A COST EFFECTIVE, NUMERIC TECHNIQUE FOR PROJECTING QUALITY OF INSULATION AND IMPENDING FAILURES Marcus O. Durham THEWAY Corp Robert A. Durham RADCo Consulting.

A COST EFFECTIVE, NUMERIC TECHNIQUE

FOR PROJECTING QUALITY OF INSULATION

AND IMPENDING FAILURES

A COST EFFECTIVE, NUMERIC TECHNIQUE

FOR PROJECTING QUALITY OF INSULATION

AND IMPENDING FAILURES

Marcus O. Durham THEWAY CorpMarcus O. Durham THEWAY Corp

Robert A. Durham RADCo ConsultingRobert A. Durham RADCo Consulting

Marcus O. Durham THEWAY CorpMarcus O. Durham THEWAY Corp

Robert A. Durham RADCo ConsultingRobert A. Durham RADCo Consulting

Abstract Hi Pot

– Cable & Machines– Numeric technique – Predicts impending failure

DC Limitations

No correlation between AC strength & DC testExperience is used

Aged insulation => AC Use => DC test Life 5 times longer if not tested

Flashover transient weakens insulation Reflected wave = 2 times peak

DC

DC Limitations

Higher test V for DC than AC = Space Charge

Etotal = Espace + Einsulation

Insulation = 5000 V, 90 mil EAC = 55 volts / mil

EDC = 10 X Result = Deterioration of insulation

To Tree or Not To Tree

Problems are recognized in polyethylene

May be in rubber as well

5000 Volt insulation ???

TestersDespite Limitations - DC Still Preferred

VOM (volt-ohmmeter) IR (insulation resistance) Hi Pot (high potential DC)

– More info about quality than other– 60,000 V field– 200,000 V labVAC

A DC

A AC OFF

Ohms V DC

120.01

Elevated Voltage

Can cause any insulation to fail

Difficult to interpret w/o destruction

Most valuable w/ historical data

Experience, skill, knowledge of local conditions - major aids to determining suitability

Resistance vs. Current Georg’s Law

Vtest = Rinsulation * Ileakage

Resistance decreases as length increases Non-linear, parallel R I exponential as length Increases Conductivity often called leakage current

A / ( k V * k ft ) => ( A / k V ) / k ft

Leakage Conductance

One number is futile Length, diameter, insulation, geometry, voltage Bulk resistivity

K log

D

d

- EPDM 20,000 M k ft

- Polyethylene 50,000 M k ft

Leakage ConductanceEPDM

Derate for temp, moisture, oil

Lower values may still be good

K = 20,000 M k ft

G = .05 A / k V * k ft0

Suitability for Reuse

Visual & 5 minute DC withstand Lack of Consensus 75 mil EPDM 11,000 - 25,000

Comparison MethodsOverpotential Test

Specified leakage current Leakage conductance Leakage w/ ratio of 3 to 1 No consistent guidelines

Leakage Current ComponentsCapacitance Charging

Between conductor & ground Starts extremely high, decrease exponentially Drops to zero in few seconds

1 2 3 4 5 60

0.1

0.2

0.3

0.4

1 2 3 4 5 6

Leakage Current ComponentsAbsorption

Dielectric insulation Result of Polarization Starts high, decreases slowly Stabilize in 5 minutes Reasonable in 2 minutes

1 2 3 4 5 60

0.05

0.1

0.15

0.2

0.25

1 2 3 4 5 6

Leakage Current Components Conductance

Steady state value Over, under, around, through insulation Corona contributes Low value is good

1 2 3 4 5 60

0.10.2

0.3

0.40.5

1 2 3 4 5 6

Leakage Current -vs- Time

i = F + (I - F) e -t/RC

F = constant, property of material

I = initial

1/RC = time constant

EvaluationMore Thorough

1 - Apply increasing voltage

2 - Plot leakage current -vs- applied voltage

3 - If R = ideal, V does not affect I

4 - Increasing I => insulation weakened

A TechniqueNot a Number

Forecast failure levels

Normalize– p Megohm - k ft– I Microamps– V Kilovolt– L thousand feet

Step 1Calculate Bulk Conductance

Contrast with measured

G = Ileak / Vtest

G1000

KiloVolts

mic

roA

mp

s

Step 2Plot A -vs- kV

I = G V

Becomes upper boundary for Itest

Step 3 Hi-pot Test

Perform hi pot test in steps

Step 4Calculate Slope

Calculate derivative at each test point

m

d i

d v

i i

v v

( )

( )2 1

2 1

Step 5Compare

Compare slope tobulk conductance

M > G => impending failure

KiloVolts

mic

roA

mp

s

Step 6Plot Forecast of Failure Point

i F e av ( )1

a

i i v v

v v

ln ln ln ln2 1 1 2

2 1

Fiav

1 e

Step 7For Next Test Point

Recalculate

steps 3 - 6

to refine forecast current

Step 8Calculate Forecast Voltage

Bulk

Test

Intersection

i G v

i F eav ( )1

va

G

aF

1ln

Step 9Calculate Comparative Quality

cq = forecast V / rated test V

Changes between tests show deterioration

Ratio < 40% indicates marginal quality

Figure 2 - Results

kiloVolts

mic

roAm

ps

Bulk Conductance

Test Data

Summary Testing

Insulation for reuse - controversial DC Hi Pot common - despite problems

Measure leakage I at test V Math technique to compare quality

Calculate bulk conductance (G) - limit Calculate rate of change of leakage G Compare derivatives (slope) Calculate coefficients of leakage curve Determine forecast V at intersection

Summary Mathematical Technique

Questions ?

Questions ?