COMMISSIONER'S DECISION
Sec. 2, Obviousness Monitoring Nuclear Power Plants
The concept of monitoring vibration characteristics of a component is shown
in the cited reference. This application establishes acceptable vibration
characteristics at various locations of the plant along with a power spectral
density analysis at selected locations to predict potential breakdown.
Final Action - Reversed Sec. 2; Amended claims submitted after the Hearing.
********
Patent application 248,995 (Cl. 349/29), was filed on March 29, 1976 for an
invention entitled Method And Apparatus For Automatic Abnormal Events Monitor
in Operating Plants. The inventors are Paul J. Pekrul et al, assignors to
Rockwell International Corporation. The Examiner in charge of the applica-
tion took a Final Action refusing to allow it to proceed to patent. In
reviewing the reject ion, the Patent Appeal Board held a Hearing at which the
Applicant was represented by Mr. Orleans.
The subject matter of this application is a method to monitor industrial plants
such as nuclear power plants. Vibration and pressure signal sensors are
located throughout the plant to continually monitor the various operating
components. Signals from these sensors are processed by a computer, which
compares them to signals that have been accepted as normal. If the difference
between the signals and the predetermined normal signals exceeds an acceptable
level, the computer actuates a warning light or an alarm to alert the operator,
who then takes corrective action or shuts the plant down.
In the Final Action the Examiner refused all the claims as obvious in view
of three United States patents and because they differ from the cited art
merely by the nature of the algorithm which the Applicant uses. These
patents are:
3,324,458 Jun. 6, 1967 MacArthur
3,694,637 Sep. 26, 1972 Edwin et al
3,778,347 Dec. 11, 1973 Giras et al
MacArthur relates to apparatus for monitoring computer controlled processes. A
computer checks the condition of a great number of process variables (pressure,
temperature etc.) and compares these with predetermined conditions; if the
variable is not within acceptable limits an alarm condition will be established
and presented on a visual display to the operator.
Giras shows a digital computer control system for operating a boiling water
nuclear reactor and its associated steam turbine in an electric power producing
plant.
Edwin relates to detecting tool wear by monitoring the vibrational
characteristic of a tool in use and comparing that characteristic by computer to
a preestablised standard. Figure 1 of that patent is shown here.
(See formula 1)
Apparatus 10 comprises a transducer 16 mounted on collar 18 in contact with
drill 12, which is linked with computer 22.
In the Final Action the Examiner stated (inter alia):
Applicant argues that a plant has complex vibrational patterns and
various components operate independently of other components thus
presenting a different problem from Edwin et al and that applicant
provides more than one comparison level.
The examiner agrees with applicant's statements that several
components of a plant operate independently and that applicant
provides more than one comparison level; however, the examiner does
not agree with applicant's conclusion that this presents a different
problem in kind from Edwin et al or that the extra comparison level
makes the disclosed system patentable. Each of the components of the
plant which interest applicant are monitored independently, therefore
each information channel feeding the computer presents exactly the
same type of problem as that facing Edwin et al. The choice of
comparison level or levels is a mere design choice which does not
confer patentability upon applicant's system.
With reference to claim 1 in particular:
(a) Monitoring a plant by computer is common knowledge as shown by
McArthur and Giras et al.
(b) The step of computer scanning a plurality of input data channels
is common knowledge in the art.
(c) Edwin et al show that it is known to process input vibrational
signals to produce power spectral density data, and to compare
the spectral data with test limits to determine the condition of
a component.
(d) Reading out the results of the comparison for use by an operator
is common knowledge.
Applicant's method therefore consists of common knowledge steps
combined with the Edwin et al method. The difference consists merely
of different input data and different test limits. In other words
the differences reside solely in the algorithm which is used to solve
applicant's problem.
Algorithms are non patentable subject matter and cannot be relied
upon to lend patentability to an otherwise unpatentable method.
Applicant's apparatus also differs over the art only in the nature of
the algorithm disclosed by applicant.
...
In response to the Final Action the Applicant stated (in part):
...
It is important to note that the invention does not reside in a
mere aggregation, for a mere aggregation of, say, the system of Edwin
et al would consist of a plurality of such simple systems each with
its own monitor. Applicant has provided for a complex plant, a
unitary monitoring system having an appropriate number of channels
and a central scanning system with data processing unit.
Nor would it be obvious to provide such an aggregation with a
central scanning system as well as a two-level control for each
channel, since to do so would require an appreciation of the problem
to which the invention is addressed and an appreciation of the
advantages to be gained.
Applicant's teachings make the problem clear and indicate the
advantages to be gained. However, there is nothing in the cited
prior art to indicate that the problem was even recognized, and there
is certainly nothing to point towards its recognition and solution.
Admittedly, referring to the Examiner's comments made in the
"Final Action", monitoring a plant by computer is known.
Admittedly, computer scanning a plurality of input data channels
is known.
Admittedly, Edwin et al shows that it is known to process input
vibrational signals to produce power spectral density data, and to
compare the spectral data with test limits to indicate the condition
of a component.
Admittedly, reading out the results of a comparison for use by
an operator is common knowledge.
However, these elements of prior art referred to by the Examiner
are taken from quite different contexts and there can be no
justification whatsoever for making a "mosaic" of them. It is quite
impermissible to show by analysing an invention that each of its
elements is separately known.
The Examiner refers to "common knowledge in the art", but does
not indicate which particular art he is referring to. The references
are taken from different arts. But in any case, it has to be noted
that there is nothing in the cited prior art to indicate any reason
for combining the references in the manner suggested, or that such
combination would result in applicant's solution to the particular
problem, or even that the particular problem had ever been recognized
previously.
The Examiner makes a blanket rejection of all the claims,
apparently without recognizing that the claims are not all of the
same scope. It would therefore be difficult to discuss the
Examiner's objections on a claim by claim basis. However, we wish to
stress that the present submission urges allowance of all the claims,
including the broad claims 1 and 11, but it is pointed out that the
remaining subsidiary claims are directed to features which the
Examiner has not discussed at all; we therefore submit that a Final
Action in respect of those claims is inappropriate.
The Examiner's reference to an "algorithm' is not understood.
There is no claim to an algorithm, and the invention is not based on
an algorithm. A computer is used in implementing the invention,
admittedly, and a computer program will involve an algorithm, but
there is nothing in the claims or the specification to suggest that
the invention resides in an algorithm as such. The invention resides
in a new monitoring process and s new monitoring apparatus in the
environment of a complex industrial plant and organization in a novel
manner to provide a novel solution to a problem which had not
previously been recognized.
The issue before the Board is whether or not the claims are patentable.
At the Hearing Mr. Orleans indicated that from his understanding of the Final
Action the rejection of the claims is on the grounds of obviousness, but found
this is somewhat obscured by other issues in which reference to algorithms and
to two United States Court decisions was presented. Since the U.S. application
issued over the same art, Mr. Orleans felt that reliance upon the two U.S. court
decisions was an additional factor for advancing his case to allowance in this
country.
On review of the application we note that "this invention relates to abnormal
event monitors, and more particularly to dynamic signal monitors for parts of
operating plants that are not readily accessible for inspection." It is
concerned with nuclear power plants in which inspection of components is
virtually impossible without shut down, so it would be desirable to continually
"seek out potential problems, analyze them as to severity and indicate what
action should be taken." Applicant describes the placing of vibration and
pressure monitoring sensors at select locations throughout the plant.
MacArthur utilizes a computer for monitoring an industrial process. Alarm
conditions are established when the actual condition does not match a
predetermined condition, such as pressure, temperature, generator out of limits
or switch position being open when it should be closed. Column 3 at line
45 ff reads as follows:
Regardless of the complexity of any particular process,
the general approach to monitoring that process will very
likely be the same. That is, simply stated, all measurable
process variables can be periodically interrogated and
compared with predetermined values in accordance with
some predetermined sequence. In response to the comparisons,
process parameters can be adjusted or, in other instances,
it may be preferable to merely notify some supervisory author-
ity or in other words, establish an alarm condition.
Giras describes the use of a digital computer system for operating a boiling
water nuclear reactor and steam turbine in an electric power plant. In this
patent the computer provides a system for controlling or operating a turbine
follow type of boiling water reactor plant with improved coordination among
the various modes of plant operation.
Edwin describes a "method and apparatus for detecting tool wear by monitoring
a vibrational characteristic of the tool while in use and comparing that
characteristic to a pre-established standard." This patent has been acknowledged
in the disclosure of this application as we find on page 4 at line 9 ff which
reads:
Monitoring the vibration energy of a cutting tool and comparing
it with a reference has been recognized as an effective way of
determining wear for the purpose of determining the optimum
time to change the tool. See for Example U.S. patents Nos.
3,694,637 [Edwin et al] and 3,841,149. But monitoring a single
tool is not the same problem as monitoring an operating plant.
Plants usually have complex vibrational patterns due to
various components operating independently. To complicate
things even more, some components operate independently and
unsynchronized, and some even operate intermittently.
As we understand the Edwin citation, obtaining the predetermined data of tool
wear analysis is easily achieved because of the limited number of variables
involved .
It is the Applicant's position that when a complex plant is first built from
components supplied by one or more manufacturers, and in a physical environment
which has no exact precedent, then the exact behavioral characteristics of the
various components are not known. He argues that since "one does not know
exactly what the critical levels are" his system enables adjustment of the
reference data according to operational experience.
While agreeing with the applicant's statements that several components of the
plant operate independently and that the applicant provides more than one
comparing level, the Final Action disagrees with the "applicant's conclusion
that this presents a different problem of the kind Edwin et al faced or that
the extra comparison level makes the disclosed system patentable". There is
no doubt that the concept of monitoring vibration characteristics is taught
by Edwin. However, this patent is concerned with tool wear vibration
characteristics in which tool change time is indicated by analyzing the change
in power frequency distribution which occurs in the "vibration acceleration
signal produced by a test tool over the life thereof." Computer monitoring
of a process by comparing with reestablished limits such as temperature,
pressure and switch position is found in MacArthur and Giras, but they are
not concerned with any vibration characteristics. We believe that the progress
the applicant has made resides in determining the acceptable vibration
(pre-established) levels for a plant, along with the power spectral density analysis
at selected points for predicting potential breakdown. Consequently we find the
subject matter of the application is not obvious in view of the cited patents.
After looking at the independent claims on file in the application we found
that their terminology lacked certain necessary characteristics to distin-
guish them over the cited art. We informed Mr. Orleans by telephone and
on July 6, 1983 and August 3, 1983 he submitted proposed amendments to the
claims of which amended claim 1 now reads as follows:
In an operating plant having significant background noise
in time dependent fluctuating signals derived from sensors
placed at selected points for continually monitoring the
operating vibration conditions of system components, a
method for scanning in real-time separate signal condition-
ing channels, one for each of said signals, to find potential
malfunctions, draw conclusions as to their severity and
indicating to an operator what act ion to take comprising
the steps of:
establishing predetermined sets of vibration frequency
dependent limits on the basis of a stored table based
upon prior experience as to malfunctions and known
characteristic spectra of operating components;
selecting each channel in sequence for spectral analysis;
processing the signal of each channel selected to produce
power spectral density data at predetermined frequencies
over a predetermined frequency range as a simple Fourier
transform;
comparing said power spectral density data of each channel
with said established predetermined sets of vibration fre-
quency dependent limits, each set consisting of at least
two limits, one for a condition requiring caution and
another for an alarm condition requiring more direct action
by the operator; and
indicating to the operator the condition of plant components
associated with each channel and the action to be taken as
a function of which set of limits and which limit of the set
is exceeded by said power spectral data.
It appears to the Board that this claim overcomes our objections to the claims
on file and also differentiates this subject matter from the cited art.
Another objection made in the Final Action was the "applicant's method differs
from the cited art merely by the nature of the algorithm which the applicant
uses." This objection relies on the examiner's assessment that the subject
matter of the application is obvious in view of the cited references. As we
have indicated, the applicant's advance in the art is in determining accept-
able vibration levels for a plant, along with power spectral density analysis
at selected points and we are guided by the view expressed in Schlumberger
that the use of a "computer to implement discovery does not change the nature
of that discovery."
In summary, on the art of record before us, we recommend withdrawal of the
objections made in the Final Action and acceptance of proposed amendments to
claims 1 and 11.
M.G. Brown S.D. Kot
Acting Chairman Member
Patent Appeal Board
I concur with the reasoning of the Patent Appeal Board. Accordingly
I am remanding the application for prosecution in accordance with the
decision.
J.H.A. Gariepy Agent for Applicant
Commissioner of Patents Ridout & Maybee
Suite 2300, Richmond-Adelaide Centre
101 Richmond St. W.
Toronto, Ontario
MSH 2J7
Dated at Hull, Quebec
this 25th. day of November, 1983