Ex Parte Mays et al

Board of Patent Appeals and Interferences

Sep 23, 2009

10278668 (B.P.A.I. Sep. 23, 2009)

UNITED STATES PATENT AND TRADEMARK OFFICE
____________

BEFORE THE BOARD OF PATENT APPEALS
AND INTERFERENCES
____________

Ex parte THOMAS GILMORE MAYS
and JOSEPH MCCLINTOCK KUNKEL III
____________

Appeal 2008-005492
Application 10/278,668
Technology Center 2100
____________

Decided: September 23, 2009
____________

Before JOHN A. JEFFERY, THU A. DANG, and STEPHEN C. SIU,
Administrative Patent Judges.

JEFFERY, Administrative Patent Judge.

DECISION ON APPEAL
Appellants appeal under 35 U.S.C. § 134(a) from the Examiner’s
rejection of claims 1-20, 23, and 24. We have jurisdiction under 35 U.S.C.
§ 6(b). We affirm.

Appeal 2008-005492
Application 10/278,668

2
STATEMENT OF THE CASE
Appellants invented a method for operating a hydrocarbon or
chemical production facility including mathematically modeling the facility.
The model is optimized with both linear and non-linear solvers such that (1)
a linear solver models behavior at the plant level, and (2) a non-linear solver
models behavior at the unit level.1 Claim 1 is illustrative with the key
disputed limitation emphasized:
1. A method for operating a hydrocarbon or chemical production
facility, comprising;

mathematically modeling the facility;

optimizing the mathematic model with a combination of linear and
non-linear solvers, wherein the linear solver models behavior at the plant
level and the non-linear solver models behavior at the unit level;

generating one or more product recipes or operating set points based
upon the optimized solution; and

storing the product recipes or operating set points in a spreadsheet or
database.

The Examiner relies on the following as evidence of unpatentability:
Meystel US 6,102,958 Aug. 15, 2000

Bechtel Corp., Process Industry Modeling System (PIMS) User’s Manual,
Ver. 8.00, Oct. 1995 (“Bechtel”).

J.M. Pinto & L.F.L. Moro, A Planning Model for Petroleum Refineries,
Braz. J. Chem. Eng’g, Vol. 17, No. 4-7, 2000 (“Pinto”).

1 See generally Abstract; Spec. ¶¶ 0014-15, 0020-23; Fig. 2.
Appeal 2008-005492
Application 10/278,668

3
THE REJECTION
The Examiner rejected claims 1-20, 23, and 24 under 35 U.S.C.
§ 103(a) as unpatentable over Meystel, Bechtel, and Pinto. Ans. 3-13.
Rather than repeat the arguments of Appellants or the Examiner, we
refer to the Briefs and the Answer2 for their respective details. In this
decision, we have considered only those arguments actually made by
Appellants. Arguments which Appellants could have made but did not make
in the Briefs have not been considered and are deemed to be waived. See 37
C.F.R. § 41.37(c)(1)(vii).
Regarding representative claim 1,3 the Examiner finds that Meystel
discloses a method for operating a hydrocarbon or chemical production
facility that mathematically models the facility. Ans. 3. According to the
Examiner, Meystel discloses all claimed subject matter except for
optimizing the mathematic model with linear and non-linear solvers that
model behavior at the plant and unit levels, respectively. The Examiner,
however, cites (1) Bechtel for teaching optimizing such a model with linear
and non-linear solvers, and (2) Pinto for using linear and non-linear solvers

2 Throughout this opinion, we refer to (1) the Appeal Brief filed March 19,
2007 (supplemented May 31, 2007); (2) the latest Examiner’s Answer
mailed March 17, 2008 (vacating the earlier Answer mailed September 26,
2007 (Ans. 1)); and (3) the latest Reply Brief filed May 13, 2008.

3 Appellants argue claims 1-20, 23, and 24 together as a group. See App. Br.
4-6. Accordingly, we select claim 1 as representative. See 37 C.F.R.
§ 41.37(c)(1)(vii).

Appeal 2008-005492
Application 10/278,668

4
to model plant and unit level behavior, respectively. Based on these
teachings, the Examiner concludes that the recited optimization with linear
and non-linear solvers would have been obvious. Ans. 7.
In reaching this conclusion, the Examiner also notes that it was known
at the time of the invention to model plant production facilities as follows:
(1) non-linearly at the unit level;
(2) non-linearly at the plant level;
(3) linearly at the unit level;
(4) linearly at the plant level; and
(5) linearly at the plant level with certain components calculated non-
linearly.
Ans. 8, 18, 19, 21, and 22.
These five techniques are said to constitute a finite number of
identified, predictable solutions. Id. As such, the Examiner reasons, it
would have been obvious to try these known techniques to obtain a method
where (1) linear solvers model behavior at the plant level, and (2) non-linear
solvers model behavior at the unit level. Id. The Examiner reaches this
conclusion emphasizing the recognized need in the art for refinery
production planning software tools that account for complex process models
and/or non-linear properties. Id.
Appellants argue that there is no motivation to combine the references
as the Examiner proposes. Appellants contend that since Meystel uses his
own mathematical model, there is no motivation to use Bechtel’s
mathematical modelling software. App. Br. 4 and 5; Reply Br. 1, 4, and 5.
Appellants add that there is no suggestion to combine Pinto’s non-linear
optimization techniques with Bechtel since Pinto is said to characterize
Appeal 2008-005492
Application 10/278,668

5
Bechtel’s model as “simplistic,” and comprises linear relations that do not
account for more complex process models and/or nonlinear mixing
properties. App. Br. 4; Reply Br. 2 and 5.
The issues before us, then, are as follows:

ISSUES
Under § 103, have Appellants shown that the Examiner erred in
rejecting claim 1 by combining the teachings of Meystel, Bechtel, and Pinto
to arrive at the claimed invention?
This issue turns on the following questions:
(1) Would it have been obvious to try various known modelling
techniques to obtain a method where (a) a linear solver models behavior at
the plant level, and (b) a non-linear solver models behavior at the unit level,
as claimed?
(2) Does the cited prior art teach away from this approach?

FINDINGS OF FACT
The record supports the following findings of fact (FF) by a
preponderance of the evidence:
Meystel
1. Meystel discloses a multiresolutional decision support system
(MDSS) for plant performance enhancement that determines optimal
trajectories (input controls) using multiresolutional analysis of acquired data.
Meystel, Abstract; col. 2, l. 66 − col. 3, l. 1; col. 8, ll. 30-32.

Appeal 2008-005492
Application 10/278,668

6
2. Meystel’s system does not use a predetermined mathematical
model or algorithm to define the process in terms of multiple variables, but
rather acquires system data and stores the data in a multiresolutional data
structure. A knowledge base is created that can be searched at varying
levels of resolution to determine optimal process trajectories. Meystel,
Abstract; col. 3, ll. 2-8.
3. “This [knowledge] base can be called ‘a model’ in a very general
sense.” Meystel, col. 3, ll. 9-10.
4. Meystel’s MDSS allows using a distributed modelling technique
with information acquired and embedded into the model. Meystel, col. 8, ll.
60-63.
5. Meystel’s modelling techniques are applicable to various process
systems including chemical facilities and refineries. Meystel, col. 28, ll. 60-
64.

Bechtel
6. Bechtel is a user manual for the Process Industry Modeling System
(PIMS). The manual indicates “Version 8.00” on the cover page and is
dated October 1995. Bechtel, cover page.
7. PIMS is a computer system that uses a “Linear Programming” (LP)
technique to optimize the configuration and operation of process industry
plants, particularly in the petroleum refining and chemical industries.
Bechtel, at 1.
8. PIMS can be used in a “very wide variety of situations” including
(1) evaluating alternative feed stocks; (2) sizing plant units in grass roots
studies; (3) organizing product mix for a given feed slate, etc. Bechtel, at 1.
Appeal 2008-005492
Application 10/278,668

7
9. Bechtel provides distributive and generalized non-linear recursion
which can address various model non-linearities, such as “pooling”
problems, process yield variations, and utility usage with unit throughput.
Bechtel, at 4, 57-58.
10. Bechtel’s non-linear recursion capabilities are controlled via
various tables including “NONLIN” and “CURVE.” These two tables must
be present to use this feature. Bechtel, at 57-58.
11. The “NONLIN” table is used to identify matrix coefficients
whose values are some non-linear function of a column activity. Exemplary
columns can indicate purchases, sales, and submodel activity (e.g., utility
consumption of electricity in the cat cracker). Bechtel, at 59.

Pinto
12. Pinto discloses a non-linear planning model for refinery
production. The model represents a petroleum refinery and allows
implementing non-linear process models and blending relations. Pinto, at 1,
14, 15.
13. Pinto notes that existing refinery production planning software
(e.g., PIMS - Bechtel (1993)) is “based on very simple models that are
mainly composed of linear relations. The production plans generated by
these tools are interpreted as general trends as they do not take into account
more complex process models and/or non-linear mixing properties.” Pinto,
at 2.
14. Pinto’s planning model includes representing various “processing
units” which include variables with non-linear attributes (e.g., “feed
properties” and “unit operating variables”). Pinto, at 2-5, 15.
Appeal 2008-005492
Application 10/278,668

8
15. Process unit optimizers based on non-linear complex models that
determine optimal values for process operating variables have become
increasingly popular. They are, however, restricted to only a portion of the
plant. Pinto, at 2.

PRINCIPLES OF LAW
In rejecting claims under 35 U.S.C. § 103, it is incumbent upon the
Examiner to establish a factual basis to support the legal conclusion of
obviousness. See In re Fine, 837 F.2d 1071, 1073 (Fed. Cir. 1988). If the
Examiner’s burden is met, the burden then shifts to the Appellants to
overcome the prima facie case with argument and/or evidence. Obviousness
is then determined on the basis of the evidence as a whole and the relative
persuasiveness of the arguments. See In re Oetiker, 977 F.2d 1443, 1445
(Fed. Cir. 1992).
According to the U.S. Supreme Court:
When there is a design need or market pressure to solve a
problem and there are a finite number of identified, predictable
solutions, a person of ordinary skill has good reason to pursue
the known options within his or her technical grasp. If this
leads to the anticipated success, it is likely the product not of
innovation but of ordinary skill and common sense. In that
instance the fact that a combination was obvious to try might
show that it was obvious under § 103.

KSR Int’l v. Teleflex, Inc., 550 U.S. 398, 421 (2007).
“[W]hen the prior art teaches away from combining certain known
elements, discovery of a successful means of combining them is more likely
to be nonobvious.” Id. at 416.
Appeal 2008-005492
Application 10/278,668

9
“A reference may be said to teach away when a person of ordinary
skill, upon reading the reference, would be discouraged from following the
path set out in the reference, or would be led in a direction divergent from
the path that was taken by the applicant.” In re Kahn, 441 F.3d 977, 990
(Fed. Cir. 2006) (citations and internal quotation marks omitted).

ANALYSIS
The dispute before us hinges on whether Appellants have shown that
the Examiner’s “obvious to try” rationale in combining the teachings of the
cited references to arrive at the invention of claim 1 is erroneous. That is,
based on the record before us, would it have been obvious to try various
known modelling techniques to obtain an optimized model where (1) a linear
solver models behavior at the plant level, and (2) a non-linear solver models
behavior at the unit level? For the following reasons, we answer “yes” to
this question.
The Examiner justifies the obvious-to-try rationale by noting that five
different techniques were known at the time of the invention to model plant
production facilities. Ans. 8, 18, 19, 21, and 22. These known techniques
are summarized below:
Appeal 2008-005492
Application 10/278,668

10

Modelling Technique Level

Taught or Suggested
in Prior Art Reference

non-linear unit

Pinto

non-linear plant

Pinto

linear unit

Bechtel

linear plant

Bechtel

linear (with certain
components calculated
non-linearly)
plant Bechtel

Table 1: Summary of Known Modelling Techniques Indicated by the
Examiner

The Examiner characterizes these five techniques as a finite number
of identified, predictable solutions (Id.), and we agree. As the table above
illustrates, two of these solutions (i.e., in Rows 1 and 4) are pertinent to the
disputed limitations of claim 1. That is, the Examiner relies on two
identified solutions as obvious to try: (1) Bechtel’s linear solver to model
behavior at the plant level, and (2) Pinto’s non-linear solver to model
behavior at the unit level.

Appeal 2008-005492
Application 10/278,668

11
These solutions are amply supported by Bechtel and Pinto. First,
Bechtel’s PIMS uses linear programming techniques to optimize the
configuration and operation of process industry plants, particularly in the
petroleum refining and chemical industries. FF 7. As such, Bechtel at least
suggests using a linear solver to model behavior at the plant level. See id.
Second, Pinto reasonably suggests using a non-linear solver to model
behavior at the unit level as the Examiner indicates, particularly in view of
Pinto’s representation of various “processing units” that include variables
with non-linear attributes (FF 14) and various non-linear modelling
techniques in connection with refinery production (FF 12).
Given the five identified, predictable solutions summarized above, we
see no reason why ordinarily skilled artisans could not have pursued these
known options to arrive at the particular recited combination of linear and
non-linear solvers that model behavior at the plant and unit levels,
respectively. See KSR, 550 U.S. at 421. That is, we see no reason why it
would not have been obvious to try the two particular known solvers in
Rows 1 and 4 of Table 1 to optimize the mathematical model as claimed.
See id.
We reach this conclusion noting that Pinto emphasizes the increased
popularity of process unit optimizers based on non-linear complex models.
FF 15. In view of this industry trend, we agree with the Examiner (Ans. 7,
18, and 21) that there was a recognized need in the industry for software
tools that accounted for these non-linear models. As such, this trend
evidences a design need or market pressure to solve the optimization

Appeal 2008-005492
Application 10/278,668

12
problem using at least non-linear solutions, and therefore further bolsters our
conclusion that skilled artisans would have pursued the known linear and
non-linear options within their technical grasp to achieve this end. See KSR,
550 U.S. at 421.
Appellants’ argument that Pinto allegedly teaches away from its
combination with Bechtel (App. Br. 4; Reply Br. 2, 5) is unavailing. First,
although Pinto states that that Bechtel’s PIMS is based on simple models
mainly composed of linear relations (FF 13), this statement appears to refer
to an earlier version of Bechtel’s software from 1993 (see id.), and not
necessarily the version described in the cited PIMS user manual which is
dated October 1995. Compare FF 13 with FF 6.
The Examiner’s point in this regard (Ans. 17) is well taken. But given
the limited information regarding these versions on the record before us, we
cannot say how and in what manner they functionally differ from each
other—if at all. As such, the Examiner’s theory that it is “possible and quite
likely” that the 1993 version of Bechtel’s software did not disclose the
“NONLIN” and “CURVE” functions in Bechtel’s 1995 user manual (see FF
10-11) (Ans. 17-18) is merely speculative without evidentiary support.
Nevertheless, Appellants’ reliance on a statement made in Pinto about
the simplicity of an apparently earlier version Bechtel’s PIMS software as
purportedly teaching away from combining its teachings with a 1995 version
of that software significantly diminishes the argument’s persuasiveness.
Despite Appellants’ contentions to the contrary (Reply Br. 3), once the
Examiner noted this discrepancy, it is Appellants—not the Examiner—who
have the burden to show with evidentiary support that any discrepancies
between the 1993 and 1995 versions of Bechtel’s software do not materially
Appeal 2008-005492
Application 10/278,668

13
affect the relied-upon passage from Pinto. Appellants, however, have
provided no such evidence in this regard, apart from asserting that Pinto’s
“statement is correct with regard to the PIMS user manual cited by the
Examiner.” Reply Br. 3. Such conclusory statements fall well short of
resolving the issue raised by the Examiner pertaining to the different
versions of Bechtel’s software.
In any event, even if we assume, without deciding, that Pinto’s
statement does apply to the cited Bechtel 1995 user manual, we still are not
persuaded that Pinto teaches away from its combination with Bechtel.
Although Pinto characterizes Bechtel’s linear modelling techniques as
simplistic (FF 13), this hardly would discourage skilled artisans from an
approach that enhanced these linear techniques with additional non-linear
capabilities such as those described in Pinto. See Kahn, 441 F.3d at 990.
Such simplicity could actually be advantageous for certain processes at a
macro level (e.g., a plant level as described in Bechtel) since the level of
complexity required to model these activities may not require the more
complex non-linear techniques needed at the unit level. Nevertheless, in
light of Pinto, skilled artisans would have ample reason to provide a non-
linear capability for certain processes at the unit level, particularly in light of
the clear trend in the industry towards non-linear modelling processes. See
FF 12-15.
In short, selecting particular linear and non-linear optimization
techniques would have been an engineering tradeoff given their relative
advantages and disadvantages—an engineering decision well within the
Appeal 2008-005492
Application 10/278,668

14
level of skilled artisans. We therefore see no reason why these techniques
could not have been used together. As such, we do not find that Pinto
teaches away from its combination with Bechtel.
For the foregoing reasons, we conclude that it would have been
obvious to try various known modelling techniques to obtain an optimized
model where (1) a linear solver models behavior at the plant level, and (2) a
non-linear solver models behavior at the unit level in view of Bechtel and
Pinto.
Although we find the Examiner’s reliance on Meystel merely
cumulative to the teachings of Bechtel and Pinto in this regard, we
nonetheless see no reason why Meystel could not have been combined with
Bechtel and Pinto to arrive at the claimed invention as the Examiner
proposes. To be sure, Meystel uses a particular modelling technique for
chemical facilities and refineries that does not use a predetermined
mathematical model or algorithm to define the process in terms of multiple
variables, but rather acquires system data and stores the data in a
multiresolutional data structure. FF 1, 2, and 5.
But Appellants have provided no evidence on this record proving that
skilled artisans would be incapable of combining these various modelling
techniques together as the Examiner proposes. At a minimum, such a
combination would provide an enhanced system with diverse modelling
schemes (i.e., a system based on a multiresolutional knowledge base as in
Meystel (FF 1-4)) along with the linear and non-linear modelling capabilities
of Bechtel (FF 6-11) and Pinto (FF 12-15).

Appeal 2008-005492
Application 10/278,668

15
Although the Examiner’s theory that it is “possible” that Meystel’s
mathematical modelling system is identical or similar to Bechtel (Ans. 20) is
merely speculative, we nonetheless see no reason why the various modelling
techniques in the cited references could not have been combined to enhance
the system with diverse modelling capabilities. Such a combination is
tantamount to the predictable use of prior art elements according to their
established functions—an obvious improvement. See KSR, 550 U.S. at 417.
We therefore find that the Examiner’s rationale for combining the cited
references together is supported by articulated reasoning with some rational
underpinning to justify the Examiner’s obviousness conclusion. See KSR,
550 U.S. at 418.
For the foregoing reasons, Appellants have not persuaded us of error
in the Examiner’s rejection of representative claim 1. Therefore, we will
sustain the Examiner’s rejection of that claim, and claims 2-20, 23, and 24
which fall with claim 1.

CONCLUSION
Appellants have not shown that the Examiner erred in rejecting claims
1-20, 23, and 24 under § 103.

ORDER
The Examiner’s decision rejecting claims 1-20, 23, and 24 is affirmed.

Appeal 2008-005492
Application 10/278,668

16
No time period for taking any subsequent action in connection with
this appeal may be extended under 37 C.F.R. § 1.136(a)(1)(iv).

AFFIRMED

pgc

FINA TECHNOLOGY INC
PO BOX 674412
HOUSTON TX 77267-4412