Distribution Fault Anticipation (DFA) Technology

boorishadamantΤεχνίτη Νοημοσύνη και Ρομποτική

29 Οκτ 2013 (πριν από 3 χρόνια και 11 μήνες)

51 εμφανίσεις

Copyright © 2011. The Texas A&M University System.

1

Distribution Fault Anticipation

(DFA) Technology

John S. Bowers, PE



Carl L. Benner, PE

Vice President of Operations


Senior Research Engineer

Pickwick Electric Cooperative


Dept. of Electrical and Computer Engineering

PO Box 49, 530 Mulberry Street


3128 TAMU, Texas A&M University

Selmer, TN 38375



College Station, TX 77843
-
3128

731
-
646
-
3766



979
-
845
-
6224

jbowers@pickwick
-
electric.com


carl.benner@tamu.edu


DFA Technology Success Stories are available at:

https://epridfa.tamu.edu/DFAReports/DFASuccess.aspx


Second Annual Smart Grid Research Consortium Conference:

“Evaluating the Business Case for Smart Grid”

sponsored by Smart Grid Consortium

Orlando, Florida, October 21, 2011

Copyright © 2011. The Texas A&M University System.

2

Overview of DFA Technology


Background
: Technology’s foundation lies in EPRI’s DFA
(Distribution Fault Anticipation) project, performed by Texas
A&M, in close cooperation with multiple utilities. DFA label
persists for historical reasons.


Substation
-
based DFA devices sense high
-
fidelity current
and voltage waveforms, using conventional CTs and PTs.


DFA devices execute on
-
line algorithms that analyze
waveforms to diagnose failures, maloperations, etc.


DFA system provides actionable reports about “important”
line activity.


The result
: Awareness of system conditions, including
incipient faults and outages.

Copyright © 2011. The Texas A&M University System.

3

Utility Partners

Hardware Demonstration Sites

Arizona Public Service

BC Hydro

Bryan Texas Utilities

CenterPoint Energy

ConEdison

CPS Energy

Exelon

KeySpan Energy

MidAmerican Energy

Northeast Utilities

Omaha Public Power District

Oncor Electric Delivery

Southern Company/Alabama Power

TVA/Pickwick Electric Cooperative

Other Previous/Current Partners

American Electric Power

Baltimore Gas & Electric (Constellation)

Central Hudson Gas & Electric

FirstEnergy

Public Service Electric & Gas

Copyright © 2011. The Texas A&M University System.

4

Conceptual Application of Intelligent

Algorithms to Electrical Waveforms


Intelligent Algorithms


(Analytics)

Reliability

Condition
-
Based

Maintenance

Asset Management

Improved

Safety

Outage Management

Forensics

Improved Power Quality

O&M Cost
Reduction

Fault Anticipation

Diagnosis of
Protection

Problems

Substation waveforms “know” about feeder activity. They can
provide system awareness, if we measure them with sufficient
fidelity and know how to interpret them.

Copyright © 2011. The Texas A&M University System.

5

Documented Failures


Voltage regulator failure


LTC controller maloperation


Repetitive overcurrent faults


Lightning arrester failures


Switch and clamp failures


Cable failures


Main substation cable


URD primary cables


URD secondary cables


Overhead secondary cables


Tree/vegetation contacts


Contacts with primary


Contacts with secondary
services


Pole
-
top xfmr bushing failure


Pole
-
top xfmr winding failure


URD padmount xfmr failure


Bus capacitor bushing failure


Capacitor problems


Controller maloperation


Failed capacitor cans


Blown fuses


Switch restrike


Switch sticking


Switch burn
-
ups


Switch bounce


Pack failure

Certain types of failures occur frequently and are well understood. Other types of
failures occur infrequently and only one or a few incidents have been
documented. Ongoing field experience enables continuous improvement and
next
-
generation DFA algorithms to diagnose those types of failures better.

Copyright © 2011. The Texas A&M University System.

6

We All Need More Data


Not!

Copyright © 2011. The Texas A&M University System.

7


On
-
Line

Signal Processing
and Pattern

Recognition

Algorithms




Performed by

DFA device in

substation

Line recloser*

tripped 8% of

phase
-
A load twice,

but reclosed and did

not cause outage

Failed 1200 kVAR

line capacitor*

(phase B inoperable)

Failing hot
-
line

clamp on phase B*

Inputs: Substation CT and PT Waveforms

*DFA system reports hydraulic reclosers, switched
line capacitors, line apparatus failures, etc, based
on substation waveforms, without requiring
communications to those devices.

DFA Reports

DFA Algorithms

Copyright © 2011. The Texas A&M University System.

8

DFA Web
-
Based Reporting Format


Reading Reported Protection Sequence (Hydraulic Line Recloser)
:


Single
-
phase line recloser operated three times and locked out.


Fault was phase
-
B and drew 745 amps.


Sequence was fast
-
slow
-
slow (3 cycles, 12 cycles, 10
-
1/2 cycles).


Open intervals were 2.1 and 2.4 seconds.


Each operation temporarily interrupted half of phase
-
B load.

(Sequence of events below was created by DFA, not by humans.)

Copyright © 2011. The Texas A&M University System.

9


Waveforms are available
for viewing and further
analysis, if desired.

DFA Web
-
Based Reporting Format

Copyright © 2011. The Texas A&M University System.

10

UTILITY EXPERIENCE WITH DFA

Utility Perspective

Copyright © 2011. The Texas A&M University System.

11

Pickwick Electric Cooperative


Rural electric coop


Headquartered in Selmer, Tennessee, home of Shiloh
National Military Park, Pickwick Dam, and Pickwick Lake


Southwest corner of Tennessee, 100 miles east of Memphis


Customer of Tennessee Valley Authority


21,000 customers; 2,000 miles of distribution


TVA’s participant in research behind DFA technology


EPRI
-
funded project at Texas A&M


Five feeders instrumented with DFA


Copyright © 2011. The Texas A&M University System.

12

MYSTERIOUS SERVICE PROBLEM
REQUIRES FOUR TRUCK ROLLS

Example

Copyright © 2011. The Texas A&M University System.

13

Example

Mysterious Service Problem


T=0 hours
: 16 customers with lights out


Blown tap fuse, but no obvious system failure.


New fuse holds. Close ticket.


T=36 hours
: Flickering lights in same area


Crew hears buzzing transformer.


Replace transformer. No buzzing. Close ticket.


T=38 hours
: Lights flickering again


Now what do I do?

Copyright © 2011. The Texas A&M University System.

14


For three weeks, the DFA system had been reporting a
failing clamp on this phase of this circuit.…




… but the dispatcher and crew

were unaware of the DFA report.


This was a missed opportunity.

Moving forward, providing

information to right people can

provide awareness and ...


Reduce complaints and truck rolls.


Reduce unnecessary equipment changeouts.

Example

Mysterious Service Problem

Copyright © 2011. The Texas A&M University System.

15

INCIPIENT FAILURES AND AVOIDABLE
OUTAGES

Example

Copyright © 2011. The Texas A&M University System.

16


6/03/06

First fault; no outage


6/10/06

Second fault; no outage


6/17/06

Third fault; no outage


6/24/06

Fourth fault; no outage


6/28/06

Similar but unrelated fault


7/04/06

Fifth fault; no outage


7/24/06

Sixth fault; outage


35 minutes, 903 customers


31,605 CMI

6/03/06

6/10/06

7/24/06

Example

Avoidable Outage


Without DFA

Copyright © 2011. The Texas A&M University System.

17


Similar faults two days apart


Next day


Alerted by DFA


Using DFA, found problem in 1 hr


Avoided consequences


Additional interruptions


Extended outage


Punch
-
through, moisture ingress


Lid launch, burning oil, fire hazard


7
th

day: Third fault prioritized repair


No outage, complaints, or further consequences

Example

Avoided Outage


With DFA

Copyright © 2011. The Texas A&M University System.

18

Step 1: Learn of recurrent fault from DFA


DFA reported identical faults, 18 days apart

Step 2: Compare DFA info to system model at
various reclosers (e.g., recloser
R
)


Protection


DFA:

Reported operation of 1ø recloser


Model:

R is bank of 1ø hydraulic reclosers


Momentary Load Interruption


DFA:

Estimated 19
-
21% load interruption


Model:

23% of load is beyond
R


Reclosing Interval


DFA:

Reported 2
-
second open interval


Model:

Reclosers at R have 2
-
second open


Conclusion
: Failure is downstream of recloser R
(26% of total feeder length).

Sub

R

Feeder NS 344

(139 circuit miles)

Detailed Example

Avoided Outage


With DFA

Copyright © 2011. The Texas A&M University System.

19


Initial patrol downstream of R visually
identified cracked dead
-
end bells (DB),
but…


DFA fault
-
current estimate @ DB: 510A


Model fault
-
current estimate @ DB: 1086A


Then, using DFA fault
-
current estimate of
510A, model targeted area outlined in oval

(~2% of total feeder).

R

Enlarged View of Line

Downstream of R

DB

X

Sub


NS 344

(139 circuit miles)

R

X

Copyright © 2011. The Texas A&M University System.

20

Sub


NS

344

(139 circuit miles)

R

Enlarged View of Oval

Manually integrating DFA information with system
model put failure in oval area, encompassing 4 spans
on either side of a tee. Searching that small area found
this failing arrestor. Replacing it avoided further
interruptions and likely outage to 53 customers.

Copyright © 2011. The Texas A&M University System.

21

Recurrent Faults


A Recap


DFA can detect failing apparatus and avoid outages.


DFA often provides sole notice of problem. Without that, nothing else matters.


Just counting faults on a feeder is insufficient. It is necessary to know that a
specific fault

is recurring.


Integrating DFA information with system model locates failure.


For the subject case, this located failure within four spans, on feeder with total of
139 circuit miles.


Process is straightforward, but manual application is tedious.


Might it be feasible to automate the process of integrating DFA
information with a system model, to streamline the process of
learning of and locating these pre
-
failures? It would seem so.


Important
: This process locates failures that have not caused
outages and that often have not generated customer calls,
thereby making it feasible to avoid many outages.

Copyright © 2011. The Texas A&M University System.

22

FAULTY CAPACITOR CONTROLLER
RESULTS IN EQUIPMENT DAMAGE
AND DEGRADED POWER QUALITY

Example

Copyright © 2011. The Texas A&M University System.

23

Example

Capacitor Problem


Without DFA


1/2004: 28 cycles per day


2/2004: 100 cycles per day


2/29/2004
-
3/3/2004


Switch contacts arced three days


Severe voltage transients


Failures in other cap banks


Final consequences


3,000 cycles in two months


Failed switch


Four damaged capacitor banks

Copyright © 2011. The Texas A&M University System.

24

Example

Capacitor Problem


With DFA


8/9/2004


Capacitor controller was changed
out during normal maintenance.


Hours after crew left, controller
began switching bank continuously.


8/10/2004


DFA reported 22 switching events.


Crew corrected controller settings.


Continuous switching stopped.

Copyright © 2011. The Texas A&M University System.

25

Feeder Lockouts Resulting from
Protection Miscoordination

(or Maybe Not!)

Example

Copyright © 2011. The Texas A&M University System.

26

Example

Improper Breaker Lockouts


Remote fault tripped mid
-
point line recloser, but
breaker tripped, too, locking out feeder.


Utility performed lengthy investigation.

Result: “Cause Unknown.”


Two years later, same thing happened again.


This time, the DFA system helped utility to
identify

and
locate

fault
-
induced conductor slap (FICS).


(
FICS
: Distant fault trips mid
-
point recloser, but
magnetic forces cause close
-
in wires to swing
together, creating second fault.)

Copyright © 2011. The Texas A&M University System.

27

Example

Improper Breaker Lockouts

Copyright © 2011. The Texas A&M University System.

28

Example

Improper Breaker Lockouts



DFA reports conductor
-
slap and enables location.


Result
: Utility can alter the offending line segment, thereby
preventing future breaker operations and feeder lockouts.

Copyright © 2011. The Texas A&M University System.

29

Another Conductor
-
Slap Example

Months

0 0.3 5 19

Facts (this case)
:


Four conductor
-
slap incidents


Eight unnecessary breaker trips


Same location on same feeder


Breaker tripped two times for each event


Phenomenon was not discovered by
conventional tools available to utilities

Ramifications (in general)
:


Conductor
-
slap and fault
will recur

in future.


Progressive conductor damage, with possibility of
conductor burn
-
down


Stress on all current
-
carrying components
(switches, connectors, …)


Fire hazard from burning aluminum particles

A significant finding has been that slap does not happen at random locations on feeders.
Locations experiencing slap once, left uncorrected, likely will experience it again and again,
repeatedly creating system stresses and the chance of prolonged outages, fire, etc.

Copyright © 2011. The Texas A&M University System.

30

INTERMITTENT FAILURE OF BREAKER
TO RECLOSE RESULTS IN MULTIPLE
SUSTAINED OUTAGES

Example

Copyright © 2011. The Texas A&M University System.

31

Example

Feeder Breaker Fails to Reclose


04/08/10: Breaker operates properly.


08/06/10: Breaker operates properly.


10/19/10: Breaker
fails to auto
-
reclose
.


12/10*: Standard tests show no problem.


01/08/11: Breaker
fails to auto
-
reclose
.


03/06/11: Breaker operates properly.


03/18/11: Breaker
fails to auto
-
reclose
.


03/26/11: Breaker operates properly.

Copyright © 2011. The Texas A&M University System.

32

Examples
:

Other Anomalies


Case 1
: Line burned down past line recloser and auto
-
sectionalizer. DFA helped determine which device failed.


Case 2
: In multiple cases, utility has used DFA to determine
whether questionable reclosers were functioning properly,
without removing them for offline testing.


Case 3
: DFA report is helping ongoing investigation to diagnose
why a self
-
healing circuit did not respond properly.


Case 4
: DFA helped utility identify that a three
-
phase recloser had
stopped mid
-
sequence, failing to reclose or lock out, creating
hazard for a crew that assumes open recloser is locked out.

Copyright © 2011. The Texas A&M University System.

33

Possible Future Algorithms


Continual process of discovering new signatures and
expanding capabilities is inherent in DFA technology’s
future roadmap.


Fundamental design of system anticipates and
accommodates future extensions of capabilities.


For example, Texas A&M currently is evaluating potential
unique characteristics for…


Lightning
-
induced faults.


Arrester
-
induced faults.

Copyright © 2011. The Texas A&M University System.

34

Benefit Estimates (Partial List)

Copyright © 2011. The Texas A&M University System.

35

Conclusions


DFA technology provides advanced on
-
line diagnostics,
creating awareness we never had before.


We are dealing with new types of information, not
previously envisioned. New processes and procedures
may be required.


Utilities are using DFA to avoid outages and better
diagnose other problems, but need better access to
reports to enable them to take full advantage.


New EPRI project is addressing issues regarding when
and how to deliver information, for best use and impact.


Field installations continue to provide data for discovery
of new “fingerprints” and better diagnosis and reporting.

Copyright © 2011. The Texas A&M University System.

36

Contact Information

John S. Bowers, PE



Carl L. Benner, PE

Vice President of Operations


Senior Research Engineer

Pickwick Electric Cooperative


Dept. of Electrical and Computer Engineering

PO Box 49, 530 Mulberry Street

3128 TAMU, Texas A&M University

Selmer, TN 38375



College Station, TX 77843
-
3128

731
-
646
-
3766



979
-
845
-
6224

jbowers@pickwick
-
electric.com

carl.benner@tamu.edu

DFA Technology Success Stories are available at:

https://epridfa.tamu.edu/DFAReports/DFASuccess.aspx