4-Fault Detection in Underground Power Distribution Cables

Fault Detection in Underground Power Distribution Cables

Igor Paprotny, Richard M. White, Paul K. Wright

University of California, Berkeley

Thursday, September 20, 2012

Introduction

57,500 miles of primary distribution cable in PG&E networkcable in PG&E network.

In urban areas 70 % of the distribution network is underground

12 KV to 35 KV voltage Cables age at different rates Cable failures Cable failures

Insulation breakdown (water trees) loss of protective grounding shield (CNs)

Need to know which cables to replace first

Need for a reliable on-line CN diagnostic technique

I. Paprotny et al. i4 Energy Symposium2

Our Vision

I. Paprotny et al. i4 Energy Symposium33

Magnetic (AMR) CN Probing

I. Paprotny et al. i4 Energy Symposium44

Magnetic (AMR) CN Probing

Measure currents in concentric neutrals (CNs) by i ti ti fi ldsensing emanating magnetic field

Anisotropic Magnetic Resistance (AMR)M t h i b l t (10% 20%) Measure return or phase imbalance currents (10%-20%)

Detect imbalances or lack of CN current as a sign of degradationdegradation

I. Paprotny et al. i4 Energy Symposium5

AMR Magnetic Field Sensors

Anisotropic magnetoresistance(AMR)

Resistance depends on the Resistance depends on the orientation of the magnetic field (B-field)

Use off-the-shelf Honeywell HMC1043 sensor

Cost $9.00 in volumes of 1,000

3-axis measurement Circumferential Circumferential Radial Axial

I. Paprotny et al. i4 Energy Symposium6

Hot-stick Deployable System

I. Paprotny et al. i4 Energy Symposium7

Fabricated Fixture

I. Paprotny et al. i4 Energy Symposium8

Jacketed ~50 A, 100 % CN return

1.2

1.4

1.6

1 CN broken ‐ 100% return ‐ 51.4 Arms CC (near)

Circumferential

2.5

34 CN broken ‐ 100% return ‐ 51.4 Arms CC (near)

Circumferential

Axial B‐field

0.6

0.8

1

1.2

Magne

tic F

ield [G

auss] Axial B‐field

Radial B‐field

1

1.5

2

Magne

tic Field [G

auss] Radial B‐field

0

0.2

0.4

0 100 200 300Angle [Deg.]

0

0.5

0 100 200 300Angle [Deg.]

Angle [Deg.]

1.2

1 CN broken ‐ 100% return ‐ 51.4 Arms CC (far)

Circumferential 2 5

3

4 CN broken ‐ 100% return ‐ 51.4 Arms CC (far)

Circumferential

0.6

0.8

1

etic Field [G

auss]

Axial B‐field

Radial B‐field

1.5

2

2.5

etic Field [G

auss]

Axial B‐field

Radial B‐field

0

0.2

0.4Magne

0

0.5

1Magn

I. Paprotny et al. i4 Energy Symposium9

00 100 200 300

Angle [Deg.]

0 100 200 300

Angle [Deg.]

Jacketed ~20 A, 50 % CN return

0.25

0.3

2.05

2.1

Circ.

1 CN broken ‐ 50% return ‐ 21.7 Arms CC (near)

Circumferential

Axial B‐field

Radial B‐field 1.4

1.6

1.8

irc.

4 CN broken ‐ 50% return ‐ 20.55 Arms CC (near)

0.1

0.15

0.2

1.9

1.95

2

Magne

tic Field [G

auss]

Magne

tic Field [G

auss] ‐C

0.6

0.8

1

1.2

Magne

tic Field [D

eg.] ‐C

i

Circumferential

Axial B‐field

Radial B‐field

0

0.05

1.8

1.85

0 100 200 300

M

Angle [Deg.]

0

0.2

0.4

0 100 200 300

M

Angle [Deg.]Angle [Deg.]

0.3

0.35

2.1

2.151 CN broken ‐ 50% return ‐ 21.36 Arms CC (far)

Circumferential

Axial B‐field

1.6

1.8

4 CN broken ‐ 50% return ‐ 20.50 Arms CC (far)

0.15

0.2

0.25

1.9

1.95

2

2.05

tic Field [G

auss]

etic Field [G

auss] ‐C

irc.

Radial B‐field

0.8

1

1.2

1.4

gnetic Field [G

auss] Circumferential

Axial B‐field

Radial B‐field

0

0.05

0.1

1.75

1.8

1.85

0 100 200 300

Magne

t

Magne

0

0.2

0.4

0.6

0 100 200 300

Mag

I. Paprotny et al. i4 Energy Symposium10

0 00 00 300

Angle [Deg.] Angle [Deg.]

Unjacketed (near)

2.5

3

Unjacketed 1 CN broken ‐ 50 A 100% return (near) 

2 5

3

Unjacketed 4 CN broken ‐ 50 A 100% return (near) 

Circumferential

~50 A 100% CN

1

1.5

2

Magne

tic field [Gauss]

Circumferential

Axial B‐field

Radial B‐field

1

1.5

2

2.5

Magne

tic field [Gauss]

CircumferentialAxial B‐fieldRadial B‐field

0

0.5

0 100 200 300

Angle [Deg.]

0

0.5

0 100 200 300

Angle [Deg.]

0 83

Unjacketed 1 CN broken ‐ 20 A 50% return (near) 

23

Unjacketed 1 CN broken ‐ 20 A 50% return (near) 

~20 A 50% CN

0.4

0.5

0.6

0.7

0.8

2.9

2.92

2.94

2.96

2.98

3

 [Gauss]

c fie

ld [G

auss] ‐C

irc.

Circumferential

Axial B‐field

Radial B‐field

0 8

1

1.2

1.4

1.6

1.8

2.9

2.92

2.94

2.96

2.98

field [G

auss]

ic field [Gauss] ‐C

irc.

CircumferentialAxial B‐fieldRadial B‐field

0

0.1

0.2

0.3

2.8

2.82

2.84

2.86

2.88

0 100 200 300

Magne

tic field

Magne

tic

0

0.2

0.4

0.6

0.8

2.82

2.84

2.86

2.88

0 100 200 300

Magne

tic f

Magne

ti

Angle [Deg ]

I. Paprotny et al. i4 Energy Symposium11

Angle [Deg.] Angle [Deg.]

Next Steps

Fabricate new magnetic field sensors Increased sensitivity to CN field Reduced susceptibility to CC field

Long-term: miniaturize and embed Miniaturize, include energy scavenging and radio for

embedding into cables or elbowsembedding into cables or elbows

Commercialization / technology transfer

I. Paprotny et al. i4 Energy Symposium1212

RF Test-point Injection Probing

I. Paprotny et al. i4 Energy Symposium1313

The Mechanism

Simplified model of a water treeCable Voltage

At V=0 voids filled with water are not connected

Ref: Hvidsten et al “Understanding Water Treeing Mechanisms in the Development of Diagnostic Test Methods

At V=0, voids filled with water are not connected

I. Paprotny et al. i4 Energy Symposium14

Ref: Hvidsten et. al. Understanding Water Treeing Mechanisms in the Development of Diagnostic Test Methods

The Mechanism

Simplified model of a water treeCable Voltage

At higher voltage micro channels open due to

Ref: Hvidsten et al “Understanding Water Treeing Mechanisms in the Development of Diagnostic Test Methods

At higher voltage, micro-channels open due to Maxwell Mechanical Stresses and water will penetrate the channels making electric contact

I. Paprotny et al. i4 Energy Symposium15

Ref: Hvidsten et. al. Understanding Water Treeing Mechanisms in the Development of Diagnostic Test Methods

RF Test-point Injection Probing

Couple an RF signal through elbow test-points Experimentally confirmed coupling grater than 1 MHz

signal

M tt ti l it f ti Measure attenuation, velocity of propagation as a function of instantaneous cable voltage.

I. Paprotny et al. i4 Energy Symposium16

Preliminary Results

See only minor changes in the transmitted waveform, as a function of voltage.

Results warrant further investigation

Likely no water-trees in the tested cable !

I. Paprotny et al. i4 Energy Symposium1717

Next Steps

Validate the method on faulty cables Validate existence of water trees on faulty cable

segments Test the technique on faulty cables under ~10 kV.q y

Commercialization / technology transfer

I. Paprotny et al. i4 Energy Symposium1818

Acknowledgements

Thi k t d b t f th C lif i EThis work was supported by grants from the California Energy Commission (CEC), contract numbers 500-01-43, 500-02-004 and POB219-B.

Equipment was donated or borrowed by PG&E, SDG&E, SCE, and A il tAgilent.

Special thanks the SECURE Cables Technical Advisory CommitteeSpecial thanks the SECURE Cables Technical Advisory Committee and the California Energy Commission

I. Paprotny et al. i4 Energy Symposium1919