U.S. patent number 3,927,308 [Application Number 05/475,987] was granted by the patent office on 1975-12-16 for monitor and results computer system.
This patent grant is currently assigned to Ebasco Services Incorporated. Invention is credited to Betty L. Christy, William A. Summers, Joseph V. Sweeney.
United States Patent |
3,927,308 |
Summers , et al. |
December 16, 1975 |
**Please see images for:
( Certificate of Correction ) ** |
Monitor and results computer system
Abstract
System used for aiding the operation of processes occurring in a
power generation plant is disclosed which includes a logging
technique for transforming input data from various pickup points
into a universal, named system variable format usable by all system
functions. Specified changes in time or in any system variable can
cause a given log to be triggered into a storing state, or an
outputting state wherein the stored data is outputted by a printer
in a page oriented log print-out format. Also included in the
system is a sequence of events interrupt technique wherein the
detection devices located within the protective circuits are
corrected for their respective time delays between the sensing time
of a measured condition and the actuation time of the associated
switching contacts. The correction is achieved by storing the
predetermined values of the time delays associated with each of the
detection devices, and subtracting such stored time delay values
from the system detected times of actuation of the respective
switching contacts, thereby deriving a corrected sequence of events
indicating the correct order of occurrence of the events at the
detection devices and, consequently, the initial cause of the
event. The system also includes an alarming technique wherein alarm
limit values are assigned for a measured system variable at a given
point in the system, and realarm values are calculated so that
realarming occurs when the measured system variable departs from
the last alarmed value by a significant amount of change, thereby
employing system discretion in selecting only the most important
alarm conditions to be outputted. A CRT device is used for
displaying the alarm information and operates with a reduced line
format for those alarms acknowledged by the operator. Entry,
removal and presentation of alarm data on the CRT screen is
designed to present the data in a simple and easily understandable
manner while maximizing the amount of data presented.
Inventors: |
Summers; William A. (North
Haledon, NJ), Christy; Betty L. (Radburn, NJ), Sweeney;
Joseph V. (Manhasset, NY) |
Assignee: |
Ebasco Services Incorporated
(New York, NY)
|
Family
ID: |
26976444 |
Appl.
No.: |
05/475,987 |
Filed: |
June 3, 1974 |
Related U.S. Patent Documents
|
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
Issue Date |
|
|
308770 |
Nov 22, 1972 |
3855456 |
Dec 17, 1974 |
|
|
Current U.S.
Class: |
702/199; 327/526;
327/355 |
Current CPC
Class: |
G05B
15/02 (20130101); G05B 23/027 (20130101); G05B
23/024 (20130101); G05B 23/0283 (20130101); Y04S
10/52 (20130101) |
Current International
Class: |
G06F
7/02 (20060101); G06G 7/00 (20060101); G06G
7/12 (20060101); G06F 7/04 (20060101); G06G
007/12 (); G06F 007/04 () |
Field of
Search: |
;235/151.1,151.2,151.3,151.35,152,153AE,153BN,156 ;307/204,211,219
;318/564 ;328/117,137,147,148,154,158 ;340/146.1BE,213G ;444/1 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Dildine, Jr.; R. Stephen
Attorney, Agent or Firm: Kenyon & Kenyon Reilly Carr
& Chapin
Parent Case Text
This application is a division of U.S. Pat. Application Ser. No.
308,770, filed on Nov. 22, 1972, and now U.S. Pat. Ser. No.
3,855,456, issued on Dec. 17, 1974.
Claims
What is claimed is:
1. A method of producing a value which is most representative of a
condition determined by a plurality of physical measurement input
devices which are responsive to the condition which can change, the
method comprising the steps of:
a. receiving signals from a plurality of input devices, the signals
representing the values which are responsive to the condition;
b. averaging the value represented by each of the received signals
with the value represented by each of the other received signals to
derive a set of averages;
c. selecting a predetermined non-linear factor which corresponds to
the probable disagreement between the received signals;
d. forming a normalized average for each of the set averages by
dividing each average by the non-linear factor;
e. subtracting from the value represented by each of the received
signals the value represented by each of the other signals to
derive a set of absolute differences:
f. dividing each of the absolute differences by the non-linear
factor to provide a first set of normalized differentials;
g. selecting a predetermined non-linear power which controls the
degree of weighting for each signal as a function of its deviation
from the other received signals;
h. raising each of the first normalized differentials by the
selected non-linear power;
i. deriving a weighting factor for each average whose value is a
function of its normalized average and its normalized differential
raised by the selected non-linear power;
j. weighting each average by the weighting factor derived for that
average; and
k. forming a composite of the weighted averages to thereby derive a
value which is most representative of the condition.
2. A method in accordance with claim 1 in which the step of
receiving signals from a plurality of input devices which are
responsive to the condition comprises receiving signals
representing at least three values which are responsive to the
condition.
3. A method in accordance with claim 1 in which the step of
receiving signals from a plurality of input devices which are
responsive to the condition receiving at least two signals
representing values responsive to the condition and receiving a
signal representing a reference value which corresponds to the
value related to the normal state of the condition.
4. A method in accordance with claim 1 in which the step of
receiving signals from a plurality of input devices which are
responsive to the condition comprises receiving a reference signal
corresponding to a value which is substantially displaced in one
sense with respect to the expected value of the condition and
receiving another value displaced in the opposite sense from the
expected value for providing a plurality of values where a
plurality of values responsive to the condition are not
available.
5. A method in accordance with claim 4 in which the step of
receiving a reference signal corresponding to a value which is
substantially displaced in one sense with respect to the expected
value of the condition comprises receiving a reference signal
approaching a zero level and in which the step of receiving another
value displaced in the opposite sense from the expected value
comprises receiving an additional signal approaching an infinite
level.
6. A method in accordance with claim 1 and further comprising the
step of selecting values to be received at least a portion of which
have substantially corresponding transient response and forming the
composite of weighted averages during the presence of transients in
the values being received, the substantial correspondence in
transient response of the values being received reducing the
discrepancy of such values from the value representing the true
condition.
7. A method according to claim 1 in which the step of weighting
includes the step of multiplying each average by the weighting
factor derived for that average.
8. A method according to claim 1 in which the step of forming a
composite includes the steps of:
i. summing the weighting factors to form a composite weighting
factor;
ii. summing the averages after each average has been multiplied by
the weighting factor derived for that average; and
iii. dividing the sum of averages of step (ii); by the sum of
weighting factor of step (i).
9. A method in accordance with claim 1 in which the step of
deriving a weighting factor for each average comprises the steps
of:
dividing each of the raised normalized differentials by its
corresponding normalized average to form a set of ratios;
summing each of the ratios with a predetermined constant; and
inverting each of the summed ratios to thereby form the weighting
factors.
10. A method in accordance with claim 9 in which said predetermined
constant has a value equal to 1.
11. A method in accordance with claim 1 further comprising the step
of tuning the forming of the composite of the weighted averages,
step of tuning including:
repeatedly varying the output of one of said input devices over a
likely range of values including a value representing a signal
received from an input device which is in a failure mode while
holding the outputs of the other input devices at appropriate
levels;
for each repeated variation selected different non-linear factors
and non-linear powers;
for each repeated variation comparing the characteristic of the
composite of said weighted average with the desired characteristic
of the value of the condition;
whereby the non-linear factor and the non-linear power result in a
characteristic most representative of the desired characteristic
can be derived.
12. A method in accordance with claim 1 and further comprising the
steps of:
a. deriving the difference between the value representing each of
the received signals and the composite of said weighted averages to
form a second set of absolute differences;
b. averaging the two smallest absolute differences of said second
set to form an average difference;
c. marking the received signals associated with the smallest
differentials of the second set thereof as good;
d. determining the larger value of the average differential and the
non-linear factor, the larger value being a test factor:
e. comparing the test factor with a predetermined multiple of the
non-linear factor to determine whether the test factor is in excess
of the predetermined multiple of the non-linear factor; and
f. additionally marking all received signals suspect for
maintenance checking when the test factor is in excess of the
predetermined multiple of the non-linear factor.
13. A method in accordance with claim 12 and further comprising the
steps of:
a. selecting the largest absolute difference of the second set of
absolute differences; and
b. comparing the selected largest absolute difference with a unique
value which is a predetermined multiple of the test factor to
determine whether the received signal related to the selected
largest absolute difference of the second set is bad or has
excessive discrepancy.
14. A method in accordance with claim 13 and further comprising the
step of:
further marking a received signal with one of predetermined
designations of good, bad and having excessive discrepancy from the
value which is most representative of the condition, the step of
marking being in response to the step of comparing the selected
largest absolute difference with the unique value.
15. A method in accordance with claim 13 in which the unique value
is a predetermined first multiplier of the test factor, the product
of the first predetermined multiplier and the test factor defining
an excessive discrepancy of a received signal from the value which
is most representative of the condition.
16. A method in accordance with claim 15 and in which the step of
further marking comprises marking the received signal related to
the selected largest absolute difference of the second set as being
bad.
17. A method in accordance with claim 16 and in which the step of
further marking comprises marking the received signal related to
the selected largest absolute difference of the second set as being
suspect.
18. A method in accordance with claim 17 in which the unique value
is a predetermined second multiplier of the test factor, the
product of the second predetermined multiplier and the test factor
defining an undesirable discrepancy which is less than an excesive
discrepancy of a received signal from the value which is most
representative of the condition.
19. Apparatus for producing a value which is most representative of
a condition determined by a plurality of physical measurement input
devices which are responsive to the condition which can change, the
apparatus comprising:
a. means for receiving signals from a plurality of input devices,
the signals representing the values which are responsive to the
condition;
b. means for averaging the value represented by each of the
received signals with the value represented by each of the other
received signals to derive a set of averages;
c. means for selecting a predetermined non-linear factor which
corresponds to the probable disagreement between the received
signals;
d. means for forming a normalized average for each of the set
averages by dividing each average by the non-linear factor;
e. means for substracting from the value represented by each of the
received signals, the value represented by each of the other
signals, to derive a set of absolute differences;
f. means for dividing each of the absolute differences by the
non-linear factor to provide a first set of normalized
differentials;
g. means for selecting a predetermined non-linear power which
controls the degree of weighting for each signal as a function of
its deviation from the other received signals;
h. means for raising each of the first normalized differentials by
the selected non-linear power;
i. means for deriving a weighting factor for each average whose
value is a function of its normalized average and its normalized
differential raised by the selected non-linear power;
j. means for weighting each average by the weighting factor derived
for that average; and
k. means for forming a composite of the weighted averages to
thereby derive a value which is most representative of the
condition.
20. An apparatus in accordance with claim 19 and further
comprising:
a. means for deriving the difference between the value representing
each of the received signals and the composite of the weighted
averages to form a second set of absolute differences;
b. means for averaging the two smallest absolute differences of the
second set of absolute differences to form an average
difference;
c. means for marking the received signals associated with the
smallest differentials of the second set thereof as good;
d. means for determining the larger value of the average
differential and the non-linear factor, the larger value being a
test factor;
e. means for comparing the test factor with a predetermined
multiple of the non-linear factor to determine whether the test
factor is in excess of the predetermined multiple of the non-linear
factor; and
f. means for additionally marking all received signals suspect for
maintenance checking when the test factor is in excess of the
predetermined multiple of the non-linear factor.
21. An apparatus in accordance with claim 20 and further
comprising:
a. means for selecting the largest absolute difference of the
second set of absolute differences; and
b. means for comparing the selected largest absolute difference
with a unique value which is a predetermined multiple of the test
factor to determine whether the received signal related to the
selected largest absolute difference of the second set is bad or
has excessive discrepancy.
Description
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to a monitor and results computer
system, and more particularly to a system which provides operation,
performance and historical data for continuously aiding the
operation of a power plant.
2. Description of the Prior Art
The conventional manner of handling log information to an operator
is to have a scanning device, that is program controlled to convey
information into the memory unit of a CPU. The log is a group of
data which you identify by designating the log number. The log has
a list of points and a value. The logging device can type out the
identification of the log, time, point Identification Number(s),
and the value(s). The storage of the logging device is the piece of
paper. If one is activating the log every 5 minutes, 5 minutes
later the log program will be scheduled to gather the data for the
log. From the operator's standpoint, the human standpoint, he sees
what is going on by looking back at the lines on the piece of
paper. To accomplish this may require one typewriter per log which
is impractical.
In such state of the art printing devices used for printing log
data, a large number n are placed in a format wherein generally a
single log occupies a line on the printer. A log having a large
number, such as 100 variables, would extend across the single line
and the resultant printing paper was very wide. On the next line, a
different log was printed. As a result of this type of format, a
relatively complicated program is required with new headings being
retyped for each new time a log is outputted. It is noted that
other than the time dependent presentation of a printout, a demand
dependent or a demand-on-program type of printout could be
employed. For example, in a demand dependent set up, a user
requests a log which is then assembled in a computer and then
printed out on a typewriter. This typewriter was, according to the
state of the art, conventionally a wide-carriage typewriter
employing mechanical printing and the printed record on paper was
the memory. The mechanical typewriters are used because of their
adaptability to printing of a character at a time. On the other
hand, printers using non-impact printing techniques are perhaps
more ideally suited from the standpoint of the relative freedom
from the high maintenance associated with the impact type printers
and from the standpoint of speed. However, non-impact printers are
not suitable to the applications in the conventional log data
storage and printing systems because such systems have a printout
format which involves the writing of a character at a time. This
character at a time printout format is incompatible with the
non-impact printers because the chemical preparation and
development processes on such non-impact printers generally
requires a relatively short interval of time between the chemical
preparation and development steps of printing. Thus, when employing
the state of the art non-impact printer in the conventional log
data system, if the time interval between the printing of adjacent
line is too long, then the chemical bath would overdevelop and wash
out those lines which were not removed from development in
time.
According to some prior art log printinig formats, a new heading
comprising the various names identifications is retyped for each
new line. Also, in logging systems used therewith, a demand panel
is used to operate in conjunction with a program and special log
programs containing lists for each special log. The program
consisted of lists in the memory of the members (the names). These
lists in the computer memory consisted essentially only of the
names while a directory contained the addresses or locations of the
names in the memory. A secondary memory device, such as a drum,
contains the contents or value of data associated with the names in
the memory.
Also, in many supervisory and control systems existent in power
plants, overload and protective devices are employed. These devices
are connected within a protective circuit and, upon their
actuation, provide the system with an indication as to the source
of a problem within the power plant, such as a loss of generator
power or a boiler trip. Contact changes from the protective circuit
network, including the overload devices and protective relays, were
incorporated into computer systems and attempts to program the
computer to indicate the relays or other protective devices which
were activated and the times of activation of their switching
contacts have met with common problems.
The protective circuits commonly include a multiplicity of trip
devices, such as relays and switches. When a trip condition exists,
one or more relays or trip devices are energized and their
associated contacts are switched. For each given relay that is
activated by a plant condition, there often is one or more
auxiliary relays within the protective network that is also
activated directly or indirectly as a result of the actuation of
the first relay. For example, a first relay switch may have
switching contacts which cascade into auxiliary switches which
eventually cause a further relay to be activated. Therefore, even
though a trip condition is initially sensed at a first relay,
signals may be produced in the system which cause an auxiliary
relay(s) or trip(s) to occur which may, in turn, indirectly switch
other contacts in the system. Because of the hardware associated
with the protective system, the contacts of the auxiliary relay(s)
may switch over before the final contact that feeds the computer
from the initial relay is energized. As a result, the computer or
control system incorrectly sees the contacts of the auxiliary
relay(s) being switched ahead of the contacts of the initial relay
and interprets this as a trip caused by the function of the
auxiliary relay. The computer also records the closing or switching
of the other devices in the protective circuit and associated or
effected plant devices, and records their respective times of
occurrence of switching and the status of all the devices.
Thus, another problem existing in the prior art monitoring and
control systems for power plants is that, because of the different
delays associated with the individual relay and tripping devices
between the precise time of the occurrence of the alarm condition
or "event" and the actuation of the contacts of such devices, the
monitoring and control system indicates the incorrect cause of the
event since the system is made aware of such event only upon the
actuation of the contacts.
Conventional alarming systems, such as those used in power plants,
provide scanning devices which compare desired system
characteristics, such as a boiler temperature, to stored low and/or
high alarm set points previously designated or calculated for that
point. The alarm information status is output on printers,
typewriters and on trouble location annunciators providing audible
and/or visible warnings of an alarm condition. One problem existing
in conventional alarming systems is that a plant emergency may
result in a multiplicity of alarms being set off almost
simultaneously, such as those due to a failure of the main turbine
which in turn interacts with other equipment so as to cause other
alarms to be actuated. In some systems, in addition to both high
and low alarm limits being set for a particular device, there is
provided one or more realarm conditions which are set to occur when
a point previously in alarm is detected as having changed a
prescribed significant amount, either above or below the last
alarmed value. Even here, often this prescribed significant change,
hereinafter referred to as "delta", is normally small to detect and
follow or track the deviations of the value once it is no longer
normal. The large transients induced by the loss of the equipment
result in the occurrence of several alarms almost continuously
being set off for the given equipment and any related plant
systems.
Alarm status information is presently being output on printers,
typewriters, annunciators and CRT devices. It has been found that
the outputting of alarm information by such devices is often
confusing or unnoticed to the human understanding because of both
the large amount of information being simultaneously displayed as
well as the format by which such information is presented.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a monitor and
results computer system which is both practical and efficient from
the standpoint of computer storage cost, processing time and
outputting.
It is another object to provide a logging system which permits the
equivalent of wide page printout on narrow paper.
It is another object to provide a logging system which is able to
easily change the log members.
It is another object to provide a logging system wherein the stored
data is universally usable by all system functions.
It is another object to provide a logging system wherein the log
page information in the memory is compacted together in an easily
printable form, with minimum processing required.
It is another object to permit storage of such information in a
second level memory, instead of on paper.
It is another object to provide a logging system which can easily
process via the log name, the defined variables of that log.
It is another object to be able to output both past and present
data without the need of continuously outputting.
It is another object to provide a logging system which can make any
log available on demand, or upon the occurrence of a condition or
event.
It is another object to provide a logging system which generates
data on a page at a time output basis, thereby making effective use
of non-impact low maintenance printers.
It is another object of the present invention to provide a
monitoring and control system which produces the actual and correct
sequence of events interrupts occurring in the plant overload and
protective devices.
It is another object of the present invention to provide outputting
of alarm information in a simple and easily understandable
manner.
It is another object to provide current, pertinent alarm
information to the human operator which is determined as being the
most significant.
It is another object to provide an intelligent reduction in the
amount of alarm information presented to the human in a given
system.
It is another object to provide a CRT display of alarm information
which is easy to understand and is formatted to provide only the
necessary alarm information arranged in a logical fashion.
It is another object to provide a CRT display for alarm information
which is formatted to maximize the amount of data presented.
It is a further object of the present invention to provide a system
which compensates and corrects for failure of any of the input
devices measuring the system variables.
These and other objects are achieved by the present invention which
provides a logging system within the monitor and results computer
system wherein analog to digital input data is presented to the
computer system from monitoring devices, such as transducers and
thermocouples located at various pickup points external to the
computer system. The input data is scanned and transformed into a
universal, named system variable format usable by all system
functions. The format includes the name identification of the
variable, the value of the system variable, the quality of the data
and the time at which data is obtained. The name or member of each
log is defined by the format. Triggering conditions are defined for
the system such that specified changes in time or specified changes
in any system variable can cause triggering of a given log into one
of two active states. Generally, a Free log state represents the
log number and is a log number list without names or page
assignments. a Defined but yet inactive log state includes names
only. An active log includes both names and values correlated with
times or time intervals, and an Active outputting log state
includes not only names and values, but also a demand for the
printing or displaying of a log. If a log is changed or triggered
to the active state or to the Active outputting state, then the log
data is stored in the proper log page field at the proper index
location. The log page format is designed for use with page
oriented printers, wherein each page includes a log number, and a
log heading containing the names or identification of each member
of the log. These names are individually arranged in the heading at
the top of the page in separate columns. In addition, each log page
includes the time indication of the data on each line.
Generally, the program is designed to handle all details following
the request by a human to place or change a point in a log, without
any further human intervention. The program is designed to define
log lists, report or remove members of the lists and to allocate
pages. The program also is designed to be triggered by events or
times, to set and control printout. Also, the program is designed
to control log status. The use of the universal format with named
system variables, the formatting and storage of data in log page
field and the final format of the page oriented printout provides a
system with the ability to easily change the types of logs, i.e.,
change the log members or add to the logs.
The present invention also provides a system and method of deriving
the ordered sequence of interrupt events sensed by detection
devices connected within a protective circuit wherein such
detection devices are characterized by a time delay between the
sensing of a predetermined condition and the actuation of their
associated switching contacts, comprising means for detecting the
change of state of the respective switching contacts associated
with each of such detection devices, means for storing values of
time delays associated with each of said detection devices, such
time delays representing the known delay for a particular contact
between its switching time and the actual time of occurrence of the
event at the associated protective device, the time of actuation of
such switching contacts being detected by the system, means for
subtracting the stored time delay values from the detected times of
actuation of the switching contacts of the respective devices to
produce a corrected time of initiation of the event at the
respective devices, whereby a corrected sequence of events is
derived which indicates the correct order of occurrence of the
events at the devices and, consequently, the initial cause of the
event.
The storage and computations are carried out by the monitor and
results computer system of the present invention which receives
indications of the real time of switching of the device contacts
and applies the associated corrective time delay for the particular
contacts to produce the actual time of activation of the protective
device. The system and method according to the present invention is
thereby able to determine the actual interrupt event which
initiated the protective devices.
The present invention also provides an alarming system and method
which includes the assigning of an alarm limit value for a measured
system variable at a given point in the system, scanning such point
for the measured system variable even after such an alarm value is
detected, calculating and applying a second alarm value
representing a significant change (a delta) in the measured system
variable from the last alarm value, and outputting the system
variable information existing at each of such realarm values.
The method also includes the setting of a fixed significant change
(a delta) associated with the measured system variable of a point
whereby the second alarm and any subsequent realarms would
ordinarily be set off when the measured system variable departs
from the last alarmed value by this fixed amount. In addition, the
method includes changing this delta for a given point by selecting
a multiplier which is applied to the fixed delta to produce a
different delta which then represents the current value of the
significant change in the measured system variable. This current
delta will require the next alarming of the measured system
variable value. The use of the multiplier to calculate a
significant change of the system variable value by which the second
or any subsequent alarm is to be set off allows the application of
discretion in selecting only those very important alarm conditions
to be outputted, thereby minimizing the number of alarm
outputs.
A CRT device is provided for outputting the alarm information and
instantly displays eachh alarm, when first detected, on the next
free line above the most recent new alarm, starting at the lowest
line of the alarm area of the CRT screen and moving upwards. When
an operator acknowledges the alarm information, the alarm message
is automatically changed to a reduced format wherein the current
value, current direction, current deviation from the alarm limit
value and current duration of the alarm is updated at the scan
frequency or the calculation frequency of the measured point. When
the CRT screen is filled with alarm information, that is, the top
line of the CRT alarm area is occupied, the oldest group of alarm
messages at the bottom of the screen is transferred into a system
memory for alarm backlog information and the alarm information on
the lines above those transferred shall be compacted downwards,
thereby permitting entry of the new alarm at the first free line
above the existing displayed alarms of the alarm area.
When the value of a measured system variable at a point was
previously detected and displayed in the new alarm condition, and
such point is further detected as having changed by a significant
amount by crossing the computed delta value above or below the last
alarmed value, and such point is not on the "no alarm" or normal
side of the alarm limit, then such point will be displayed on its
present line of the CRT screen in the new alarm format until
acknowledged by the operator, at which time the alarm information
is returned to the reduced format. However, if the measured system
variable alarm message had been previously moved into the backlog
of alarms, the alarm message shall be removed from the backlog and
placed on the next free line determined in the alarm area.
It is to be understood that, as used herein, the term "named system
variable" includes a measured system variable, and vice versa.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a general block diagram of the monitor and results
computer system of the present invention;
FIG. 2 shows a printer output format for a log page, illustrative
of the present invention;
FIG. 3 is a block diagram showing the four log states;
FIG. 4 is a generalized system operation flow diagram;
FIG. 5 is a more detailed operations flow diagram of the diagram
shown in FIG. 4;
FIG. 6 is a representation of the operator/engineer panel;
FIG. 7 shows a flow chart of the operation for defining a log;
FIG. 8 shows a flow diagram of the operation for changing the
status of a log;
FIG. 9 shows a flow diagram of the operation for controlling the
collection and/or outputting of log data;
FIG. 10 shows a protective circuit connected to the computer
control portion of the system which computes the actual sequence of
events detected by overload and protective devices;
FIG. 11 shows a graphical representation of a curve plot of a
measured system variable at a point, with the set alarm values, the
set realarm values and alarmed values drawn to illustrate the
alarming system;
FIG. 12 shows a CRT screen having an alarming area for displaying
alarm information in accordance with the format of the present
invention;
FIG. 13 shows a generalized functional block diagram of the
alarming system;
FIGS. 14A, B, C, D, and E show the sequence of presentations of the
log trend;
FIGS. 14F, G, H, I and J show the sequence of presentations in a
prior art trend display;
FIG. 14A is a graphical representation of a tuned output signal
(Believed Value) as the input signals are varied;
FIG. 15B is a block diagram of a tuning system for deriving an
output (Believed) value used for system maintenance; and
FIG. 16 shows the method for deriving a Believed Value and
accordingly applying system maintenance.
DESCRIPTION OF THE PREFERRED EMBODIMENTS
FIG. 1 shows a block diagram of a monitoring and results computer
system used for aiding the operation of processes such as occurring
in power generation plants. The system is generally shown by a
block diagram. Block 10 refers to the sensing devices throughout
the plant that provide the information to the computer system in
analog, digital and/or pulse form. Block 11 represents the
interface and filter circuits for filtering and conditioning the
digital and pulse type signals and interfacing them with the
computer system either through the CPU and/or the input/output
processor. The block 12 represents the filter equipment and the
scanning devices to handle the analog signals. Block 13 includes
the control of the multiplexing for scanning. Block 14 is the
representative of the logic and arithmetic circuits associated with
the CPU. Block 15 is the programming and programming system that
directs the logic arithmetic circuits 14 and memory accesses to
perform the desired sequence of instructions yielding the
monitoring services. Block 16 represents the actual CPU internal
control. The computers used in this function are mainly interrupt
type machines and represent a small to medium level CPU with
respect to the present art. Multiple CPU' s are also utilized where
required. The systems shown usually include multiple levels of
memory and high memory transfer rates. Block 17 shows the high
speed working memory. Block 18 represents a secondary level memory
and block 19 represents tertiary level memories. Various levels of
memory are utilized depending upon the speed of response required
for the programming contained therein. Block 20 represents the
input/output processor having the logic, control and associated
memory mechanism for processing, formatting, etc., all system input
and output data, and for regulating the connections among system
information units and transfer control of information needed to be
exchanged or transmitted between external peripheral devices and
the internal memories. Block 21 represents the hard copy output
devices such as printers and/or printer plotters of various types.
Some of these devices 21 are used for permanent historical records,
others for the printing information of temporary or "tear off" use
in operations. Block 22 is a group or groups of alphanumeric
presentation devices, such as CRT's for display of information of
immediate or transient rise that doesn't require hard copy. Block
23 is a group or groups of combination alphanumeric and graphic
display types of CRT's or graphic types, for use similar to devices
22 but with pictorial capability. Block 24 represents a combination
type CRT that is used both for display and in conjunction with the
operator/engineer keyboard console requirements which are shown by
numeral 25. The block 26 represents interfacing circuits providing
translation and adaption equipment required to pass information
from the subject system to the other data systems.
Generally, the system shown in FIG. 1 permits the functioning as
will be described in detail hereinbelow.
The logging system included in the system of the present invention
is designed for use with page oriented printers. An example of the
printer output format for a page is shown in FIG. 2. In page
oriented printing, each of the system logs is identified by a
unique number indicated at numeral 30 as "Log 66," and a log
constitutes a group of data associated with such log. Each numbered
log has associated therewith an English identification indicated at
31, which is assigned from changeable programmer input equipment
and from an on-line change or compiling system. This English
identification, i.e. "Turbine Inlet Conditions," is prined on each
page in the log. Each log is made up of a series of pages, such as
up to 10 pages, with each page containing a multiplicity of system
variables, such as 16. In FIG. 1, the corners of the page define
the log number 66 and page number 1 as 66/1. Each system variable
in the page has a capacity of, for example, 40 lines of
time-sequenced or other readings.
Included in the log heading are the year, month, day of the month,
day of the week, hour, minute, seconds, and time interval between
logs for the data on the log page. Also, as indicated by the
bracketed area by numeral 32 is a list of the point numbers
together with an english description of each of these points. In
the example shown, there are 16 points respectively set forth in
the 16 columns listed across the page. As shown by the heading area
32, column 1 is a listing of times, column 2 is a listing of the
main steam at throttle, etc.
In addition to the log heading, each page includes heading and data
identification information for each variable assigned to that page
in accordance with the sample shown in FIG. 2. Thus, the page 1 of
each log may include as one system variable the time of the data on
that line. The logs can be assigned at repeat collection
frequencies extending, for example, from one second to one day.
The system specifies an organized naming system for system
variables that is constant throughout thereby simplifying the task
of selecting and requesting information from the system. The system
provides a method of grouping associated operating information,
which is organized on the basis of functional lists, referred to
herein as Log pages, so that simple requests will generate all
associated information. As will become apparent from a reading of
this specification, the Log page lists can be easily created by
non-computer trained personnel and modified by the plant operators
and engineers. The logging system is organized so that hard copy is
generated on a "page at a time" log output, thereby making
effective use of non-impact low maintenance printers. Also, the
page oriented logging system permits continual acquisition of
desired log data with the option to receive the latest page or
pages of data upon demand or upon the occurrence of a predefined
event.
This system also facilitates the generation of a graphic display of
the trends of selectable data from active log pages. A Log Trend
CRT can be used to trend information on the screen of one or more
points in any active log or logs. A trend information display can
be initiated by the demand panel to display trend information such
as, identifying heading, Pen ID, Point ID, English identification,
quality, present value, present engineering units, scales, and
others. The request for Trend Pen Recording and for CRT log trend
display can be generated and controlled from the operator's panels.
All points assigned to the operator's recorders can be identified
and the current value displayed in a specified format on the Trend
Information CRT. It is noted that the CRT display of log output
data can be used for outputting as an addition to or substitute for
a log page printer.
The page oriented printed format according to the present invention
comprises a given log content in a full page of information with
several points of information located on one line for a given time.
Thus, a request by a user for a given log will provide a page of
printed information with the present log time and the previous 30
or 40 log lines of information for the same log. This is
ccontrasted with some prior art logging formats which print out
substantially a different log on each successive line intermingling
unrelated sets of information with one another.
The logging format makes it more practical to have a larger
quantity of information in storage in memory such as the names, the
addresses or location in memory and the contents or value of the
names in memory. Of course, the amount of information contained in
memory will depend to a great extent on the trade-off between the
amount of memory available as against the processing time
limitations. Generally, with the page oriented printing technique
according to the present invention, the memory comprises a pool of
pages of information with each page of information similar in
content to the data to be printed out on the printer for a given
log. Of course, the data in memory can be assembled to a lesser
degree than that described above in which case processing and
assembling of the data into final form for a page will be
required.
In the page oriented printing format a printer may typically
accommodate 16 columns of information. Since many of the logs
contain more than 16 items of information at a given time, then the
pages of print are arranged in a fashion to accommodate a large
number of columns of information for a given log, such as 160
columns of information. In this example, 10 pages of print will be
arranged in a strip in series, with each page comprising 16 columns
of print for the same log all on a time synchronized basis. Thus,
page 1 may contain data values for a first 16 names or points of
information at various times indicated, page 2 will contain data
points 17 through 32 taken at the same time as the data on page 1,
page 3 will contain information for points 33 through 48, etc.
Further, the heading on a given page need not be changed for each
line on such page since each of the page lines contains data
associated with the points or names in the heading at the columns
on the top of the page.
As an example, assuming that a single log contains 273 characters
or 33 names with data for each point at a given time and that
single page contains 16 columns, then 3 log pages will be required
to present this data. While this ordinarily would require a very
wide paper, the page oriented printer arranged the pages in a strip
series arrangement.
The system includes a naming scheme such that related items have
one unique basic or root name throughout the system. The root name
and suffixes are also referred to as point identication or ID
number. All references to a point by the operator and the source
language programs use the basic name plus any desired suffixes. As
an example, a basic name can have the form:
Pnnnn,
where:
P = a prefix from the list below
Nnnnn = a number from 0 to 9999
The prefix can be assigned according to the following:
A (or Blank) Analog Input
B boolean/bit
C constant
D digital Input
F fast Calculation
G spare (Reserved for GROUP NUMBERS)
H spare
I summation Boolean Changes
L spare
M memory Location
P performance Calculation
R relay Output
S special Analog (Composed point)
T maximum of N Points
W minimum of N Points
X variable from Common
The suffixes applied to the basic name for indicating processing or
transformation on a point is instituted by appending the
transform-related identifier to the basic root name. Suffixes can
be multiple-level. At that time, the new point is created. The
suffix can be consistently assigned according to alpha
representations, such as A=Periodic Average, B=Periodic Minimum,
C=Continuous Running Average, D=Daily Average, F=Daily Minimum,
etc.
All analog and digital type variables are stored in the computer
memory with a quality code associated with the value which is, for
example, in range of O through 3. The quality code is generated
along with the value and is always propagated through transforms
and calculations. Data quality levels can vary between the numbers
O and 3 where the O represents good data quality and the 3
represents bad data quality. All printed or displayed messages are
adapted for including a quality identifier along with the
value.
In order to make the system operable the following needs of the
operator must be met: (1) requests that the log information be
presented immediately; (2) requests that the log content may be
changed; or (3) requests that the log information be presented at
specified time intervals, i.e., every hour. Each log may be in one
of the following four states.
Every log carries with it a built in trigger, i.e., IF statement
which is selected by the operator by, for example, activating log
44 if D103 (pump contacts) go off and if an analog value goes above
a certain defined state or value, or system activity previously
determined or changes of quality.
Referring to FIG. 3, the four log states are shown in a block
diagram form. Here, the Free (Undefined) Log State O represents the
log number without labels in the log and without page assignment.
Generally, the Free Log State O is a log number list without names.
The Defined (Inactive) Log State 1 includes names. The Active
(gathering but Non-printing) Log State 2 includes both names and
values at specified time intervals of collection. The outputting
(gathering and Active Printing) Log State 3 includes not only names
and values but is outputting as well as collecting.
In the Free State, the log is available for definition and uses no
pages. A request to print or report a Free Log shall report to the
requesting station immediately the message that the log is
"Free."
In the Defined State, the log has some system variables assigned to
its columns but it is not gathering or printing. Logs are moved
from the Free State to the Defined State by panel function and/or
the on-line background compiler. As the log is Defined, the
Operator designates the column and the ID of the system variable
assigned to that column. Column number in excess of 16 causes the
system to automatically utilize additional page(s), and the
mechanism continues for multiples of 16 up to 10 pages. Logs with
high collection/printing frequency require 2 memory pages per log
page to provide time to print. The Operator can insert blank
columns for formatting as he desires. A request to print or report
a Defined Log shall cause the standard log headings, "variable"
heading, and one line of the latest available value of the
corresponding system variables. All of this is printed on a Page
basis by the printer associated with the requesting panel, unless a
different printer is requested or is automatically substituted by
the system itself.
A Defined Log may be moved to the Active State by either panel
function or change of state to a prescribed direction of a flag bit
set by an event triggered program. In moving from Defined to Active
State, a collection frequency is specified either by the panel
function or the initiating program or the previously designed
frequency if a new one is not supplied. In the event of a conflict,
the program prevails. The Active State causes the logging system to
collect the values associated with the system variables contained
in its page list(s) at the respective frequency defined for the
owning log. As this data is collected it is written to the system
drum/disc (secondary) memory. Each log page contains, for example,
enough storage to provide heading information and forty lines of
data. Output has a total of 52 lines of l32 characters/line. Three
additional lines of storage are provided to buffer for the print
time. The logging program continues to overwrite the oldest line,
thus keeping a full Page of log available for instant printing on
demand. When the "fast" logs (i.e.: interval less than 30 seconds)
are moved to the printing state, two Pages can be alternated-- one
being printed while the other is being filed by the logging
program.
The state change from Active to Outputting is initiated by panel
function or by prescribed direction change of state of a second
Flag bit. When output is requested by panel function, the present
contents of the log is immediately printed, and each new Page of
data (forty lines) is automatically printed until the panel
function requesting outputting is cancelled. Where printing is
initiated by event, the digital action program controlling the
print determines both the number of readings to be taken before
printing is initiated and the duration of the printing. In any
case, where printing is initiated by event or event delayed, the
time of the initiating event will be stored as a line of data in
Column 1, and asterisks written into that corresponding line for
all other columns of that line. All logs with event triggering are
capable of manual triggering with or without delay to print.
Either the Active or the Output Report of an Active log will cause
full headings and data of the current page to be printed on the
printer associated with the requesting panel, unless otherwise
requested.
As an example, Pages to be assigned for logging can be taken from a
system pool of 280 pages. Assuming that the logs described
previously are numbered 1-99, this pool includes the Pages for
Series 100 logs as well as Special Logs.
Series 100 logs include;
A: Tabularized log 101 which is the formal log of up to 10
Pages;
B: Log 102 which includes the formal Daily Logs of up to 10
Pages;
C: Log 103 which is the Operational and TurbineTrip Analysis of two
Pages, and one-second frequency;
D: Log 104 which is the operational Data and Turbine-Trip Analysis
Log up to five Pages of Fifteen-second frequency;
E: Log 105, 106, 107 which are assigned to the Turbine Data Log for
operation at predetermined value openings once a day, and turbine
roll to half load, or trip to turning gear plus 10 minutes, and
also from certain analog or contact variables associated with the
turbine exceeding preset limits; and
F: Log 108 to 110 are spare.
Every log has the abiltiy to take on the above defined four states
and to move forward and backwards within such four states as shown
in the arrows interconnecting the blocks. With the above logging
system, no programmer is needed on call since the program in
operation is under the full control of the plant operator or
user.
There are generally three types of system variables, these being of
the analog type, the boolean type and the reference type. The
analog type varibles consist of points having values represented by
more than one bit plus quality. Analog type variables include
analog inputs, constants which are generated by the program or
entered via the operator's panel, fast calculations, performance
calculations, common variables and others. Boolean type variables
consist of points having values represented by one bit plus quality
and include variables such as digital inputs, relay outputs and
boolean bits (flags). Reference variables provide memory location
identification and include memory locations in core or drum/disc
for the purpose of references to the contents of those
locations.
The logging system can be used to provide both tabular data and
graphic or curve representations of the system variables. The curve
representations are generally used where the logging system is in
the active state and the data is moving rapidly or has just gone
into alarm status. Here, the system is programmed to produce curve
vectors to define the various operations. The system finds the
desired information via the log lists and produces a log trend of
the point at the present time and during the period of 40 logged
readings preceeding the present time, such as for the last 40
seconds or 20 minutes, etc. The generation of a log trend for a
point is, therefore, made by a system which can be characterized as
an on line real time system.
The system includes a log trend CRT display which can show trends
of up to four points, for example, in any active log or loop. When
trend is initiated by operation of ANALOG TREND (Button 126b shown
in FIG. 6) the system displays the equivalent, suitably scaled
analog tract from the values stored in the appropriate log page
(FIGS. 14A - 14E). FIGS. 14F-J represent a prior art display
(identified as "WITHOUT") in which there is no trend presentation
at the time of commanding a trend presentation since such a display
included no structured stored data.
A time scale grid can be displayed according to log frequency. The
trace can be suitably updated and "advanced" with each new log line
stored into the log pages.
The log trend CRT display enables operating difficulty to be
displayed immediately by back tracking a number of prior log
readings, to log readings, by way of example. As a result, effects
of immediate operating action are instantly displayed for tracking
correction action.
Referring to FIG. 4, there is shown a system operation flow
diagram. The system can be characterized by four general
operations. The first operation comprises the transforming of
analog and digital input data on lines 46 and 48, respectively, to
the computer into a universal format usable by all system
functions, indicated by the numeral 50 in the FIG. 4. The universal
formats, indicated by the numeral 52, includes the name, the
variable (quantity), the quality and the time. Thus, the log
members are defined by the name, value, quality and the time.
Therefore, the log members are defined by the universal format, and
the input data is transformed into such format.
The second general operation comprises the defining of triggering
conditions for the system, such as a specified change in time
.DELTA. T or a specified change in a variable in the format. In the
system, log triggering, indicated at 54, can be initiated by any
condition at pickup points under measurement, such as a boiler,
turbine or a generator, or by operator set tables of
adjustment.
As indicated above, every log has associated with it a built in
trigger (or IF statement) which is activated upon the occurrence of
the preset trigger conditions. Upon the occurrence of a trigger for
a given log, the third general operation of the system involves a
means of controlling the changes of log status with respect to four
log states defined above, as indicated by numeral 56. That is, when
a trigger ocurs, control means are provided which effects a change
of status from the Active Non-printing state to the Active Printing
State, from the Defined state to the Active state, etc., as per the
arrows shown in FIG. 3. If the log status is designated, then the
system triggers a change of status of the log to such designated
state. It is noted that there can be different types of triggering;
the trigger need not be instantaneous, but can be delayed.
The fourth operation of the system involves the formatting and
storage of data in the proper log page field as indicated by
numeral 58. If the data is a number of a log and it is the proper
time to store the data on an active log list, then such data is
formatted and stored in the proper log page field at the proper
index spot, thereby providing a set up for an efficient output on
all proper log pages.
The fifth operation comprises the data retrieval from the log
pages, final format and print out, indicated at numeral 60, of the
log page in accordance with the page oriented printing format
described above.
Referring to FIG. 5, there is shown a more detailed operations flow
diagram of the diagram shown in FIG. 4. In the FIG. 5, identical
numerals are used to denote identical portions of the system shown
in FIG. 4. Specifically, analog and digital input data is received
in the computer on lines 46 and 48, respectively, and presented as
scanned input data at 62. The signals are received from
transducers, thermocouples, etc. located at various parts of the
measuring system. The scanned input data at 62 can be transformed
by special and programmed calculations into calculated data at 64.
Both the scanned input data at 62 and the calculated data at 64 are
applied as quality and variable into the named system variable
format at 52. Also, data relating to time is provided from a time
data block 66 to the named system variable format 52. In addition,
memory cells 68 containing names are provided for the named system
variable format 52.
If there is a change of state of the named system variable 52 it
must be determined at 74 whether this change of state should cause
the log to trigger. On the other hand, if there is no change of
state at 70 as indicated at 72, then the pickup point must be
examined for other items. It is noted that the named system
variable package which includes the name, the variable, the quality
and the time is stored in the current data field.
Each analog type and Boolean type variable can be assigned separate
limits or a limit index to trigger an associated program. For
Boolean type variables it shall be possible to specify the
direction of state change to trigger the associated program. These
programs will be used to manipulate scan frequencies, log events or
events delayed, and other special system actions. Once a program is
triggered, retriggering shall not occur until the triggered program
has been serviced.
If the change of state at 70 is programmed to cause the log to
trigger at 74, then the program will control the log status as
generally indicated by numeral 56. More specifically, the program
is designed to control the log status by either changing the log
from the active, non-printing state to the active printing state at
76, or from the defined inactive state to the active state at 78.
If the program is such that there should be a change from the
active non-printing state to the active printing state at 76, then
the log is changed at 80 to print status for a period .DELTA. T. If
the program is such that there should be a change from the defined
inactive state to the active state at 78, then the clock is changed
to the active state at 82 for the time interval .DELTA. T.
If the log is changed to the print status at 80 or to the active
status at 82, the log package is then stored at 58.
There are two alternate approaches which can be employed for
storing data. The first approach requires that if the data is on
any active log list at 84 and it is time to store the data in a
defined list at 86, then the data is formatted and stored at 88 in
the proper log page field at the proper index spot. In the second
approach, if the data is on any active lists Y at 90 and it is time
to store the data in this log list Y at 92, then the data is
gathered at 94 from a storage buffer or tape. The first approach,
84, 86, 88 represents a separate job technique, whereas the second
approach, 90, 92, 94 represents a tagged technique.
The use of the universal format with named system variables, the
formatting and storage of data in log page field and the final
format of the page oriented printout provide a system with the
ability to change the types of logs, i.e., change the log members
or add to the logs.
The logging system is designed to known: the facilities to define
log lists, to report or to remove members of the lists, and to
allocate pages. The logging system also is designed to be triggered
by events or times, to set and control printout. Also, the logging
system is designed to have the ability to control log status.
Generally, two types of logs are known;
1. Formal logs which are produced periodically and are not
changeable by the operators but may be changed by special function
of the plant engineer; and
2. Special logs wherein members are placed into or removed from the
logs by the operators and the logs are of a periodic nature, or the
logs are triggered.
In defining a log, there can be assigned any system variable point
to any special log. Any value having a system name shall be
permitted on a special log. Also, any point (names system variable)
can be assigned to any column of the special log. As shown in FIG.
6, the Assign Log Button 119 of operator/engineer panel 118 is
pushed. The log N number, column and system variable name are
entered by means of buttons 120. Then the EXECUTE command button
121 is operated to cause the specified named system variable to be
assigned to a log column. In the example, column 17 automatically
starts the second page, column 33, the third, etc. A similar
sequence, with the CANCEL command button 122 causes the removal of
a system variable name from the specified column. This causes
blanks to appear if the column is printed or displayed. A special
case of CANCEL command is provided to permit assigning blanks to
all the columns of a particular log, thus in effect making the log
"Free." Log assign (button 119), log number (buttons 120) and
REPORT operation (button 123) causes the entire heading and system
variable names to be printed or displayed with the latest value of
each column.
Any log can be activated by demand. The Operator may insert or
change the period collection (from 1 second to 1 day), or
alternately specify it to be set up for activation by change of a
specific bit. This is accomplished by pressing button 119,
tabulating with the tab buttons 120 to set the frequency field,
entry of the frequency number of buttons 120, and pressing of
EXECUTE button 121.
Any number of logs can be activated at any time and at frequencies
compatible with system variable creation states. It is noted that
named system variables may appear in more than one log at the same
or different log collection rates.
Referring again to FIG. 6, operation of COLLECT LOG (button 124),
LOG NUMBER (buttons 120) and a CANCEL command (button 122) will
stop collection of log data and return that log to the Defined, not
Active state. If printing, it also stops printing on completion of
the present page.
When activated, a similar sequence of operations plus operation of
a REPORT command (button 123) will cause the full log headings,
system variable descriptions, etc., and contents in any log
storage, including latest system value of all assigned named system
variables and pages to be printed. If the display button, 125a of
button array 125, is down, the first page shall be written onto the
selected screen frame in normal group output format. Subsequent
pages can be seen upon operation of "Page For More" button
126a.
Referring again to FIG. 6, the operator can request the output of
any log, by number from the panels. Special buttons (FIG. 6) are
supplied for the FORMAL LOG (button 127) and SHIFT REVIEW (buttons
128). When immediate output is required, an EXECUTE command (button
121) is selected. The selected log will immediately be printed on
the printing devices associated with panel 118 unless otherwise
requested at EXECUTE time.
At EXECUTE time, any log can be alternately specified for event or
event delayed printing and an appropriate bit change specified to
trigger this. An event or event delayed activation time-of-event
occurrence, shall be stored in column one of the next line of the
next log page and all other columns blanked. The output should then
proceed for event types, or proceed after collection of additional
lines in the event delayed type.
When outputting an OUTPUT LOG (button 129) and a LOG NUMBER
(buttons 120), a CANCEL command (button 122) stops outputting and
returns the log to Active collecting state.
When outputting, OUTPUT LOG (button 129), and LOG NUMBER (button
120) are activated and if the DISPLAY selector (button 125a) is
selected, the report is outputted to the selected frame in group
format, one page at a time (via PAGE FOR MORE button 126a of array
126 of buttons).
According to the present invention, the log is both flexible and of
the triggering type. That is, one can create a trigger from any
variable you are working with, such as a boiler or generator pickup
point(s). Because of the provision of the named system variables in
the universal format described above, the system is able to set any
conditions for triggering. This universal format comprises a name,
a variable (quantity), a quality and time. The universal format
containing the named system variable is usable by all system
functions and enables the computer user to access the overall
information of all the specified pickup points upon demand or in
accordance with the programmed times or changes in state of the
named system variables.
Depending on the particular computer machine employed, the
formatting process can be divided in various manners, that is, if
there is a shortage or limited amount of computer storage space
then a format is employed which requires less information in the
central storage, or less channel transfer capacity, but on the
other hand, requires additional processing time. However, if there
is plenty of storage space and transfer capacity and processing
time limitations are strict, then a format will be chosen wherein
the data is stored in the central storage in a form more closely
resembling the format of the print out page, thereby requiring
relatively little processing time.
Referring to FIG. 7, there is shown a flow chart of the operation
for defining a log. A log definition is initiated by a request to
enter, change or delete a named system variable of a log. The
beginning of the flow chart for operation is indicated by the
symbol at 130. The system retains at 132 the details for selecting
the log number, the system variable name or blank, and the selected
column assignment, if such a column assignment is requested. If
there is as yet no selected column assignment, then the system will
go to the next available column on that or the following page.
Ordinarily, the human operator defines the log and selects the
details thereof. However, the operation can be designed to receive
the details of the log at 132. After the details of the log are
held at 132, the status description for that log must be located at
134 to determine what the current state of that log is at 136. If
the particular log N is currently defined, is active and storing or
is active and printing, then the page I of such log N must be
located for the selected or determined column J at 138. The purpose
of learning at 136 whether the particular log is currently defined
is that the condition must be known that the log exists before a
particular log is altered. If the log N is not currently defined at
136, and it is a request for a blank column at 140, then it is not
necessary to make a log assignment at 141 because nothing is being
entered into a log or changed in a log. However, if the entry is
not blank at 140, the status of log N is set to the defined state
at 142 to select a free page from the pool.
On the other hand, if the log N is currently defined at 136, the
page of the log N is located at 138 for the selected or determined
column J. With the log page number and column located at 138, then
a check is made at 144 as to whether or not the initial request was
to enter or change a point (i.e., not a blank) or whether the
request was for a blank log. If the request was for a blank log,
then it must be known at 146 whether there exists any other named
system variable on the log page. If there are other points on the
page, then blank spaces are entered in the column for those points
at 148 and a collector of values is notified at 150 that the old
point is no longer needed in this particular log N. If, however,
there are no other points on the page at 146, the system can remove
that page from the log and return it to a Free Page Pool at 152,
and thereafter notify the collector of values at 150 that the
particular point is no longer needed in the log N. After the page
is removed from the log and returned to the Free Page Pool at 152,
if such page was the last page in the log at 154, then the status
of such log page is set to the undefined state at 156 and that
portion of the operation for the log page is terminated at 158.
Returning now to the determination at 144 that the request was to
enter or change a log point (not a blank), this indicates that
there is a value that is going to be entered or changed in the log.
Due to this, a check is made at 160 as to whether the particular
page I that has been calculated for this log has already been
assigned. If the page I has not already been assigned, then a page
is selected at 162 from the Free Page Pool and the selected page I
is entered at 164 into the log N definition list.
If the page I is already assigned at 160 then a check is made at
166 as to whether there is a named system variable already in the
column J of such page I. If there is a named system variable
already in a particular position or column J, the collector of
values is notified at 168 that the old log point is no longer
needed in the log N. After the collector of values is notified at
168 or a free page is selected at 162 and the selected page I is
entered into the log N definition list at 162, then the computer
will enter at 170 the new named system variable into the log page I
of log N at the point or column J of the format. Thereafter, the
computer will notify the collector of values at 172 of the new
named system variable for collection so that the collector is on
notice that this new point is needed for the particular log N. This
operation is terminated at 158.
Referring to FIG. 8, there is shown a flow diagram of the operation
for changing the status of a log. A log status change is initiated
by an external or internal request to change the status of a log at
190. This input request is saved at 192 and the current log status
is located at 194. A check is made at 196 to determine whether the
current status of the log is defined, active or active printing. If
the log status is not one of these three conditions, (i.e.,
undefined) then the system characterizes this as an error condition
at 198, because the log must be in the defined state before its
status can be changed to any of the other states. If there is an
error condition at 198, and the request is from a human at 200, the
system informs a requestor at 202 that the log is not defined so
that the human is aware that he will have to define the log if he
is to change its status. If the input request is not in the form of
a human at 200, then the system sends a message at 204 to an error
handler stating that the log is inactive, is not defined, and that
the error handler must identify the user making the illegitimate
request.
If there is a positive condition at 196, then the system goes from
one state to another to examine the log for the current status, and
to set the status of the log to the desired condition. The three
desired log status conditions which can be requested at 206 are: to
make the log active in storing at 208; to make the log active in
printing at 210; and to make the log inactive and defined at 212.
To make the log active and storing at 208, a check is made to see
whether the log is currently printing at 214 and, if so, the
printing indicator is reset at 216 together with a request for
termination of any printing that is in progress. In other words,
the log has been printing at 216 and it is desired to terminate the
printing and set the log status to the active storing condition at
218.
If the log is not currently printing at 214, a check is made to see
whether the log is currently defined at 220 and, if so, the storing
frequency is set up at 222 as per the input request and the log
state is set to active and storing at 218. However, if the log
status is not currently defined at 220, the implication is that the
log is already actively storing, which is the very condition that
is being requested. This infers that the input request was made in
order to change the storing frequency, this being done at 224 per
the input request. In this case, since the desired log status at
208 is to make the log active and storing, and it is determined at
214 and 220 that the log is currently active and storing, then
there is no discontinuity in the storage status, but rather there
is a shift in the frequency or time period at 224.
If the desired log status at 210 is the active and printing state,
it is first determined at 226 whether the log is currently storing
and, if it is currently storing, a check is made at 228 to see if
the log is currently printing. If the log is currently printing at
228, the print/store frequency is changed at 230 as per the request
of the input information and the log status is set at 232 to active
and printing. On the other hand, if the log is not currently
printing at 228, the printing frequency is set up at 234 and a
request for printing to occur at the desired time is made via the
printing control mechanism.
If the log is not currently storing at 226, then the control flag
is set at 238 for the collector to begin storing values at the
designated frequency. Also, the flag is set at 234 for printing to
occur at the designated frequency, and the log status is set to
active and printing at 232.
If it is desired that the log be placed in the active/ defined
state at 212, this simply means that the log is to be returned to
the defined state wherein both printing and storing cease. To
accomplish this, the active printing request flag is reset at 240
and the active storing request flag is reset at 242 so that the
status of the log is set to the defined condition at 244.
Termination of the operation flow chart for changing the log status
is indicated by the symbol at 246.
Referring to FIG. 9, there is shown the operation flow diagram for
controlling the collection and/or outputting of log data. This
control action of log data is initiated at 252 by either a value,
an event or time. The status word of a particular log L is then
located at 254 and a check is made at 256 whether the status of the
log is either active and storing or active and printing. It is
noted that the term "printing" as used herein, is intended to
include other outputting operations, such as a visual display on a
CRT device. If the log L status is active and storing or active and
printing at 256, and it is determined at 258 that it is time to
store data for log L, then a list of all named system variables
required for log L is collected at 260, all of the log L values are
stored at 262 into appropriate locations for this time frame, and
this stored set of log L values is added at 264 to the buffer
provided for log L.
If it is not time to store data for log L at 258, or the stored set
of values are added to the buffer provided for log L at 264, then
it is determined at 266 whether log L is active and printing. If
the log is active and printing at 266, and it is time to print the
log L at 268, then the heading and format of log L is collected at
270 for outputting on the desired device and the prepared log L is
outputted on the requested or designated device on request at
272.
Action on the collection and/or outputting of log data is
terminated at 274.
SEQUENCE OF EVENTS INTERRUPTS
Referring to FIG. 10, there is shown a protective circuit connected
to a system in accordance with the present invention. Protective
circuit 280 includes several overload and protective devices
illustrated by the FIG. 10. Specifically, there is shown a trip
initiating device 282, a trip actuation device 284, an intermediate
device 286 and a protective relay 288 for other independent trip
initiation. Intermediate device 286 represents a chain of
intermediate devices and the cumulative delay resulting therefrom.
The device 288 is another protective device installed for a
different purpose, but which, due to the natural coupling of
equipment not part of protective circuit 280, can in fact be
actuated by the physical item that caused device 282 to actuate.
Since device 288 has a low cumulative delay time, this undesired
cross-coupling can cause the contact from device 288 to actuate the
final trip actuation device 284 before the recorded signal coming
through chain device 286 arrives at device 284. This causes an
erroneous allocation of the initiation of the trip.
Other switching elements may be connected in the protective circuit
280. The specific arrangement of the protective devices is not
pertinent to the invention. Rather, the protective circuit 280 is
shown for the purpose of illustrating that various protective
devices connected in a power plant system are ordinarily
interconnected in various cascade and series circuits with varying
attendant delays. Therefore, the activation of any one of the
protective devices will frequently affect the other devices causing
them to be activated and their associated contacts will be switched
For instance, the energization of the relay 286 might cause the
protective relay to be energized.
The problem that exists with such a protective circuit 280 is that
the control system 300 will detect only the actual occurrence of a
switching operation of the contacts, which will ordinarily occur
many milliseconds after the occurrence of the event affecting the
device 282. Also, due to the external interaction of the devices in
the protective circuit, certain ones of the contacts can be
actuated before the actuation of the contact relay string reporting
the particular protective device which sensed the event. For
example, referring to the Table A below, there is shown in the left
hand column a sample listing of the inherent time delays in each of
the devices 282, 284, 286, 288 and 290 (expressed in
milliseconds).
TABLE A.
TIME DELAY BETWEEN ACTIVATION TIME AND CONTACT SWITCHING TIME
Time Delay (282) = 40 milliseconds
Time Delay (284) = 25 milliseconds
Time Delay (286) = 80 milliseconds
Time Delay (288) = 35 milliseconds
Time Delay (290) = 70 milliseconds
Thus, if the chain represented by device 286 is characterized by an
80 millisecond delay between initiation of the interrupt event
initially sensed by device 282 and the actual switching of its
contacts 282a, and a device 284 has a 25 millisecond delay, then
the occurrence of an interrupt at the device 286 may at the same
time by way of contacts 286a cause an auxiliary trip at the device
284. Because of the various external interaction or chain of
connections, the contacts 288a and 288b for device 288 may be
switched ahead of the contacts of the device 282 which was the
actual cause of the event, such as a boiler explosion. The control
system 300 will detect the switching of the contacts 284a and 288b
for devices 284 and 288, respectively, before it detects and
records the switching of the contact device 282. This error is
commonly found in protective circuits having a plurality of
hardware devices.
To determine the actual or real cause of the event, the control
system 300 includes a storage section 302 for storing the time
delay characteristics of each protective device used in the
protective circuit 280 for each path. Also included is a logic
arithmetic section 304 for applying the corrective delay times held
in the storage section 302 and timing circuit 308 of the control
system 300. The input section 306 is connected to the protective
circuit 280 for receiving signals from each of the switching
contacts. The logic arithmetic section 304 receives signals from
the input section 306 indicating the identification of the
switching contacts which were actuated and the times of occurrence
of the actuation. Logic arithmetic section 304 subtracts the time
delays for the proper path stored in storage section 302 and
associated with the protective devices, from the contact actuation
times indicated on line 310 from the input section 306 and provides
at its output 312 a corrected reading indicating an identification
of the device and the actual corrected initiation time and
corrected sequence of events. A plurality of devices may be listed
in order of their occurrence of computed initiation times as
compared with the time of switching of their contacts. For this
purpose, the logic arithmetic section 304 or a separate logic
circuit 314 may be employed.
The Table B illustrates the manner of operation of the logic
arithmetic section 304 wherein the detected times of switching of
the contacts of the various devices are received on line 210, the
time delays associated with the respective devices are received
from the storage section 302 on line 311, and the corrected actual
time of activation of each device (event interrupt) is computer and
provided at the output line 312 of section 304.
TABLE B
__________________________________________________________________________
DETECTED CONTACT DEVICE TIME CORRECTED ACTIV- SWITCHING TIME -
DELAY = ATION TIME
__________________________________________________________________________
T.sub.Detect (282) - 40 msec. = T.sub.Correct (282) T.sub.Detect
(284) - 25 msec. = T.sub.Correct (284) T.sub.Detect (286) - 80
msec. = T.sub.Correct (286) T.sub.Detect (288) - 35 msec. =
T.sub.Correct (288) T.sub.Detect (290) - 70 msec. = T.sub.Correct
(290)
__________________________________________________________________________
After the logic arithmetic section 304 computes the corrected time
of activation of each of the devices, it re-orders the data
received by the computer control 300 according to the corrected
sequence so that the actual interrupt event is determined. That is,
the computer control 300 determines the correct sequence of
activation of the devices by the alarm or interrupt event.
ALARMING AND DISPLAY DEVICES
Referring to FIG. 11, there is shown a graphical representation of
a curve of a measured system variable at a point, for example the
temperature of a boiler. The ordinate 330 represents temperature
values (degrees Fahrenheit) while the abscissa 332 represents time.
The curve of the values of temperature with respect to time is
indicated by the numeral 334. The example shows the high limit
alarm value set at 930.degree.F for the boiler and indicated by the
numeral 336. It is noted that a low limit alarm value is also set
at 820.degree.F, as indicated by the numeral 338. The temperature
range between the high limit alarm value 336 and the low limit
alarm value 338 is referred to as the "no alarm" or normal side of
the alarm limits 336 and 338. A fixed value for the significant
change, also referred to as delta, is set at 5.degree.F. This
significant change, or delta, when added to the most recent alarmed
value of the measured system variable, defines a band having an
upper bound which is 5.degree. above the most recent alarmed value
and a lower bound which is 5.degree. below the most recent alarmed
value. When the value of the measured system variable reaches or
exceeds this band, the measured point is realarmed and the band
reestablished about the current value of the measured system
variable. This current value is now the most recent alarmed value.
If the measured system variable value is within the normal range
between alarm limits 336 and 338, the delta does not apply.
In addition, each point in the system has a multiplier factor which
is used to expand the delta value at times when the operation of
the system requires a decision as to which of the alarms are the
most important as to warrant being outputted on a CRT or other
output device. Therefor, the multiplier factor is applied to the
fixed delta value for those alarms which have been determined to be
of less significance with regard to the current power plant status.
The multipliers are uniquely selectable and applicable at each
individual measuring point as a part of an algorithm and can be
applied by program or by a human operator. The algorithm is
comprised of the actual significant change or delta which is equal
to the specified or fixed delta times the multiplication factor M.
Thus, the use of the multiplication factor is essentially the
application of a correction factor to an already existing computer
used value in order that the alarm system be apprised of only the
significant changes in the measured values after the initiation of
the first alarm.
In the example, it is assumed that a multiplication factor M of two
(2) is applied to a delta of 5.degree., thereby changing the
effective delta to 10.degree.. If, due to the scanning frequency of
the measuring device, the temperature of the boiler tube is first
measured as crossing the alarm limit value of 930.degree. at a
value of 932.degree., as indicated by the cross at numeral 340,
then 932.degree. represents the most recent alarmed value. Since
the new effective delta is 10.degree. (5.degree. times 2), then the
next alarm value or condition at which the alarm message will occur
is greater than or equal to 942.degree., as indicated by the
numeral 342. If the measured value is subsequently scanned at
943.degree., as indicated by the cross at numeral 344, this
represents the new alarmed value. This process continues in the
same manner so that the next alarmed value will be greater than or
equal to 943.degree. plus 10, or 953.degree., as indicated by the
line 346, or less than or equal to 943.degree. minus 10, or
933.degree., as indicated by the line 347.
Referring to FIG. 12, there is shown a screen 350 of a CRT device
employed for outputting the alarm information obtained by the
system disclosed with reference to FIG. 11. Each alarm, when first
detected, is instantly displayed on the next free line, L1-Lx,
starting at the bottom of the alarm area 351 of the screen 350.
Each of the lines L1-LX may contain alarm information for a single
measured system variable in the system, as is indicated with the
alarm information presented. The first alarm for a given measured
system variable in the system is first displayed on a line in a
"new alarm" format which utilizes the full width of the screen 350.
As an example, the line L2 is shown as displaying a new alarm
format wherein alarm information is displayed across the entire
line L2 along sections 352, 354 and 356. When an operator
acknowledges the alarm information displayed on the line L2, the
alarm message is automatically changed to a reduced format wherein
only the current value, the current deviation from the alarm limit,
the current direction and the current duration of the alarm is
updated at the scan or calculation frequency of the measured system
variable. This reduced format comprises the portion of line L2
indicated by numeral 354. The heading for the data displayed in the
alarm area 351 is indicated by numeral 358. As shown in the alarm
heading 358, the new alarm format includes the point
identification, an English description of the system variable,
i.e., Boiler Tube 68 Temperature, the S representing the
identification of the alarm limit which is being violated high or
low, the D representing the direction in which the current measured
value moved with respect to the last measured value, the current
measured value, the engineering units of the measured value, the
difference between the alarm limit value and the current value, the
quality Q, the .DELTA. T representing the length of time the
measured value has exceeded the alarm limit value (336 or 338), the
low limit value, the S representing an indication that the low
limit detector is on or off, the high limit value, and the S
representing an indication that the high limit detector is on or
off.
When the alarm area 350 is filled with alarm information, that is,
the top line LX is occupied, then a group of the oldest alarm
messages at the bottom of the alarm area is transferred into a
system memory for alarm backlog information and the alarm
information on the lines above those transferred shall be compacted
downwards. This readjustment permits entry of the new alarm at the
bottommost but unused line of the alarm area.
Referring again to both FIG. 11 and FIG. 12, it is noted that the
occurrence of a new alarm at point 340 of FIG. 11 can be displayed
in a new alarm format including sections 352, 354 and 356 of FIG.
12. After being acknowledged by the operator the alarm information
at point 340 is reduced to the acknowledged alarm format covering
the section 354 of the section 350. However, if the same measured
system variable, shown graphically in FIG. 11, undergoes a
significant change in the measured value to exceed the delta of
10.degree. set for such point, such as being measured at
943.degree. as shown by the numeral 344, then the line of the alarm
area 350, which is displaying the alarm information for this
measured system variable will expand to the new alarm format until
it is acknowledged, whereafter this second alarm information is
reduced to the narrow format.
Thus, it is seen how the alarm area 350 is used to convey two
different states or conditions of alarm, these being the new alarm
or wide format, and the acknowledged alarm or narrow format. Also,
the alarm information remains in the same line position in the
alarm area 351 as the original alarm information for the same
system variable up until the time that the area must be rearranged
for accommodating a new alarm message for which there is no free
line at the top of the alarm area. This permits the operator to
recognize a position or line on the screen as representative of a
specific alarm message until the point in time at which the alarm
area must be reorganized.
If a particular measured system variable, previously in the alarm
condition, returns to the normal condition, the system and operator
are notified on the return to normal and the alarm information is
removed from the alarm area 351. At this point, however, the other
lines are not shifted downward to fill in the removed line of
information until the alarm area 351 is filled to the top line. The
purpose of this technique is to maintain the alarm data in the same
line position for as long as possible, as opposed to continuously
shifting the data on the screen. Of course, as mentioned above,
when the alarm area 351 is filled to the top line and the content
of the alarm area is to be maximized, then a group of the oldest
alarm information is removed from the alarm area 351 and placed in
a backlog in the system memory.
FIG. 13 shows a generalized functional block diagram of the
alarming system. The alarm data is received on lines 370 by a
control unit 372 which controls the entry and removal of alarm
information on the CRT screen 350. Control unit 372 also controls
formatting and the position of the alarm information on each of the
lines on the alarm area 351 in accordance with the described
method, as well as controlling the removal of alarm information
from the alarm area 351 into a backlog condition in a system memory
374. The alarm backlog in system memory 374 is made available for
display on a visual information presentation device 376. Alarm data
in the system memory 374 is made available for presentation by the
device 376 upon request via an operator keyboard console 378
connected to the control unit 372. The measured system variables
held in the alarm backlog of system memory 374 are updated in a
similar manner as updating of the points on the CRT screen 350.
When a measured system variable in the alarm backlog changes
significantly past its designated delta, it is removed from the
alarm backlog and reinserted once again on the alarm area 351 in
the new alarm format at the next available line for an alarm
message.
When a point which has been in alarm is determined to have returned
to the normal condition, such point is removed from the CRT screen
350 or the alarm backlog of system memory 374 and indicated as
returning to the normal condition on another area of the CRT screen
350 other than the alarm area 351.
The visual information presentation device 376 includes on its
screen a display area 377 which functions as a keyboard input guide
and verification. Display area 377 includes the four bottom lines
of the screen of device 376 for presenting a memory of what the
human was working on, even though the demand functions of the
operator have been changed. The device 376 presents information
relating to what was last carried out with the particular blank of
information when it was used by the human. The device 376 enables
the human to go back and see what he had been doing with a
particular function, and then modify his last operation or
condition. The memory information produced in the display area 377
is stored in a local memory 379. Thus, the display area 377 acts as
a guide of the human inputs previously entered in the system and
enables the human to simply fill in the fields to update the
information rather than reentering a new set of information
relating to the same item.
When a point has been in alarm is determined to have become a bad
or faulty point, it is removed from the CRT or alarm backlog and is
indicated as being a bad input. On the other hand, if a point has
not been in alarm and is determined to be faulty, this is marked as
a bad input and all user programs using that particular input are
notified, and it is put into the bad input summary and name and
status of this point is temporarily indicated on an area of the
screen reserved for bad inputs. The system includes a bookkeeping
arrangement which accounts for each of the points that activated an
alarm and subsequently went bad. This is useful since the alarms
must only be activated and maintained by good or suspect system
variables.
BELIEVED VALUE
In the conventional plant condition measurement systems of the type
wherein several inputs are connected, if one of the inputs is in
error, then the output of the system is significantly affected by
the error of the input device. In addition, sudden jumps occur in
the output of the system, as the input device changes state from
unacceptable to acceptable, and vice versa. With this incorrect
output, the system erroneously signals for a corrective action
leading to incorrect operating actions by either human operators or
an automatic control system.
The present invention provides a system which compensates and
corrects for the failure of an input or input device measuring
certain conditions and involves a method of timing a circuit for
the characteristics of the input measurements, thus permitting one
input device to fail, be repaired and restored to use without
significantly affecting the resulting output value of the
system.
The present invention provides a "Believed Value" which is a
mathematically determined number representative of the most likely
value of an important plant measurement. A Believed Value is a
special form of an analog composed point, and is a calculated value
which may require manipulation of one or more analog inputs and/or
other variables. For a critical plant measurement, multiple input
devices, (direct or inferential) are installed to provide
independent measurements. Generally, the Believed Value involves a
technique wherein the relationships among the various inputs are
compared for validity by combinations of the differences between
any two of their indications. This method permits analyzing the
disagreement of the input values and confirming the "agreed on"
value as a "true" value for the critical item being measured. The
system is designed to mathematically compensate for, and tag as
Bad, the particular input device which is in error by a large
amount. A lesser degree of error will be tagged as Suspect.
Each of the three input devices has reasonably compatible dynamics,
such as time lags or responses to measurements in the same order of
magnitude and each of the devices is designed to sense related
conditions, such as pressure or flow. As an example, the described
arrangement includes three devices for sensing pressure or three
devices for sensing flow or three devices for sensing a liquid
level. Since the three inputs are related to each other, when the
readings in one input device vary, then similarly, the readings in
the other input devices should vary if they are working properly.
Where the devices are not measuring directly the same condition,
they will be suitably scaled and linearized to give compatible
numbers.
Tuning of the Believed Value is carried out by varying the output
of one of the input devices while holding the other two at
appropriate levels. The output of the input device that is varied
is taken over the range of values likely to be produced by the
input device and includes values that would represent the failure
mode. At the same time, the output of the system is recorded over
such ranges as these values are run, so that the performance of the
Believed Value system can be evaluated. The tuning of this system
is accomplished by adjusting the non-linear factor (NLF) which
designates a band of reasonable variation and the non-linear power
(NLP) which controls the degree of weighting provided for each
value as a function of its deviation.
More specifically, tuning of the filter circuit to a particular set
of input devices or transmitters for a particular process
measurement involves the varying of the values of a non-linear
factor (NLF) and a non-linear power (NLP). The non-linear factor
(NLF) is a number derived from the probable disagreement between
the readings of the input signals or devices, since there is
involved a statistical uncertainty of known readings. The
non-linear power (NLP) is a function of the type of Believed Value
that one is seeking. For example, as shown in FIG. 15A, the curve
382 shows a particular Believed Value curve for a non-linear power
(NLP) equal to four. Here, the inputs one (1) and two (2) are held
constant at the values V1 and V2, respectively, while the value of
the third input (3), shown as V3, is varied in accordance with the
curve indicated by numeral 384. On the other hand, the Believed
Value curve indicated by numeral 386 represents an NLP equal to one
and has a smoother slope than the relatively steep slope
characterizing the Believed Value curve 382. In this fashion, the
tuning method is used to derive the values of the non-linear power
(NLP) and the non-linear factor (NLF) which provide the desirable
characteristics or shape of the Believed Value. Also, as will be
explained in more detail hereinbelow, the tuning method produces
those NLP and NLF values which are used to compute the Believed
Values for the real values of the input signals.
Referring to FIG. 15B, the three inputs V1, V2 and V3 are shown as
being applied on lines 388A, 388B, and 388C leading into a tuning
circuit 390. Tuning circuit 390 includes means for adjusting the
values of the NLF and NLP factors which are received on lines 392
and 394 from a tuning coefficient generator 395. The output of
tuning circuit 390 provides a Believed Value on lines 397 which is
fed to maintenance checking circuits 399 which in turn provide a
Believed Value output on line 396 and maintenance output signals on
line 398.
The test method and system described above permits such tuning to
be accomplished with realistic system values. When properly tuned,
the system includes provisions for retention of the tuning
coefficients such that if the input device comes into error or is
inoperable, then the output of the system (the Believed Value) will
stay in a reasonable range of value. By means of these adjustments,
a Believed Value can be produced which compensates for a failing of
one of the input components, so the output reading remains useable
for the operator. Stated another way, the Believed Value is an
output number derived from a system which is valid in the presence
of a fault with one of the input components.
The Believed Value is generally produced for a given point for a
specific condition or characteristic of the power plant, such as
pressure, flow or fluid level. The basic difference between a
Believed Value and a standard system value is that, due to the
tuning circuit and the computation of the Believed Value against
the set of input devices, a failure of one of the input devices
will not produce an erroneous indication of a plant operating
failure. In addition, the Believed Value is a number that does not
have sudden shifts in it due to its input members taking sudden
wide excursions. Connected to the output of the Believed Value
tuning circuit is a test circuit for testing the Believed Value
against acceptable quality standards, marking those deficient as
requiring maintenance. The output of the test circuit includes
maintenance signals for correcting the faulty input device. Also, a
pre-processor can be connected between the input devices and the
Believed Value tuning circuit for adapting the system to permit
more than the given number, in this case, three, of inputs to be
handled, such as n inputs. The output value of the system is a
number that represents the best or most probable measurement for
that value from an operating standpoint. For example, the steam
pressure can be most probably represented as having the value X.
The Believed Value is the number that the operator is supposed to
believe and, therefore, should use in operation.
It is to be noted that the Believed Value technique is not
analogous to the well known redundancy technique whereby all inputs
must be identical and the system does not operate all inputs
simultaneously. The conventional redundancy technique is one in
which the identical lines are considered bi-stable as being in
either the good or bad state and, a large excursion of one of the
lines is totally ignored where the other two lines have the same
values. Conventional techniques are very intolerant of non-matching
dynamics of signals. By contrast, the Believed Value technique
involves, the this case, three or more inputs or input devices
which have reasonably compatible dynamics and the various
measurement techniques associated with the three devices have time
constants in the same order of magnitude, but has toleration for
non-matching transients or dynamics. Also, according to the
Believed Value technique, the tuning circuit takes into
consideration variations in values of one of the input devices for
every reading that the other input devices provide so that where
such one input device should go faulty, the Believed Value will
stay within its computed range.
Thus, the Believed Value technique includes the steps of tuning a
filter circuit which is connected to receive signals from the input
devices, such tuning being accomplished by calculating output
values for the tuning circuit as a function of various values of
the input devices over a range wherein an input device varies
between its maximum and minimum excursions with the other two input
devices being held to their expected values, such calculating
includes weighing the effect of variations in value of one of the
input devices on the output of the filter, and computing a Believed
Output representative of the most likely value of a plant
measurement and output of the tuning circuit, such Believed Value
providing a reliable value of proper operating conditions under
measurement whereby a fault in one of the input devices will not
change the output of the system out of the range of the Believed
Value. The system also includes a trigger mechanism and a
maintenance, marking and bookkeeping system.
FIG. 16 shows the steps involved in the determination of the
Believed Value. The first block of steps indicated as 400 is the
basic Believed Value determination, the second block of steps
labeled 410 is the post processing check required for reporting
deviant inputs so that proper maintenance can be applied. After
preprocessing has reduced the remaining values to three, they are
processed by steps 400 and then by steps 410. Within the block of
steps 400, the first step 401 is to prepare Averages of each pair
of the three values. These are called Average one (1), Average two
(2) and Average three (3). Average one (1) consists of the average
of Values one and two. Average two (2) consists of the average of
Values one and three. Average three (3) consists of the Average of
Values two and three. Step 402 normalizes these averages by
dividing each of them by a non-linear factor (NLF) which is one of
the tuning coefficients described for this system. Step number 403
determines the absolute value of the differences between input
Value one and two, Value one and three, Value two and three, and
divides each of them by the non-linear factor NLF.
Step 404 defines the computation of weighting factor (WF)
calculated for each average (401) and the corresponding normalized
absolute difference (403) and the other tuning factor NLP,
non-linear power. The denominator of the equation in 404 is
computed by adding to one (1) the absolute normalized difference
(403) raised to the non-linear power (NLP) and the result divided
by the Average (401). The numerator of the equation is one. This
equation (404) is repeated for each pair of corresponding Averages
and Differences.
Step 405 sums each Average multiplied by its weighting factor for
the numerator, and then divides that by the summation of the
weighting factors to produce the Believed Value.
Block 410 in the post processing sorts out the transducers that
require maintenance attention and designates them as Bad or Suspect
and, if they do not require maintenance attention, marks them as
good. This is accomplished by step (411) for calculating the
absolute differential between Value one and the Believed Value
previously determined (405), Value two and the Believed Value
previously determined, and Value three and the Believed Value
previously determined. In step 412, the two smallest of the
absolute differentials are averaged and their associated values are
marked as good. The average of these two smallest differentials is
compared (413) with the non-linear factor and selects which ever of
these numbers is the greater. This greater number is referred to as
the test factor or TF. A further check is made to insure that the
test factor (TF) has not exceeded five times the non-linear factor.
Such a deviation indicates a broad discrepancy in the input
readings and all are marked Suspect for a maintenance checking. In
step 414, the largest or remaining of the absolute differences
(411) is checked to see if it is greater than Z times the test
factor (TF), Z being an adjustable tuning coefficient. If the
absolute difference is greater than Z times the test factor, its
associated Value is marked Bad and flagged for maintenance. In step
415, the largest absolute difference is again checked against Y
times the test factor, Y being an adjustable tuning coefficient,
and if it is found to be greater than Y times but less than Z times
the test factor (TF), it is marked Suspect and again flagged for
maintenance. If it passes both tests, the associated value is
marked good in step 416 and the function has been accomplished.
Although the above description is directed to the preferred
embodiments of the invention, it is noted that other variations and
modifications will be apparent to those skilled in the art, and,
therefore, may be made without departing from the spirit and scope
of the present invention.
* * * * *