U.S. patent application number 14/267345 was filed with the patent office on 2015-11-05 for self learning radio frequency monitoring system for identifying and locating faults in electrical distribution systems.
This patent application is currently assigned to Siemens Energy, Inc.. The applicant listed for this patent is Siemens Energy, Inc.. Invention is credited to Jon Patrick Oak, Edward David Thompson.
Application Number | 20150316606 14/267345 |
Document ID | / |
Family ID | 54355105 |
Filed Date | 2015-11-05 |
United States Patent
Application |
20150316606 |
Kind Code |
A1 |
Oak; Jon Patrick ; et
al. |
November 5, 2015 |
SELF LEARNING RADIO FREQUENCY MONITORING SYSTEM FOR IDENTIFYING AND
LOCATING FAULTS IN ELECTRICAL DISTRIBUTION SYSTEMS
Abstract
Electrical faults are detected in electrical distribution
systems (EDS) by detection and location of radio frequency (RF)
emissions generated by the fault with multiple time-synchronized
radio frequency monitors (RFM) distributed about the EDS. The RFMs
are coupled to a self-learning, electrical fault monitor (EFM) that
characterizes and/or locates electrical faults based on operating
state (OS) patterns learned from transmission of test signals
generated within the EDS. RF emissions data samples are
characterized as safe operation (SO) states or potential electrical
faults by accessing a base of stored knowledge concerning fault
emission characteristics and/or synchronized time of arrival at
each RFM. Information in the base of stored knowledge is updated to
include new EDS operating states (OS). Confidence level
associations, location of new radio frequency emission patterns and
whether those patterns are indicative of safe operating (SO)
conditions or electrical faults are stored in the base of stored
knowledge.
Inventors: |
Oak; Jon Patrick; (Fern
Park, FL) ; Thompson; Edward David; (Casselberry,
FL) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Siemens Energy, Inc. |
Orlando |
FL |
US |
|
|
Assignee: |
Siemens Energy, Inc.
Orlando
FL
|
Family ID: |
54355105 |
Appl. No.: |
14/267345 |
Filed: |
May 1, 2014 |
Current U.S.
Class: |
324/537 |
Current CPC
Class: |
G01R 31/085 20130101;
G01R 31/14 20130101; G01R 17/02 20130101; G01R 31/2834
20130101 |
International
Class: |
G01R 31/28 20060101
G01R031/28; G01R 17/02 20060101 G01R017/02 |
Claims
1. A method for monitoring electrical faults in a power plant
electrical distribution system, comprising: monitoring and sampling
in real time with an EDS electrical fault monitor (EFM) operation
state (OS) radio frequency (RF) emissions data from electrical
distribution system (EDS) equipment that are obtained with a
plurality of time synchronized radio frequency monitors (RFM) that
are distributed about the EDS and coupled to the EFM, the OS RF
emissions potentially indicative of electrical faults in the EDS;
storing in an automated data storage device that is coupled to the
EFM OS data samples for each RFM, sampled data including the RFM
identification, data sample time and data sample waveform
characteristics; determining with the EFM in real time a likelihood
of whether one or a combination of the respective OS data samples
is indicative of an electrical fault by: referencing in an
automated data storage device coupled to the EFM previously stored
information associating OS data with any one or more of EDS safe
operation (SO), electrical faults or RF emission location within
the EDS; comparing at least one stored OS sample reading from each
RFM with respective stored association information relevant thereto
and making respective EDS safe operation (SO) first confidence
level determinations; comparing a combination of the at least one
stored OS sample reading from each respective RFM with respective
stored association information relevant thereto, if such
combination information is available, and making respective EDS
safe operation (SO) second or more confidence level determinations;
and combining all prior sequentially determined confidence levels
information to derive an EDS safe operation (SO) overall confidence
level; and causing the EFM to generate an alarm when any of the
sequentially determined first through overall SO confidence levels
is indicative of an electrical fault within the EDS.
2. The method of claim 1, further comprising updating previously
stored SO information in the automated data storage device with
additional information derived from performing any of the previous
steps.
3. The method of claim 2, further comprising performing
sequentially one or more additional confidence level determinations
after the second confidence level determination.
4. The method of claim 1, wherein the determining step is performed
by the EFM using any of fuzzy logic, self-learning, algorithmic,
statistical, neural network logic, case-based reasoning or
heuristic methods.
5. The method of claim 1, wherein the previously stored information
in the automated data storage device comprises any of rules, set
points, absolute reading limits, RF waveforms, time differences
between receipt of data samples by different RFMs, historical
information, weightings, human inputs or statistical
probabilities.
6. The method of claim 1, the previously stored information
comprising sampled data obtained by the RFMs resulting from
generating RF test signals with an RF test generator.
7. The method of claim 1, further comprising the EFM locating the
RF emission source within the EDS based on correlation of
respective data sample time and/or respective data sample
attenuation obtained from multiple RFMs.
8. The method of claim 7, further comprising: the previously stored
information in the automated data storage device including sampled
data obtained by the RFMs resulting from generating RF test signals
with an RF test generator; updating previously stored SO
information in the automated data storage device with additional
information derived from performing any of the previous steps with
the EFM; the determining step performed in the EFM using any of
fuzzy logic, self-learning, algorithmic, statistical, neural
network logic, case-based reasoning or heuristic methods; and the
previously stored information in the automated data storage device
including any of rules, set points, absolute reading limits, RF
waveforms, time differences between receipt of data samples by
different RFMs, historical information, weightings, human inputs or
statistical probabilities.
9. A system for monitoring electrical faults in a power plant
electrical distribution system, comprising: a plurality of time
synchronized radio frequency monitors (RFM) that are distributed
about the electrical distribution system (EDS), for monitoring and
sampling in real time EDS operation state (OS) radio frequency (RF)
emissions data from electrical EDS equipment, the OS RF emissions
potentially indicative of electrical faults in the EDS; an
electrical fault monitor (EFM) coupled to the RFMs and an automated
data storage device having stored therein sampled data readings for
each RFM, sampled data including the RFM identification, data
sample time and data sample characteristics, the RFM determining in
real time a likelihood of whether one or a combination of the
respective OS data samples is indicative of an electrical fault by:
referencing in the automated data storage device coupled to the EFM
previously stored information associating OS data with any one or
more of EDS safe operation (SO), electrical faults or RF emission
location within the EDS; comparing at least one stored OS sample
reading from each RFM with respective stored association
information relevant thereto and making respective EDS safe
operation (SO) first confidence level determinations; comparing a
combination of the at least one stored OS sample reading from each
respective RFM with respective stored association information
relevant thereto, if such combination information is available, and
making respective EDS safe operation (SO) second or more confidence
level determinations; combining all prior sequentially determined
confidence levels information to derive an EDS safe operation (SO)
overall confidence level; and causing the EFM to generate an alarm
when any of the sequentially determined first through overall SO
confidence levels is indicative of an electrical fault within the
EDS.
10. The system of claim 9, the automated storage device having
stored therein additional SO information derived from performing
any of the prior EFM determining steps.
11. The system of claim 9, the automated storage device having
stored therein additional SO information derived from performing
prior electrical fault monitoring during EDS operation in the power
plant.
12. The system of claim 9, the EFM sequentially performing one or
more additional confidence level determinations after the second
confidence level determination.
13. The system of claim 9, the EFM performing the determining step
using any methods including fuzzy logic, self-learning,
algorithmic, statistical, neural network logic, case-based
reasoning or heuristic methods.
14. The system of claim 9, wherein the previously stored
information in the automated data storage device comprises any of
rules, set points, absolute reading limits, RF waveforms, time
differences between receipt of data samples by different RFMs,
historical information, weightings, human inputs or statistical
probabilities.
15. The system of claim 9, the previously stored information
comprising sampled data obtained by the RFMs resulting from
generating RF test signals with an RF test generator.
16. The system of claim 9, further comprising: the automated data
storage device previously stored information including sampled data
obtained by the RFMs resulting from generating RF test signals with
an RF test generator and additional SO information derived from
performing prior electrical fault monitoring during EDS operation
in the power plant; the determining step performed in the EFM using
any of fuzzy logic, self-learning, algorithmic, statistical, neural
network logic, case-based reasoning or heuristic methods; and the
previously stored information in the automated data storage device
including any of rules, set points, absolute reading limits, RF
waveforms, time differences between receipt of data samples by
different RFMs, historical information, weightings, human inputs or
statistical probabilities.
17. A method for monitoring electrical faults in a power plant
electrical distribution system, comprising: monitoring and sampling
in real time with an EDS electrical fault monitor (EFM) operation
state (OS) radio frequency (RF) emissions data from electrical
distribution system (EDS) equipment that are obtained with a
plurality of time synchronized radio frequency monitors (RFM) that
are distributed about the EDS and coupled to the EFM, the OS RF
emissions potentially indicative of electrical faults in the EDS;
storing in an automated data storage device that is coupled to the
EFM OS data samples for each RFM, sampled data including the RFM
identification, data sample time and data sample waveform
characteristics; determining with the EFM in real time a likelihood
of whether one or a combination of the respective OS data samples
is indicative of an electrical fault by: referencing in an
automated data storage device coupled to the EFM previously stored
electrical fault information obtained by the RFMs, resulting from
generating electrical fault RF test signals with an RF test
generator in the EDS; and comparing the sampled OS data with the
previously stored electrical fault information in the automated
data storage device; and causing the EFM to generate an alarm when
the referenced stored information associates the sampled OS data
with an electrical fault within the EDS.
18. The method of claim 17, the previously stored information in
the automated data storage device including any of rules, set
points, absolute reading limits, RF waveforms, time differences
between receipt of data samples by different RFMs, historical
information, weightings, human inputs or statistical probabilities;
and the determining step performed in the EFM using any of fuzzy
logic, self-learning, algorithmic, statistical, neural network
logic, case-based reasoning or heuristic methods.
Description
BACKGROUND OF THE INVENTION
[0001] 1. Field of the Invention
[0002] The invention relates to electrical fault detection in
electrical distribution systems (EDS) by detection and location of
radio frequency (RF) emissions generated by the fault. More
particularly the invention relates to self-learning radio frequency
electrical fault monitor (EFM) systems and methods, having time
synchronized distributed radio frequency monitors, wherein the
monitoring system is trained to characterize and/or locate
electrical faults based on operating state (OS) patterns learned
from transmission of test signals generated within the
distributions system. During operation, radio frequency emissions
from the EDS are evaluated by multiple radio frequency monitors
(RFM) as safe operation (SO) or potential electrical faults by
accessing a base of stored knowledge concerning fault emission
characteristics and/or synchronized time of arrival at each
monitor. During evaluation the EFM system establishes one or more
levels of confidence factors whether the monitored RF emission
characteristics are indicative of electrical distribution system
safe operation (SO) condition or an electrical fault its likely
location within the EDS. Information in the base of stored
knowledge is updated to include new operating states (OS) of the
electrical distribution system. Confidence level associations,
location of new radio frequency emission patterns and whether those
patterns are indicative of safe operating (SO) conditions or
electrical faults are stored in the base of stored knowledge.
[0003] 2. Description of the Prior Art
[0004] Electrical conductors produce an electric field with the
application of an electrical current. The electrical field profile
can be affected by a variety of characteristics of the conductor.
One characteristic that can affect the field is cracking in the
conductor causing small electrical fault arcs to jump across the
crack. Additionally variations in the field may also be caused by
imperfections in the insulation that can allow partial electrical
fault discharges consisting of very small mini-arcs on the surface
of the insulation. An electrical fault arc produces a burst of
broad spectrum radio frequency energy as the arc is made and again
as it is broken. This applies to various types and severity of
electrical fault partial discharge, discharges across bearing oil
films due to shaft grounding or pedestal insulation problems,
cracks in conductors, or any other kind of arcing. The signature
characteristics of the RF emission will be different depending on
the source.
[0005] Currently, there are several commercially available devices
that monitor RF emission activity, specifically for power plant
electrical generators. Partial discharge detectors monitor RF
activity on the main conductors that carry electricity away from
the generator (phase buses) and are typically located between the
generator and the main transformer within the power plant
electrical distribution system (EDS). They can consist of single
sensors or multiple sensors on each phase. The use of multiple
sensors can determine if an RF pulse is emanating from the
generator or outside the generator by determining the direction the
pulse is traveling. An existing known radio frequency monitor (RFM)
available from Siemens Energy Inc., detects RF activity at a single
point on the generator neutral ground. This approach is lower cost
than others are, and will detect electrical fault and other RF
activity from other pieces of electrical equipment in the power
plant electrical distribution system (EDS), because all the
equipment is typically tied to a common EDS ground or earthed phase
at some plant location. However, the RFM does not determine the
location of the equipment that is causing the RF emission activity:
only its existence. Various manual tests and inspections must be
made to locate specifically the source of the RF emission anomaly.
The fault locating Utility customers find the process cumbersome
and do not want the additional burden of finding the source of the
problem. Typical plant maintenance practice indicates that the
source of such problems become known eventually, usually when the
electrical equipment in the EDS fails.
[0006] RFM's and other detection systems, such as partial discharge
monitors, utilize either bus couplers or other means of voltage
isolation so they can directly measure voltage waveforms of power
plant equipment. The RFM output is typically observed as a series
of RF spikes with quieter periods of lesser activity in
between.
[0007] Existing electrical fault detection systems and methods do
not efficiently isolate location of RF emissions in power plant
electrical distribution systems. Those systems do not efficiently
characterize whether an RF emission is an indicia of an existing or
incipient electrical fault or whether the RF emission is merely
part of an EDS operating state that does not negatively impact its
safe operation (SO). If the latter, the detected RF emission OS can
be discounted as a SO that is acceptable for power plant EDS.
SUMMARY OF THE INVENTION
[0008] Accordingly, a suggested object of the invention is the
creation of a power plant electrical distribution system (EDS)
electrical fault monitor (EFM) system and method, for detecting
radio frequency (RF) emissions and their location within the EDS
with distributed radio frequency monitors (RFM) and in real time
associating the RF emissions with EDS safe operation (SO) or
electrical faults, that is flexibly employed and reconfigurable
within the power plant, to accommodate changes and upgrades in the
EDS configuration and component hardware.
[0009] Another suggested object of the invention is the creation of
a power plant electrical distribution system (EDS) electrical fault
monitor (EFM) system and method for detecting such EDS electrical
faults that self-adapts and "learns" programmed and real time
operational radio frequency (RF) emissions in the EDS that are
detected by radio frequency monitors (RFM) and determines whether
it associates the RF emissions with EDS safe operation (SO) or
electrical faults in real time.
[0010] An additional suggested object of the invention is the
creation of a power plant electrical distribution system (EDS)
electrical fault monitor (EFM) system and method for detecting such
EDS electrical faults that stores "learned" or programmed RF
operational state (OS) data in a base of stored knowledge for
future reference by the EFM, so that the stored knowledge is
subsequently used by the EFM to establish confidence levels whether
the OS data are indicative of safe operation (SO) or an electrical
fault.
[0011] These and other objects are achieved in one or more
embodiments of the invention by a power plant electrical
distribution system (EDS) electrical fault monitor (EFM) system and
method, for detecting radio frequency (RF) emissions with
distributed radio frequency monitors (RFM) and in real time
associating the RF emissions with EDS safe operation (SO) or
electrical faults, that is flexibly employed and reconfigurable in
conjunction with a distributed RFM network within the power plant,
to accommodate changes and upgrades in the EDS configuration and
component hardware. The EFM stores the RF operational state (OS)
data in a base of stored knowledge for future reference by the EFM.
The OS data in the base of stored knowledge can be pre-loaded
therein based on operational observations of EDS in other power
plants. In some embodiments, OS data in the base of stored
knowledge includes rules, set points, absolute reading limits, RF
waveforms, and time differences between receipt of data samples by
different RFMs, historical information, weightings, human inputs or
statistical probabilities. The EFM subsequently uses OS information
in the base of stored knowledge for associating the RF emissions
with EDS safe operation (SO) or electrical faults in real time. In
embodiments of the invention, the EFM determination of whether RF
emissions are associated with SO or electrical faults is performed
by the EFM using any of fuzzy logic, self-learning, algorithmic,
statistical, neural network logic, case-based reasoning or
heuristic methods. In embodiments of the invention the EFM updates
the base of stored knowledge with new OS data, determinations of
whether the OS data are associated with SO or an electrical fault,
RF emission location within the EDS and confidence level
determinations.
[0012] Embodiments of the invention feature methods for monitoring
electrical faults in a power plant electrical distribution system
(EDS) by monitoring and sampling in real time with an EDS
electrical fault monitor (EFM) operation state (OS) radio frequency
(RF) emissions data from EDS equipment. The emissions data are
obtained with a plurality of time synchronized radio frequency
monitors (RFM), which are distributed about the EDS and coupled to
the EFM. The OS RF emissions are potentially indicative of
electrical faults in the EDS. OS data samples for each RFM are
stored in an automated data storage device that is coupled to the
EFM. Exemplary data samples include identification of the RFM that
obtained the sample, data sample time and data sample waveform
characteristics. The EFM determines in real time a likelihood of
whether one or a combination of the respective OS data samples is
indicative of an electrical fault by referencing in an automated
data storage device coupled to the EFM previously stored
information associating OS data with any one or more of EDS safe
operation (SO), electrical faults or RF emission location within
the EDS. The EFM compares at least one stored OS sample reading
from each RFM with respective stored association information
relevant thereto and makes respective EDS safe operation (SO) first
confidence level determinations. The EFM compares a combination of
the at least one stored OS sample reading from each respective RFM
with respective stored association information relevant thereto, if
such combination information is available, and makes respective EDS
safe operation (SO) second confidence level determinations.
Optionally additional confidence level determinations can be made.
The EFM combines all prior sequentially determined confidence
levels information to derive an EDS safe operation (SO) overall
confidence level. The EFM generates an alarm when any of the
sequentially determined first through overall SO confidence levels
is indicative of an electrical fault within the EDS.
[0013] Other embodiments of the invention feature a method for
monitoring electrical faults in a power plant electrical
distribution system by monitoring and sampling in real time with an
EDS electrical fault monitor (EFM) operation state (OS) radio
frequency (RF) emissions data from electrical distribution system
(EDS) equipment that are obtained with a plurality of time
synchronized radio frequency monitors (RFM) distributed about the
EDS and coupled to the EFM. The OS RF emissions are potentially
indicative of electrical faults in the EDS. An automated data
storage device that is coupled to the EFM stores OS data samples
for each RFM. The sampled data including the RFM identification,
data sample time and data sample waveform characteristics. The EFM
determines in real time a likelihood of whether one or a
combination of the respective OS data samples is indicative of an
electrical fault. The determination is performed by referencing in
the automated data storage device that coupled to the EFM
previously stored electrical fault information obtained by the
RFMs, resulting from generating electrical fault RF test signals
with an RF test generator in the EDS. The sampled OS data are
compared with the previously stored electrical fault information in
the automated data storage device. The EFM generates an alarm when
the referenced stored information associates the sampled OS data
with an electrical fault within the EDS.
[0014] Additional embodiments of the invention feature a system for
monitoring electrical faults in a power plant electrical
distribution system, including a plurality of time synchronized
radio frequency monitors (RFM) that are distributed about the
electrical distribution system (EDS). The RFMs monitor and sample
in real time EDS operation state (OS) radio frequency (RF)
emissions data from electrical EDS equipment. The OS RF emissions
are potentially indicative of electrical faults in the EDS. An
electrical fault monitor (EFM) is coupled to the RFMs and an
automated data storage device, and has stored therein sampled data
readings for each RFM. The sampled data includes the RFM
identification, data sample time and data sample characteristics.
The EFM determines in real time a likelihood of whether one or a
combination of the respective OS data samples is indicative of an
electrical fault by referencing in the automated data storage
device previously stored information associating OS data with any
one or more of EDS safe operation (SO), electrical faults or RF
emission location within the EDS. The EFM compares at least one
stored OS sample reading from each RFM with respective stored
association information relevant thereto and makes respective EDS
safe operation (SO) first confidence level determinations. A
combination of the at least one stored OS sample reading from each
respective RFM is compared with respective stored association
information relevant thereto, if such combination information is
available, in which case the EFM makes respective EDS safe
operation (SO) second or more confidence level determinations. The
EFM combines all prior sequentially determined confidence levels
information to derive an EDS safe operation (SO) overall confidence
level; and causes the EFM to generate an alarm when any of the
sequentially determined first through overall SO confidence levels
is indicative of an electrical fault within the EDS.
[0015] When utilizing the EFM system and method embodiments of the
invention, faulty electrical equipment in the EDS, including the
generator, can be repaired during a scheduled period of down time
resulting in higher power plant reliability.
[0016] The disclosed invention utilizes an initialization testing
method (see FIG. 2) that will produce a series of typical RFM data
signatures that can be used to locate the source. A portable RF
source will be used to produce a pulse at different locations
around the power plant, for example, at the generator neutral
ground, at the main transformer primary side phase A, at various
motors, etc. The portable device will also have a precise time
measurement similar to the RFMs, so the exact time of the pulse
initiation is known. The arrival times and magnitude of the
generated spikes at each RFM location will form a vector for
storage in one or more of the storage devices 32, 34, 58.
[0017] The respective objects and features of the present
invention, including the exemplary embodiments described herein,
may be applied jointly or severally in any combination or
sub-combination by those skilled in the art.
BRIEF DESCRIPTION OF THE DRAWINGS
[0018] The teachings of the present invention can be readily
understood by considering the following detailed description in
conjunction with the accompanying drawings, in which:
[0019] FIG. 1 is a schematic diagram of an embodiment of a power
plant electrical distribution system including an electrical fault
monitor system (EFM) of the invention; for identifying and locating
faults in the electrical distribution system;
[0020] FIG. 2 is a block diagram of an embodiment of the EFM of
FIG. 1;
[0021] FIG. 3 is a block diagram of an embodiment of a radio
frequency test generator of the invention that is used to introduce
RF test signals into the EDS during EFM "learning" mode, in order
to populate the EFM's stored information associating the EDS
operating state (OS) data with any one or more of EDS safe
operation (SO), electrical faults or RF emission location within
the EDS;
[0022] FIG. 4 is a flowchart schematic diagram of an embodiment of
a method for operating the radio frequency test generator of FIG. 3
and the EFM of FIGS. 1 and 2 within the power plant EDS, in order
to populate the EFM's stored information associating the EDS
operating state (OS) data with any one or more of EDS safe
operation (SO), electrical faults or RF emission location within
the EDS; and
[0023] FIGS. 5 and 6 are flowchart schematic diagrams of
embodiments of methods for operating an EFM in a power plant EDS
using the system of FIGS. 1 and 2.
[0024] To facilitate understanding, identical reference numerals
have been used, where possible, to designate identical elements
that are common to the figures.
DETAILED DESCRIPTION
[0025] After considering the following description, those skilled
in the art will realize that the by the teachings of the present
invention electrical faults are detected in electrical distribution
systems (EDS) by detection and location of radio frequency (RF)
emissions generated by the fault with multiple time synchronized
radio frequency monitors (RFM) distributed about the EDS. In
embodiments of the invention the RFM location is reconfigurable to
accommodate changes in the EDS configuration and/or component
hardware. The RFMs are coupled to a self-learning, electrical fault
monitor (EFM) that characterizes and/or locates electrical faults
based on operating state (OS) patterns learned from transmission of
test signals generated within the in the distributions system
and/or that are preloaded into an EFM accessible base of stored
knowledge. RF emissions data samples are characterized as safe
operation (SO) states or potential electrical faults by accessing a
base of stored knowledge concerning fault emission characteristics
and/or synchronized time of arrival at each RFM. In embodiments of
the invention information in the base of stored knowledge is
updated to include new EDS operating states (OS). In embodiments of
the invention the confidence level associations, location of new
radio frequency emission patterns and whether those patterns are
indicative of safe operating (SO) conditions or electrical faults
are stored in the base of stored knowledge.
[0026] In embodiments of the invention, specific RF emission spikes
in the EDS are detected by the RFM network under control of the EFM
30, recording the precise and synchronized measurement of time in
addition to at least the magnitude of measured RF spikes, saving
the series of RF magnitudes and time readings in a central database
as a vector, and comparing that vector to a set of vectors recorded
during initialization testing with the TE that will allow advanced
pattern matching algorithms to determine a series of likely
locations with confidence factors (see FIG. 6). The initialization
testing will automatically account for timing not only of the RF
spikes through the power plant's structure, but also the timing of
the RFM signals themselves through internal or external data
transfer wiring, fiber optic cables, wireless data transfer
technologies, etc. This testing will compensate for such
miscellaneous delays inherent in the entire EFM system, as the RFM
system will be recording the initialization test signals in exactly
the same way it records the actual RF emission spikes or signals
during subsequent EMF live operation within the EDS.
[0027] In embodiments of the invention during operation of the EDS
the EFM determines in real time a likelihood of whether one or a
combination of the respective OS data samples is indicative of an
electrical fault by referencing in the automated data storage
device previously stored information associating OS data with any
one or more of EDS safe operation (SO), electrical faults or RF
emission location within the EDS. The EFM compares at least one
stored OS sample reading from each RFM with respective stored
association information relevant thereto and makes respective EDS
safe operation (SO) first confidence level determinations. A
combination of the at least one stored OS sample reading from each
respective RFM is compared with respective stored association
information relevant thereto, if such combination information is
available, in which case the EFM makes respective EDS safe
operation (SO) second or more confidence level determinations. The
EFM combines all prior sequentially determined confidence levels
information to derive an EDS safe operation (SO) overall confidence
level; and causes the EFM to generate an alarm when any of the
sequentially determined first through overall SO confidence levels
is indicative of an electrical fault within the EDS.
[0028] Electric Fault Monitor (EFM) Hardware Description
[0029] An embodiment of the system of the invention as employed in
an exemplary known power plant electrical distribution system (EDS)
10, having three phases .PHI.A, .PHI.B, .PHI.C and neutral N is
shown in FIG. 1. The EDS 10 includes may include one or more of
exemplary known gas turbine 12, generator 14, turning gear motor
16, motors 18, transformers 20, pumps 22, EDS protection
distribution and control switch gear/circuit interrupters/circuit
breakers 24, power plant instrumentation and control system 26 and
any other EDS equipment components.
[0030] Referring to FIGS. 1 and 2, an embodiment of an exemplary
electrical fault monitoring system (EFM) comprises a plurality of
radio frequency monitors RFM-1-RFM-N that are selectively
distributed about the power plant EDS 10, where they monitor RF
emissions from EDS equipment components (e.g., generator 14, motors
16 or 18, transformers 20, pumps 22, switchgear 24, etc.). The EFM
also comprises an EFM control system 30 that is communicatively
coupled to the RFMs; an EFM network database 32 that includes a
base of stored knowledge for determining wither an RF transmission
received by an RFM is associated with an EDS 10 safe operation (SO)
state or an electrical fault; any other EFM-accessible database(s)
34; wireless transceiver 36; and EFM data network 38.
[0031] The EFM data network 38 that is shown in FIGS. 1 and 2
includes exemplary known bi-directional networked data busses,
wireless transceivers (e.g., RFM-N to transceiver 36), and other
known types of conductive communications pathways (fiber optic,
hard wired, twisted pair cables, coaxial cables, etc.). In the data
network 38 the radio frequency monitors RFM-1-RFM-N are selectively
distributed about the power plant EDS 10 and can be reconfigured
selectively to accommodate alterations in the EDS configuration
and/or electrical components by coupling sensors 40-1-40-N to
electrical equipment or power distribution phase busways between
electrical equipment. As shown in FIG. 2, exemplary RFM sensors
include multi-phase arc sensor 41-1 to RFM-1 and single-phase
sensor 41-3 to RFM-3. Also in FIG. 2 an exemplary wireless
transceiver 36 is communicatively coupled to RFM-N and the
controller 30 by way of an Input/output interface 48. As shown the
same I/O interface 48 is also coupled to the network database known
automated storage device 32 (e.g., data processing system hard
drive or other mass data storage device) by communications pathway
44. Communication between the controller 30 and data busses of the
data network 38 is also facilitated by communications module
interface 46.
[0032] As shown in FIG. 2, the EFM control system 30 (also referred
to herein as the RFM Network Controller) includes a controller 50
that has a processor 52, memory 54 coupled to the processor and
software modules 56 that when executed by the processor causes the
system to monitor power plant EDS operating conditions, including
respective RF emissions waveforms detected by the RFMs, and
determine the likelihood of whether the monitored RF emission
operating states, alone or in combination, are indicative of an EDS
safe operating (SO) condition or an electrical fault, using the
steps subsequently described herein. The controller 50 may be
incorporated in the power plant instrumentation and control system
26 or as a stand-alone dedicated hardware platform, such as a known
programmable logic controller or a personal computer utilizing a
known operating system. The controller 50 and its processor 52 have
access to a data storage device 58, which may include a disc drive,
non-volatile solid state memory, firmware or the like. The storage
device 58 may be incorporated into the controller 50 or be coupled
to and remotely accessible to the controller. Data storage device
58 includes previously stored information (loaded or "learned"
during EDS operation) that associates respective types of power
plant EDS RF emission operating states (OS) with EDS safe operation
or a potential electrical fault. The stored information may include
by way of non-limiting example rules, set points, absolute reading
limits, historical information, weightings and statistical
probabilities. The storage device may extract, transfer or share
previously stored information with any of the other communicatively
coupled data storage devices 32 or 34. A human machine interface
(HMI) is coupled to the controller 50, and may include a visual
display or monitor 60, a keyboard 62, a mouse or joystick 64 or
other known HMI devices (e.g., tablet or personal computer). The
HMI may be used to configure or monitor the EFM, modify the stored
information in the system data storage device 58 or other storage
devices 32, 34, and allow human operator interaction and control
over the EFM system.
[0033] In response to an observed operating condition indicating an
electrical fault in the EDS, the EFM system, directly or indirectly
through the plant control system 26 may automatically operate,
enunciate or alarm for human operation output devices. For example
the EFM system may automatically communicate electrical power
disconnection commands to the switchgear 24 to trip circuit
interrupters/circuit breakers if the electrical fault conditions
warrant such action but under lesser detected fault conditions send
an alarm communication to the power plant operators for their
evaluation and subsequent action. As previously noted the
controller 50 is respectively in communication with the various
plant system detection and control devices by respective
communications pathways, which may include by way of non-limiting
example wireless signal, known metallic cable including twisted
pair wiring, co-axial, multiple conductor wire cable, optical cable
and data busses employing known communications protocols.
Additionally the controller 50 may be in communication with other
devices within or remote from the power plant including by way of
non-limiting example wireless receiver/transmitters, other facility
command and control systems, remote monitoring stations and remote
data storage/web hosting/data access sites.
[0034] In accordance with embodiments of the invention the EFM
control system 30 "learns" RF emission operation states (OS) of the
EDS 10 by generating test RF signals S.sub.IN at selected locations
and times about the EDS with one or more known calibration testers
TE-1-TE-N by tester output 70 and test signal transducer 72 into
one or more of the EDS phases .PHI.A-C, N. One or more time
synchronized RFMs monitor the resultant emission signal or
signature S.sub.OUT via associated sensor 41 and transmit OS
information about that resultant emission to the EFM control system
30 by way of a previously described data network 38 communications
pathway. Resultant RF emission S.sub.OUT exemplary OS information
from the RFMs may include time of signal receipt, which when
synchronized with the TE S.sub.IN time of emission provides
time-of-flight OS information useful for the EFM control system 30
to time associate location of the RF emission S.sub.IN relative to
the associated RFM that recorded the emission signature S.sub.OUT.
Other emission signature S.sub.OUT exemplary OS information
includes emission waveform capture potential, frequency
distribution and amplitude characteristics. Additional OS
information and parameters can be added to RFM data capture in
order to enhance potential electrical fault identification, such as
the OS RF pulse main frequency component, pulse shape
classification, descriptive parameter and rates of attenuation for
various frequency components. For convenience the S.sub.OUT test RF
information and any other RF emissions monitoring information
conveyed by the RFMs to the EFM control system 30 may comprise
vectors including the synchronized time information t.sub.1-N and
waveform characteristics V.sub.1-N in any known communications
packaging and transmission protocol. The term "signal" is used
herein interchangeably with the term spectral characteristics
respecting the RF emissions and is not intended to imply that the
RF emission is being used as a carrier medium for other types of
communication any more than intention the raw power waveform
distributed within the EDS 10 acts as an information communications
medium.
[0035] Electric Fault Monitor (EFM) Operation
[0036] As previously noted the EFM system 30 monitors operation
states (OS) of power plant EDS 10 power transmission equipment in
real time and stores OS samples. The EFM system 30 determines a
likelihood of whether a monitored OS reading, alone or in
combination with other readings, is indicative of an EDS 10 safe
operation (SO) condition with sufficient safety margin for
statistical variations or whether the OS reading constitutes an
electrical fault. As a general overview, the EFM 30 performs the
following tasks in any order. The EFM system 30 references
previously stored information in an information storage device that
associates respective types of OS readings from each monitored RFM
with EDS operating or operation states (OS), and establishes first
safe operation (SO) determination confidence levels (CL). If such
information is available optionally the system also compares a
combination of OS sample readings from multiple RFMs and/or over
multiple time sampling windows and establishes a second SO
confidence level, and potentially additional sequential SO
confidence levels. The EFM system 30 also combines all desired
confidence levels information to derive an overall confidence level
of whether the EDS 10 RF emissions monitored by the RFMs jointly or
severally are within a safe operating (SO) condition or are
indicative of an electrical fault. An enunciation, alarm or
automated reaction response is initiated if the overall confidence
level does not indicate a likelihood of safe operation, depending
upon the severity of the monitored situation. For example, a fault
determination above a defined threshold may cause automatic
switchgear 24 tripping to interrupt electrical power transmission
within all or a defined sub-portion of the EDS 10, whereas a fault
determination above a lower defined threshold may initiate a lower
level of urgency response (e.g., alarm or other enunciation),
depending upon the overall confidence level.
[0037] Preferably, the present invention is implemented in software
as a program or plural program modules tangibly embodied on a
program storage device. The program may be uploaded to, and
executed by, a machine comprising any suitable architecture. As
previously described herein, preferably the controller 50 is
implemented on a computer platform having hardware such as one or
more central processing units (CPU), a random access memory (RAM),
and input/output (I/O) interface(s). The computer platform also
includes an operating system and microprocessor instruction code.
Both the various processes and functions described herein may be
part of the microprocessor instruction code or part of the program
(or combination thereof), which is executed via the operating
system. In addition, various other peripheral devices may be
connected to the computer platform such as an additional data
storage device 32, 34 or 58 and various output devices.
[0038] It is advantageous to initialize the EFM system 30 base of
stored knowledge (e.g., in the automated storage devices 32, 34 or
58) with preloaded operating state (OS) data and/or other data.
FIG. 4 shows and exemplary embodiment of the test sequence for the
test equipment for transmitting an RF emission S.sub.IN within the
EDS 10 and capturing the RF emission characteristics S.sub.OUT with
the RFMs. In step 80 one or more test equipment apparatus TE is
moved about to different designated locations TE-1-TE-N within the
EDS 10 and coupled to EDS equipment, as shown in FIGS. 1 and 3. In
step 82 the TE generates one or more RF pulses S.sub.IN at
designated times and the transmitted data S.sub.OUT, is recorded by
the RFM. S.sub.IN data qualities are verified at step 84 by either
or both of the RFM and the controller 50 of the EFM system 30.
Examples of data quality verification include by way of
non-limiting example whether the RF spike pulse was detected,
amplitude of the waveform and timing within designated limits,
network data transmission accuracy and waveform characteristics.
Unsatisfactory data are rejected and the pulse transmitted
repeated. Satisfactory quality S.sub.OUT data are saved at step 86
in one or more of the data storage devices 32, 34, 58 in a desired
format, such as a vector including: the RFM identification; time t
of the pulse receipt (synchronized with TE pulse generation time
and other RFM sampling and/or clocks); and desired waveform
characteristic(s) V, and transferred to the EFM 30 controller 30.
The RFM OS data can be represented as single or multidimensional
arrays or vectors. The data can be stored in any format such as
flat files, relational database, or other means. Various techniques
for matching operational data sets to the TE testing acquired
initialization OS data sets can be used. These variations in
techniques result in a similar outcome of determining a list of
potential RF sources with confidence factors, by comparing a
pattern of RF data from a network of RFM's to expected patterns
determined by initialization testing with the TE and resultant EFM
system 30 "learning". Additional data derived from other sources
(such as learned from operation of other power plant electrical
distribution systems within the same or different power plants or
laboratory experimentation) may be preloaded into the EFM 30 data
storage devices 32, 34, 58 to pre-populate the base of stored
knowledge before or any time after initial operation of the EFM/RFM
network controller 30.
[0039] Monitoring sequence of the EFM/RFM network controller 30
during EDS 10 electrical fault monitoring are shown in the flow
charts of FIGS. 5 and 6. In step 90 the RFMs feed operational state
(OS) data to the EFM 30 and data quality are verified in step 92 as
previously described in connection with the test equipment TE test
pulse sequence. Satisfactory quality OS data meeting amplitude,
timing, network quality, waveform analysis and vector archiving
standards are stored in the data storage devices, such as internal
data storage device 58 of the EFM 30 or the shown data storage
devices 32, 34. If data quality do not meet standards an alarm
system message is sent to the instrumentation and control system 26
or other output location in step 94. Satisfactory data are
archived. When an RF emission spike within the EDS 10 is detected
by one or more RFMs the transmitted operation state OS data are
evaluated by the EFM 30 at step 98. The evaluation includes
received RF pulse waveform analysis of RF spikes detected at plural
locations by different RFMs, RF pulse location mapping by pulse
receipt time synchronization interpolation between different RFM
nodes or other known location interpolation methods. All RFM's
preferably having a synchronized time reading or synchronized
sampling rates, so an exact time, such as GMT in nanoseconds, can
be recorded and compared. Such a time representation can be
accomplished by the use of atomic clocks, which are now readily
available and low cost and sometimes exist on a single chip, or by
the use of multi-channel RFM analyzer that is capable of reading
all the RFM signals from multiple sensor locations within the power
plant EDS simultaneously. Typically in EDS operation the RF pulse
is series of salient spikes that can be tracked by multiple RFMs at
precise times as the spikes propagate through the power plant and
is read at various locations, so that the RF spike source can be
identified by subsequently reviewing the track path and timing.
[0040] OS data vectors from prior EFM system 30 initialization TE
test locations are stored in a database within one or more of the
data storage devices 32, 34, 58 that will also house the live
monitored RFM operating data. If an RF pulse is detected by any RFM
in the network of RFM's, the EFM 30 analysis software will attempt
to evaluate and match the resulting pulse location vector(s) with
those stored during the initialization testing. The OS electrical
fault or OS evaluations are performed by the EFM 30 controller 50
by known evaluation methodologies, such as rules, set points,
absolute reading limits, historical information (including human
experiential and judgmental historical information) fuzzy logic,
case based reasoning, self-learning neural networks, Principle
Component Analysis, or other input or rule-based process. The
evaluation of OS data taken at different RFM times and locations
may be performed in parallel or sequentially for each OS, or in any
combination or sub-combination desired for a monitoring function.
The RFM will then produce a list of one or more locations, each
with a confidence factor based on how closely the vector(s) match
initialization data. If the RFM pulse vector does not precisely
match any of the TE test stage or other pre-loaded initialization
vectors, a list of potential sources will be produced, each with a
confidence factor based on how closely the data vector matches
those in the set of the initialization vectors. A location between
initialization locations, or related to specific locations, can be
surmised by a data vector that produces lower confidences in two or
more locations.
[0041] In embodiments of the invention the OS evaluation process
performed in step 98 includes one or more sequential confidence
factor determinations of whether the RF emission OS data samples
for each RFM are indicative of EDS 10 safe operation (SO) or an
electrical fault, as further described in conjunction with the more
detailed flow chart of FIG. 6, in which step 98 is further
explained with the 9800 et seq. series of steps. In step 9805 the
OS data are analyzed by referencing the previously stored
information associating OS data with electrical faults and safe
operation (SO). The EFM 30 determines whether the OS data exceeds a
"hard" set point or absolute ceiling reading (e.g., voltage,
current, and rate of current change, detected frequencies or
frequency spectrum) or a previously learned safe threshold in step
9820, in which case the OS data meet alarm criteria in step 100.
When meeting alarm criteria, the EFM 30 may undertake a range of
action, depending upon the severity of the determined electrical
fault, that warrants action ranging from tripping or interrupting
electric power transmission in the relevant parts of the EDS with
switchgear 24 or in step 102, sending an alarm with location of the
detected fault to the plant operators and/or the instrumentation
and control system 26 for further evaluation. As part of the alarm
criteria determination of step 100 the event is logged by the EFM
30 and the relevant OS waveform and/or other OS data are captured
for storage.
[0042] Referring back to FIG. 6, if OS data do not exceed hard set
points or learned safe thresholds, a first confidence level of safe
operation (SO) is established and stored and added to the SO model
in the base of stored knowledge. Confidence level determination is
performed by the EFM 30 without human intervention using any one or
more of fuzzy logic, self-learning, and algorithmic, statistical or
heuristic methods in the controller 50. The confidence level
evaluations are performed in real time for other OS data received
from the same RFM and other RFMs. In step 9830 one or more
combinations of different OS data from the same RFM and/or other
RFMs are evaluated by referencing the base of stored information
associating combinations of OS data with SO and electrical fault
determinations. For example, a combination of RF spike OS detected
in the gas turbine 12 electrical system by RFM-2 and RF spike OS
simultaneously detected in the generator 14 electrical windings by
RFM-3 may be indicative of an impending generator malfunction that
exceeds a learned state threshold in step 9835, necessitating
taking the turbine/generator pair offline. If the combination
learned state threshold is not exceeded one or more second
confidence levels of safe operation (SO) may be established and
added to the SO model in the base of stored knowledge as new EFM
system learning. Additional sequential N confidence levels (NCL)
may be analyzed in steps 98N-1-98N-3. In step 9850 all prior
confidence levels are analyzed in combination by referencing
previously stored information, which may include other power plant
EDS or EDS in other power plant locations to establish an overall
confidence level in step 9855. If the overall confidence level
indicates sufficient safe operation SO safety margin the stored
information in the base of stored knowledge (e.g., in the data
storage devices 32, 34 or 58) is updated to include the newly
learned additional SO confidence information for future operational
use by the EFM.
[0043] Although various embodiments, which incorporate the
teachings of the present invention, have been shown and described
in detail herein, those skilled in the art can readily devise many
other varied embodiments that still incorporate these teachings.
The invention is not limited in its application to the exemplary
embodiment details of construction and the arrangement of
components set forth in the description or illustrated in the
drawings. For example, in order to facilitate a clear understanding
of the present invention, illustrative examples have been provided
herein which described certain aspects of the invention system
hardware and architecture. However, it is to be appreciated that
these illustrations are not meant to limit the scope of the
invention, and are provided herein to illustrate certain concepts
associated with the invention. It is also to be understood that
embodiments of the present invention may be implemented in various
forms of hardware, software, firmware, special purpose processors,
or a combination thereof. It is to be understood that, because some
of the constituent system components and method steps depicted in
the accompanying figures are preferably implemented in software,
the actual connections between the system components (or the
process steps) may differ depending upon the manner in which the
present invention is programmed. Specifically, any of the computers
or devices may be interconnected using any existing or
later-discovered networking technology and may also all be
connected through a larger network system, such as a corporate
network, metropolitan network or a global network, such as the
Internet. Also, it is to be understood that the phraseology and
terminology used herein is for the purpose of description and
should not be regarded as limiting. The use of "including,"
"comprising," or "having" and variations thereof herein is meant to
encompass the items listed thereafter and equivalents thereof as
well as additional items. Unless specified or limited otherwise,
the terms "mounted," "connected," "supported," and "coupled" and
variations thereof are used broadly and encompass direct and
indirect mountings, connections, supports, and couplings. Further,
"connected" and "coupled" are not restricted to physical or
mechanical connections or couplings.
* * * * *