INTRODUCTIONPower cable and electrical equipment such as motors and

transformers must operate as long as possible within reliabilityand safety standards. Critical plant equipment must be moni-tored and maintained without sacrificing plant availability. Acritical component is one whose failure could endanger plantsafety, could cause an extended forced outage, or has a longlead time. An influence component is one whose failure wouldprobably not result in an extended outage, would not endangerplant safety, and is unit specific. In a power production plant,cable and station main transformers are critical components.

Electrical systems do not last forever. At some stage, PlantOperations personnel must decide when to replace cable com-ponents, electrical equipment, or an entire electrical system.This is a difficult and potentially costly decision. Wholesalerefurbishment of an electrical plant is too costly to be a practi-cal economic option. Common practices used to schedule elec-trical system maintenance are as follows.

• Operate until the electrical systemfails.

The least involved practice is to operatethe electrical system without any preventivemaintenance and repair when the cable fails.Replacement is scheduled when repair costsbecome more expensive than replacement.

• Replace based on specified failure rateor life span

Replacing equipment when the failurerate reaches a pre-established level orreplacing the cable system after a certain life span is still wide-ly used. This assumes that cable and equipment age uniformly,which is not the case. Different segments of the cable systemoften age non-uniformly along its length. If Plant Operationshas the means to identify only the components that needreplacement, significant savings can be achieved.

• Conduct laboratory diagnosticsAnother way of assessing cable condition is to perform a

laboratory evaluation on cable samples removed from field.The diagnostic tests conducted in the laboratory may revealvaluable information about the condition of the removed sam-ple but the results have to be extrapolated to the rest of the sys-tem. High voltage DC (Hi-pot) and VLF tests are sometimesused to identify the cable components that are on the verge offailure. However, these are destructive tests.

• Perform Diagnostics testing.Of all the diagnostic methods available today; condition

assessment diagnostic testing provides the most detailed infor-mation about the performance of an electrical system, down toindividual components. Condition assessment provides earlyidentification of weak components of the cable system. Itlocates degraded components and determines the extent of

degradation. This is essential to maintain system reliability.Cost savings can be realized by prioritizing the replacement ofweak sections of a circuit. Condition assessment is often moreexpensive than other diagnostic methods. However, the value ofthe detailed test results and the savings achieved far exceed thecost of condition assessment.

DTE ENERGY TECHNOLOGIES ONLINE CONDITIONASSESSMENT

On-line, in-situ testing to estimate future performance ofoperating cable systems and electrical equipment represents anadvance in diagnostics technology for the cable industry. Thisadvance is possible due to novel technology developed andpatented by DTE Energy Technologies (DTECH), includingadvances in signal processing and interpretation. Of severaldiagnostic methods available today, the Cable/Wise conditionassessment diagnostic testing is most effective, as it provides

early identification of weak compo-nents of the cable system while the sys-tem remains energized. It can locatedegraded components of the systemand determine the extent of degrada-tion. This is essential to maintain sys-tem reliability. On average an 80% costreduction in cable and equipmentreplacement costs can be realized byDTECH condition assessment.

The overall objective of diagnostictesting, of course, is to identify defects

that could cause a system failure, and estimate the time remain-ing before these defects progress to failure and cause an outageof the electrical system. The test should be economically justi-fied and should not cause additional degradation to the systemunder test. Hence, testing performed at over-voltages is alwaysof some concern.

Diagnostic test methods that detect partial discharges (PD),which are active during the time of testing, can only detectdefects or imperfections that produce partial discharge greaterthan the sensitivity of the test method. Cable accessories suchas splices and terminations are most likely to fail because of PDthat causes degradation. For cables, not all degradation phe-nomena are associated with PD.

Power cables are used to supply power to plant equipment,such as large motors, auxiliary transformers, precipitators, andback-up diesel generators. These cables are rated from 2001Vto 15 kV and are single conductor or three-conductor cables,shielded or unshielded construction. Before 1970 power cableswere insulated with extruded dielectrics, such as XLPE, andbutyl rubber. In some cases PILC cable was also used.

Electricity Today28

GENERATION

CONDITION ASSESSMENT OF ELECTRICALEQUIPMENT IN POWER PLANTS

By Nagu Srinivas and Dr. Oscar Morel, DTE Energy Technologies

NomenclatureEPR Ethylene Propylene RubberPD Partial DischargePILC Paper Insulated, Lead CoveredXLPE Cross-Linked PolyethylenePD Partial Discharge

Continued on Page 30

March ET 2/23/09 3:15 PM Page 28

The majority of cable failuresin an extruded cable system arerelated to water treeing, which failthe cable when they progress toelectrical trees. Once a water treeprogresses to an electrical tree, thetime to failure normally is veryshort because the initiated electri-cal tree propagates rapidly throughthe already weakened dielectric.Thus, the only window for detec-tion is during the conversionprocess. Under normal operatingconditions, such conversion iscaused by prolonged activity incavities created in the water treechannel as the result of heat gener-ation caused by ionic current.

In PILC cables, failures arecommonly associated with mois-ture ingress, which normally fails the cable through thermalrunaway. Moisture in PILC cables increases the dielectric loss-es resulting in localized heat generation that thermallydegrades the paper insulation and normally leads rapidly to acable failure. PD may only be present at advance stages of suchdegradation.

An integral method that provides detection, location andcondition assessment of both PD and water content is neededto ensure reliable cable system operation. Any PD testing tech-nique that requires a cable system or equipment to be isolatedwill require shut down coordination to switch off and isolatethe equipment under test, and reroute power to other equipmentnot under test. This may not always be convenient or econom-ical. Also, critical circuits remain in service while being tested.In contrast, the Cable/Wise (DTECH) technique is capable ofdetecting and locating PD and moisture content while the entiresystem remains energized.

Note that measuring only the magnitude component of PDdoes not provide enough information to reliably assess thecable condition. The severity of the PD condition depends onthe:

• Insulation material in which PD occurs;• Environment in which the cable system is operating;• Type of defect producing the PD;• Location of the PD within the insulation wall.Hence, any specific reported PD value diminishes in sig-

nificance as that activity is further removed from the conductorshield.

DTECH TEST METHODOLOGY - CABLEAs noted above, when cables and accessories age, the

resulting changes do not take place uniformly along the systemlength. Hence, for any diagnostic tool to provide truly mean-ingful information, one must be able to assess the cable systemby length. Non-uniform aging may be due to one or more ofmany factors; manufacturing issues, localized contaminationleading to weak boundary layers at an insulation-contaminantinterface, water migration to high stress sites, loose shields at

discrete locations, microcracks produced by mechanicalfatigue, and so forth. Of great significance is the exact locationof the defect within the cable insulation wall.

Signal and partial discharge detection in the field is signif-icantly different from partial discharge testing of extrudedcables shortly after manufacture. The objective of the latter isto detect manufacturing defects (voids, shield-interface imper-fections) as a result of the extrusion process. This testing isintentionally performed at an over-voltage. Testing in the fieldrequires suitable sensors, a noise filtration system and signaldetection and processing capability. The DTECH approach(Cable/Wise) provides this at operating voltage, hence elimi-nating the need for a system shutdown and deleterious effectscaused during off-line testing.

Due to defects, all cable components emit signals duringoperation, but the nature of those signals changes dependingupon the cause (i.e., the defect type, as noted above); for exam-ple, loose shields yield different signals than do internaldefects. The Cable/Wise totally passive technique utilizes the(RF) emissions to provide an assessment of the remaining lifeof the cable sections, splices, terminations, and other electricalequipment connected to the circuit.

One challenge is to be able to measure such small signalsand transmit the information, and another is to be able to inter-pret them (relative to type and location). DTE EnergyTechnologies has been able to perform this and relate the infor-mation from such signal detection to future reliability in oper-ation. The signals measured are generally due to small conven-tional partial discharges; however, DTECH has been able tomeasure other age related signals also.

The signal developed during aging induces a current flowin the cable shield and conductor; this consists primarily ofhigh frequency components of the signal. The magnetic fieldresulting from the current flow is measured. The DTECHmethod couples energy from the magnetic field of the signalpulse into the measurement system (Figure 1). The sensors,developed and constructed by DTECH, are designed to pick upthe magnetic field and are capable of detecting signals in the

Electricity Today30

Condition AssessmentContinued from Page 28

Figure 1 Data acquisition setup

March ET 2/23/09 3:15 PM Page 30

Level 5: The system has a highprobability of failure within thenext two years. Consider immedi-ate replacement.

Phase resolve studies ofsignals from several hundred thou-sand feet of cables (PILC, XLPE,EPR) over the past six years haveenabled DTECH to ascertain pat-terns from the data developed,which, in turn, facilitates estima-tion of future performance.

Test Methodology - ElectricalEquipment

The condition of electri-cal equipment such as transformer,switchgear, and motors connectedto power cable can also beassessed with Cable/Wise technol-

ogy. Again, no shutdown of equipment is required while tak-ing readings (Figure 2).

Measurements are taken while the equipment is operatingat operating voltage under normal conditions.

Throughout the life of a transformer several operating andphysical conditions degrade the transformer cellulose and oilinsulation system.

• High temperature and high load currents (above nameplate rating)

high frequency range. The readings are taken atintervals of several hundred feet along the cable.This is preferred since, as noted above, cablesage unevenly and knowledge of the aging con-dition of the system over discrete sections isdesired. It is to be emphasized that this testing isperformed while the system remains energized.Noise reduction is accomplished through signalprocessing in the frequency domain. The non-destructive test procedure is not limited by cablelength, operating voltage, insulation type, cableconstruction, or branching of the system. A sig-nificant feature of Cable/Wise technology is thatit can distinguish between cable and accessorydegradation activity.

The development of an extremely fast,affordable digital signal oscilloscope and wave-form digitizer, and recent advances in signalprocessing and computer technologies have ledto better understanding of the role and signifi-cance of individual partial discharges and relat-ed signals in degrading insulation systems.

Frequency domain testingKey to Cable/Wise technology is Frequency domain test-

ing. Frequency domain testing has several major advantagesover time domain testing, including:

1. PD can be detected, characterized, and located withouthaving to trigger on the first pulse.

2. Since the frequency domain testing is usually carriedout in service, and the PD is detected at various points alongthe cable, the cable between the point of detection and thecable termination acts as a high frequency filter which removesmuch of the noise which interferes with sensitive PD detection.

3. Since the PD detector is closer to the PD source andmuch of the interfering noise is filtered out by the cable, thebandwidth of the PD signal at the sensor can be used to judgelocation, and very sensitive PD detection is possible (Figure 5).

4. If the analyzer is triggered synchronously with thepower frequency, the analyzer display becomes a phase/fre-quency fingerprint of the PD signal.

5. Because PD detection is undertaken in service, one canassume that any PD source which could be active is likely tobe active. As explained above, if the cable is taken out of ser-vice to do off-line PD testing, the voltage must be raised to 2pu in order to assure that all PD sources which could be activewill be active.

Thus PD testing at 2 pu off-line is roughly equivalent totesting at normal operating voltage in service.

Since the signal magnitude cannot be used by itself toassess the significance of the detected signal, condition levelshave been assigned to the signal information. A description ofeach level is provided below.

Level 1: The system is not degraded. No action needs tobe taken.

Level 2: There is a small amount of aging related signalsin joints and terminations. This amount of signal is normal andthus, no action needs to be taken. However, in extruded cables,retesting is recommended within the next two years.

Level 3: The system has a low probability of failure with-in the next two years. Consider retesting at a one-year interval.

Level 4: The system has a medium probability of failurewithin the next two years. Consider replacement.

31March 2009

Figure 2 Data acquisition for a station orunit transformer

March ET 2/23/09 3:15 PM Page 31

• Moistureinside the trans-former

• Oxygen –oxidation of cellu-lose and oil

• Voltagesurges, lightning

• Physicaldamage due tomoving/re-locating

As celluloseand oil ages, theirinsulating proper-ties degrade, andproduce gases thatcan form voids inthe insulating oil.These voids, inturn, are sites forpartial discharge.

C a b l e / Wi s edetects partial dis-charge in thesesites. Aging canprogress to a levelthat will eventuallyfail the trans-former, most com-monly in the coilstructure.

In addition toinsulation aging,the laminated coreof a transformercan degrade.

D i s c h a r g emay take placebetween lamina-tions at locations where thelamination insulation (usual-ly a metal oxide layer) hasbroken down. Core dis-charge can take place over along time period, but usuallydoes not lead to catastrophicfailure of the transformer.Transformers typically failin the core structure.Cable/Wise can discriminatebetween core and trans-former insulation discharge.

Transformers, motors,and switchgear operate withsome level of internal dis-charge, even when new. Asingle test can indicate thepresence or absence of dis-charges as well as their rela-tive intensity. The dischargeintensities and patterns mea-sured in the initial test areunique to that motor only.

Electricity Today32

Figure 3 Surface dischargedetected in a pad mountedswitchgear

Figure 4 Analysis of 15-kV terminations

March ET 2/23/09 3:15 PM Page 32

More than one test within a cer-tain period of time is necessaryfor an accurate diagnosis, withthe period of time between testsdepending on the intensity ofthe discharge. Trend analysis ofsuch a series of tests is amethod that provides an accu-rate assessment of the condi-tion of electrical equipment.This also applies to transform-ers.

CASE STUDIESPad mounted switchgear

Surface tracking in a padmounted switchgear was locat-ed by the RF sensor located200 feet away in an manhole(Figure 3). The switchgear wasnot part of the original condi-tion assessment contract.However, after notifying theclient about what was detected,the switchgear was investigatedfurther. The switchgear wasopened and a parabolic coronadetector was used to confirmthat the RF signals appeared to be emit-ted by surface contact between cable ter-minations and barriers inside theswitchgear. The discharge also registeredas heat in a thermal image of the termi-nations.

15-kV Distribution FeedersRF measurements were taken at the

base of six terminations of a double cabledistribution feeder circuit at the substa-tion end of the cables. The terminationshad been in service for 20 years. Analysisof the signals indicated that all six of ter-minations had an advanced degree ofinternal tracking (Figure 4). It was rec-ommended that the client replace all sixterminations as soon as possible. Theclient replaced the terminations and theexisting cable. The client shipped the ter-minations, each with a 25 ft length ofcable attached, to DETCH.

Laboratory RF measurements con-firmed the results of the field measure-ments. The terminations were then dis-sected (Figure 5). Heavy tracking wasobserved in the bore of each elastomerstress cone in all six terminations.Discharge activity eroded some the elas-tomer at the embedded wire stress conearea. Some of the metal componentsinside the terminations were heavily cor-roded, an indication of moisture entryover a long period of time. Siliconegrease was found at the termination of

the cable semiconducting insulationshield in all six terminations. The grease

in some of the terminations had begun towax, an indication that discharge activity

33March 2009

Figure 5 Dissection results of 15-kV terminations

March ET 2/23/09 3:15 PM Page 33

THE WUNPEECE DUCT SPACERTHE WUNPEECE DUCT SPACERONE PIECE DOES IT ALL!ONE PIECE DOES IT ALL!

was in progress at the end of the shield. In one termination adouble impression in the elastomer stress cone was evidencethat the cable had moved 1/4” inside the termination.

It was concluded that although these terminations did notfail in service, they certainly were near end of life.

MotorsA motor operates with some level of internal discharge,

even when new. A single test can indicate the presence orabsence of discharges as well as their relative intensity. Thedischarge intensities and patterns measured in the initial testare unique to that motor only.

More than one test within a certain period of time is nec-essary for an accurate diagnosis, with the period of timebetween tests depending on the intensity of the discharge.Trend analysis of such a series of tests is a method that pro-vides an accurate assessment of the condition of a motor.

In Figure 6 two motors in a refinery facility had high lev-els of RF discharge as measured in 2002. Maintenance wasperformed on the motors based on the 2002 measurements.The same motors had significantly less discharge when retest-ed in 2003 as shown Figure 6.

On-line Discontinuity Locator (ODL)ODL is used to detect direct buried splices and neutral

corrosion in cables. The technology used for the tests isDTECH (proprietary) On-line Discontinuity Locator (ODL). Asmall pulse signal is injected into a cable by a sensor. The sig-nal reflects back to the sensor at changes in the cable imped-ance. The reflected signals are picked up with the same sensor.The pulse reflection shape indicates impedance changes alonga cable. Neutral corrosion, cable faults, both open and shortcircuit (short circuit, open circuit), and insulation damage, canbe detected and located by analyzing the reflected signals.

Figure 7 shows an example of severely corroded coppershield tapes located by ODL in a nuclear power plant.

25 kV Rated, Heat Shrink JointsCable/Wise testing was performed on a 24 kV cable sys-

tem in two utilities. The less than 20 year old, XLPE insulat-ed, direct buried systems exhibited a significant number ofsites with elevated levels of degradation. The source of degra-

Electricity Today34

Figure 6 Trend analysis in motors

Figure 7 Neutral corrosion

March ET 2/23/09 3:15 PM Page 34

CONCLUSIONC a b l e / W i s e

online condition assess-ment has a proven trackrecord of being a costeffective, timely, andversatile diagnostic toolthat assesses the condi-tion of each componentof an energized electri-cal system, including allsections of cable, cableaccessories, transform-ers, motors, andswitchgear. It canincrease electrical sys-tem reliability, andavoid unplanned out-ages and power plantelectrical problems.

ACKNOWLEDGEMENTThe authors

thank Dr. Nezar Ahmed, Principal Technical Consultant, forproviding much of the information presented in this paper, andto David Bogden, Engineering Manager, for compiling datafor this paper.

dation was identifiedto be the cablesplices. The customerelected to remove 12joints with a 10-meter cable sectionattached and sentthem to the laborato-ry for visual inspec-tion and tests. Thesplices exhibiteddegradation levels 2through 5. Thesplices were identi-fied as A through L.

The laboratoryevaluation includedRF detection, ACbreakdown, thermo-graphic and visualexamination. Theresults are given intable 1. The laborato-ry evaluation con-firmed the field data. In addition a correlation was observedbetween degradation levels and ac breakdown.

15 kV rated Butyl Rubber CablePartial Discharge measurements were performed in the

field on 15 kV rated butyl rubber insulated, armored/PVCjacketed cable installed at a production facility. Visual obser-vation of the cable revealed various degrees of jacket crackingfrom light to severe cracking (In some places there was noouter jacket at all). Subsequently, the 250 meter section ofcable, which showed discharges of high level distributed uni-formly throughout the length of the cable, was removed fromservice. This section was then cut into ten-meter long samplesand brought to DTECH for further examination. The sampleswere subjected to various electrical and chemical tests in thelaboratory. The samples exhibited a breakdown strength in therange of 8.0 to 16.00 volt/mm. Chemical testing also indicateda uniform thermal degradation both radially and longitudinal-ly. This indicates that the cable had degraded along its entirelength, independent of the condition of the jacket.

15 kV rated Submarine CablesCondition assessment was performed on two circuits of

submarine cables (rated at 24 kV and operated at 13.2 kV).Each circuit consisted of three single conductor, 4/0, EPR insu-lated, armor cables. The circuits were in service 12 years. Asmall portion of the circuits was installed on land, while themajority was in 150-meter deep water. A system overload con-dition lead to thermal degradation of the cables. Both circuitssuffered several failures within one week. As a result, the cir-cuits were tested by the DTECH system. Data from testingrevealed that the land portion of the cables at both shores haddegraded completely.

The data also showed that the water portion of the cableswas still usable due to the cooling effect of the water. The landportions of the cables were replaced, and the circuit returned toservice. This action saved the remainder of the tourist seasonfor the island. The cables were completely replaced the fol-lowing year prior to the tourist season.

35March 2009

March ET 2/23/09 3:15 PM Page 35