." Appeal Br. 12 (Claims
App'x). Appellants argue the Examiner's interpretation of the phrase
10
Appeal2017-009709
Application 13/624,039
"system level" in the claim "is misplaced" in light of the disclosure. Appeal
Br. 8 (citing Spec. i-fi-1 44--45).
The Examiner states "[t]he broadest reasonably interpretation of a
level in light of Appellants' specification is a position within a system."
Ans. 15 (citing Fig. 2). To extent the Examiner construes the recited
"system level" with this same breadth, we agree with Appellants that this
interpretation is too broad. Figure 2 depicts "system level reasoner 72" and
shows this reasoner at a level above subsystem modules 70a-n. Spec., Fig.
2. Paragraph 44 similarly describes model 78 as a "high-level model" that
"may be considered a qualitative reasoner, system level aggregator, or
system-level reasoner" (Spec. i1 44), and that provides feedback to "lower-
level" modules 70a-n (Spec. i1 45). See also Spec. i1 53 (describing
subsystem modules as "local information" and the "system level" having
"global awareness"). In light of the disclosure, we determine one skilled in
the art would have understood the phrase "system level" to be located at the
level where the recited "system level reasoner" is found.
Nevertheless, although arguing the Examiner misconstrued the
limitation "system level," Appellants do not further assert the references fail
to teach claim 24's limitations. Appeal Br. 8. In fact, as discussed above
and as proposed, the references collectively teach or suggest diagnostic
information fed back to a second subsystem can be replaced (e.g.,
substituting a diagnostic value for the sensor data (see Bellows 4:8-33)) at
the system level reasoner (e.g., at the system level) with information
collected from first sensor information. We refer above for more details.
For the foregoing reasons, Appellants have not persuaded us of error
in the rejection of claims 22 and 24.
11
Appeal2017-009709
Application 13/624,039
THE REMAINING OBVIOUSNESS REJECTION
Claims 14 and 19 are rejected under§ 103(a) over Novis, Bellows,
Wiseman, and Kikta. Final Act. 19--24. Appellants do not separately argue
these claims or the rejection. See generally Appeal Br. We sustain the
rejection for reasons similar to those discussed above for claims 6 and 15,
from which claims 14 and 19 respectively depend.
DECISION
We affirm the Examiner's rejection of claims 1, 4, 6-9, and 13-25
under§ 103(a).
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(l )(iv).
AFFIRMED
12