Smart Grid Monitoring by Duval

8/12/2019 Smart Grid Monitoring by Duval

1/65

1

Smart Grid Monitoring of Transformersby DGA

Michel Duval

CIGRE Thailand, Bangkok 2013


2/65


3/65

3

Power Transformers


4/65

4

Catastrophic Failures


5/65

55

Failures in Service

- The failure rate of power transformers in service (internalfailures needing repairs) typically is 0.3% per year.

- For a population of 2000 transformers, this means 6transformers will fail in the next year.

- Less than 1 will fail catastrophically.

- 1994 will not fail.

- 200 (i.e., 10% of the population at or above IEEE/IECcondition 1) will form abnormal amounts of gases becauseof faults.


6/65

66

The Monitoring Dilemma

- Nobody knows which 6 of the 2000 transformers will failnext year.

- To identify them, all the transformers need to be monitored,including the 1800 operating normally, just for the purpose ofdetecting the 6 that will fail and need repairs and the lessthan 1 that may eventually fail catastrophically.

- In economic terms, the cost of monitoring is justified as long

as it does not exceed the cost of not detecting the 6 failuresand the catastrophic one (typically, >20M$).


7/65

77

Smart Grid Monitoring

- A smart grid is a modernized electrical grid that usesinformation and communications technology in anautomated fashion to improve the efficiency, reliability andsustainability of the production and distribution of electricity.

- It implies a re-engineering of the electricity services industry.

- It requires monitoring tools for evaluating on a real-timebasis the condition of electrical equipment, so as to optimizeasset utilization, system reliability and load capabilities


8/65


9/65

99

Basic Types of FaultsDetectable by DGA

-PD: partial discharges of the corona-type in voids in paperinsulation, as a result of poor drying or impregnating with oil.

-D1: low-energy discharges, such as partial discharges ofthe sparking-type in oil or paper, tracking on paper, smallarcing, arc-breaking activity in LTCs.

-D2: high-energy discharges, e.g., flashovers, high-energy

arcing, short-circuits with power follow through, withBuccholz alarms and tripping.


10/65

1010

Basic Types of FaultsDetectable by DGA

-T1: thermal faults of low temperature T < 300C, becauseof overloading, insufficient cooling, design problems.

-T2: thermal faults of 300 700C, becauseof high circulating currents in core and coil, short circuits in

laminations, often in oil only.


11/65

1111

Additional Sub-Types of FaultsDetectable by DGA

-S: stray gassing of oil (in oil only) at T


12/65

1212

Additional Sub-Types of FaultsDetectable by DGA

-T3/T2 in oil only: at T >700/ 300C, of minor concern aslong as they do not evolve into faults D1, D2 or C.

-C: carbonization of paper at T >300C, potentially moredangerous (loss of insulating properties of paper).


13/65

1313

DGA Diagnosis Methods

-Key Gas, Rogers and IEC methods. Limitations are high %of wrong diagnosis (50%) or undiagnosed cases (30%),respectively.

-Duval Triangle 1, allowing to detect the 6 basic types offaults (PD, D1, D2, T3, T2, T1 + DT).

-Duval Triangles 4 and 5, allowing to detect the 5 sub-typesof faults (S, O, R, T3/T2 in oil, C), and to distinguish

between faults of lesser concern in oil and more seriousfaults in paper.


14/65

1414

Duval Triangles 1, 4 and 5

Triangle 1

Triangle 4 Triangle 5


15/65

151515

Use of Triangles 4 and 5

-Triangles 4 and 5 should never be used in case of faultsidentified with Triangle 1 as faults D1 or D2.

-Triangle 4 should be used only for faults identified firstwith Triangle 1 as low temperature faults PD, T1 or T2, or

when there is a high level of H 2.

-Triangle 5 should be used only for faults identified firstas high temperature thermal faults T2 or T3.

15


16/65

16

Mixtures of Faults

-mixtures of faults sometimes occur rather than pure faults and may be more difficult to identify with certainty.

-for instance, mixtures of faults D1 and T3 may appear asfaults D2 in terms of gas formation.


17/65

17

New Faults vs. Old Faults:

-when a new fault appears, as evidenced by a change ingas pattern, a more precise identification of the new faultmay be obtained by subtracting the gas concentrationscorresponding to the old fault from those corresponding tothe new one (incremented values).

-this, however, introduces additional uncertainty on thesubtracted value.


18/65

1818

Interpretation of CO and CO 2

-Until recently, CO and CO 2 were considered as goodindicators of paper involvement in faults. Recentinvestigations at CIGRE, however, have shown that this isnot always the case.

-High concentrations of CO (>1000 ppm) and/or lowCO 2/CO ratios (


19/65

191919

CO2

and CO from Closed Transformers

Ref: I.Hoehlein, CIGRE TF15 (2010)

56 MVA, 220kV

Manufactured 2006

Rubber Bag

19


20/65

2020


-High concentrations of CO (>1000 ppm) and low CO2/CO

ratios (10,000 ppm), high CO 2/COratios (>20) and high values of furans (>5 ppm) are anindication of the slow degradation of paper at relatively low

temperatures (


21/65

2121


-Low concentrations of CO and CO 2, below condition 1 ofIEC or IEEE (750 and 7500 ppm, respectively), correspondto normal gassing in transformers without faults.

- Intermediate concentrations of CO, CO2

and CO2/CO ratios

may indicate a slow degradation of paper and intermediateDPs of paper, of no concern at all for the normal operationof the transformer.


22/65

2222


-Zero or very low rates of change of CO and CO 2 do notnecessarily mean the absence of a fault in paper. Localizedfaults in paper often do not produce detectable amounts ofCO and CO 2 against the usually high background of thesegases in service.

-However, they do produce significant amounts of the otherhydrocarbon gases, allowing the detection of faults in paperwith Triangles 4 or 5.


23/65

2323

Example of a Localized Fault in Paper


24/65

24242424

CIGRE Risk of Failure vs. CO 2

-The risk of failure is very low at high CO 2 values, which arestrongly correlated with paper degradation and low DPs ofpaper, suggesting that the risk of failure at low DPs of paper isalso very low, not very high as generally mentioned.

-Indeed, large numbers of transformers have been observed atCIGRE to operate quite normally with DPs of paper < 200.

-And no cases have been reported so far of transformers withDPs < 200 that failed because of the mechanical weakness ofpaper, even when subjected to external short-circuits.

24


25/65

25252525

Transformers at Risk of Failure

-So, in a large majority of cases, low DPs of paper do notmean the end-of-life of transformers as generallyassumed.

-The main concern with low DPs of paper is the shrinkage ofpaper and loosening of windings, not the mechanical (tensile)strength of paper. This can be mitigated by reclampingtransformers.

-Transformers most at risk of failure are gassing transformersthat cannot be fixed.

25


26/65


27/65

27

Typical / Condition 1 Values

-Typical /condition 1 values of IEC/ IEEE correspond to agiven percentile (90%) of the population of DGA results

-They mean that 90% of DGA results for dissolved gasesare below these 90% Typical values

-They are used to concentrate maintenance efforts on the10% of the population with the highest gas levels and

therefore most at risk

l d l


28/65

28

Typical / Condition 1 Values

-Below typical/ condition 1 values, gas formation isconsidered not to be a concern for the equipment.

-Below these values, it is recommended to use normalsampling frequency (monthly, semi-annual, etc.,..) and notto attempt a diagnosis.

- Above these values, it is recommended to use increasedsampling frequency (e.g., monthly or weekly) and a DGA

diagnosis may be attempted.


29/65

29

90% Typical (Condition 1) Valuesfor Concentrations at IEC (2007), in ppm

(vs. source)


30/65

30

90% Typical (Condition 1) Valuesfor Concentrations at IEEE (2013), in ppm

(vs. kV, MVA, age, %O 2)


31/65


32/65

32

Pre-failure (Condition 4) Values

-CIGRE has evaluated the probability of having a failure-related event (PFS) in service vs. gas concentration andgassing rate.

-Based on these PFS curves, pre-failure (condition 4)values have thus been established, as well as intermediateconditions 2 and 3.


33/65

33333333

Risk of Failure vs. Gases Formed

(PFS = Probabilityof Failure in Service)

33


34/65

343434

CIGRE/IEC Sampling Intervals vs.Concentrations in Service, in ppm

34


35/65

353535

CIGRE/ IEC Sampling intervals vs.Gassing Rates in Service, in ppm/month

35

Actions Recommended by IEC at


36/65

36

Actions Recommended by IEC atConditions 1-4

-Condition 1: increase oil sampling frequency for DGA

-Conditions 2-3: consider complementary tests(infrared scans, acoustic tests, PD tests, effect of load).

-Conditions 3-4: consider transformer inspection.

-Condition 4: consider transformer repair or replacement.

Transformer Parameters Influencing


37/65

37

Transformer Parameters InfluencingConditions 1-4

-CIGRE (2006)/ IEEE (2013):-Operating conditions (load, climate)-Age (new, old)-Type (power, core, shell, instrument, reactor).

-MVA, voltage-Open or closed

-CIGRE WG47 (2013):

-Fault type


38/65

Effect of Type of Thermal Fault


39/65

3939

on Condition 1 Values at CIGRE

(ppm)

(ppm)

Effect of Type of Electrical Fault


40/65

4040

ypon Condition 1 Values at CIGRE:

(ppm)

(after deletingC2H2 < 2 ppm)

(ppm)

Effect of Type of Fault


41/65

4141

ypon Condition 4 values at CIGRE:

(in ppm, using previous adjustment factors)

Comparison with Cases of High Gas Levelsi h F il CIGRE


42/65

4242

without Failure at CIGRE:

(in ppm)


43/65

4343

Monitoring with DGA

-Monitoring off-line (by manual or laboratory DGA) is mostlyused but cannot detect faults occurring between two oilsamplings (e.g., every year, month or week).

-On-line multi-gas or hydrogen monitors can detect abnormaland/or quick-developing faults occurring within days or hoursbetween oil samplings.


44/65

4444

Abnormal and Quick-Developing Faults

-Abnormal gassing (above condition 1 for concentrationsand gassing rates) will occur in 200 of the 2000transformers.

-Quick-developing faults (above condition 4 for gassingrates) will typically occur in 20 to 40 of them.

Detection of Quick-Developing Faults with a Multi-


45/65

4545

Q p gGas Monitor in a 3-Phase GSU Transformer

Day 2 16:00

Day 2 20:00

Day 3 00:00

Day 3 04:00

Day 2 12:00

Day 3 12:00

Day 3 16:00

Day 3 08:00

Day 23 04:00 toDay 24 08:00Followed bytransformer failure

C2H2 = 800 ppm/day!

700 MVA Transformer


46/65

4646

C2H2 = 45 ppm/day!

336 MVA Transformer


47/65

4747

336 MVA Transformer (Placed in Service -1969)

C2H4 = 300 ppm/day!

1100 MVA Transformer


48/65

4848

C2H4 = 300 ppm/day!


49/65

4949

-Gassing rates were all significantly above condition 4values in the previous 4 examples.

-The corresponding transformers were removed fromservice 1 to 3 days after looking at monitor readings, beforepotential catastrophic failure.

-However, it might have been better to remove them fromservice earlier.

-Without an on-line monitor, these transformers wouldpossibly have suffered catastrophic failures.


50/65

5050

On-Line Monitoring with Multi-Gas Monitors

- Multi-gas monitors will detect all types of faults, even intheir early stages at condition 1, and without false alarms.However, they are more expensive than hydrogen onlymonitors.

- The recommendation of CIGRE (TB # 409, 2010) istherefore to use multi-gas monitors in critical transformers(GSU, nuclear, transmission) and in abnormally gassingtransformers.


51/65

5151

Fault Detection with Hydrogen Monitors

Note: for faults T3 in paper (C), curve for H 2 is a bit higher.Ref: Duval, TSUG 2013.



52/65

5252


-Hydrogen monitors are most sensitive to stray gassing ofoil S (occurring in ~ 25% of cases), and to corona partialdischarges PD (occurring in only 0.3% of cases).

-Such faults will commonly produce thousands of ppm of H 2without being a concern for the transformers. If the limit inhydrogen monitors is set at average condition 1 values forH2 (100 ppm or 7 ppm/month), this may result into manyfalse alarms.


53/65

5353


-Faults D1/D2 at dangerous condition 4 of CIGRE willproduce 0.5 ppm/day of C 2H2 together with only 1 or 2ppm/day of H 2.

-If the limit for H 2 in the monitor is set at average condition 1(100 ppm), the monitor will detect these faults only in theirlate stages (condition 3 or 4), when dangerous levels of 25to 50 ppm of C 2H2 have already formed.

-If it is set at 5 ppm over a period of 3 days, this may resultinto many false alarms.



54/65

5454

- In case of thermal faults T3/T2/T1/O the main gas formed is

C2H4, CH 4 or C 2H6, together with 3 to 10 times less of H 2. Ifthe limit for H 2 is set at 100 ppm, the monitor will detectthese faults only in their late stages (condition 3 or 4).

- Decreasing the limit for H 2 in the monitor (e.g., to 50 or 20

ppm) will increase the number of false alarms due to faultsS or corona PD of lesser concern.

- The recommendation of CIGRE (in TB # 409, 2010) istherefore to use hydrogen monitors in non-critical

transmission and distribution transformers, and intransformers with no previous gassing history.

Examples of On-Line Gas Monitors


55/65

55

Examples of On Line Gas Monitors

Multi-Gas Monitors


56/65

56

Multi Gas Monitors

Monitors of the chromatographic type:

-after gas extraction, will separate individual gases on a GCcolumn, then measure them with GC detectors.

-TM8, TM3 (Serveron)

-Calisto 9 (Morgan Schaffer)


57/65

57

Monitors of the Chromatographic-Type:

-use the same standardized, NIST-traceable techniqueas laboratories.

-provide automatic recalibration at fixed intervals aslaboratories do.

-require some maintenance (change of carrier gas,calibration gas mixture, GC columns every 3 to 5 years).


58/65

58

Monitors of the Infrared-Type:

-after gas extraction, will measure directly individual gases withan infrared detector, and H 2 with a solid state sensor

-Transfix 8, Transport-X 7 (GE-Kelman) use a photo-acoustic(PAS) detector.

-LumaSense 9 uses a non-dispersive IR detector.

M it f th i f d t


59/65

59

Monitors of the infrared type:

-do not require change of carrier gas and gas mixture.-cannot measure H 2 , O 2 by infrared, requiring the use ofrelatively inaccurate solid state sensors for that purpose.

-some may need recalibration because of contaminationsin the ambient air (SF 6, oil vapours, solvents) and lampfade with time; some cannot be recalibrated in the field.

-require change of infrared lamp ~ every 5 years.-contain several moving parts.

Hydrogen Monitors


60/65

60

-Hydran (GE): measures 100% of the H 2 + 18% of the COpresent in oil with a PTFE membrane and fuel cell detector.

-Calisto 2 (Morgan Shaffer): measures H 2 only with a PTFEmembrane, GC and TCD detector.

-TM1, Qualitrol, Weidmann: measure H 2 with an inorganicmembrane and an H 2Scan Pd solid state sensor.

-TM1 (Serveron): improved version of H2Scan.

Other Applications of DGA


61/65

6161

Other Applications of DGA

-DGA can also be used to detect faults in LTCs, using forexample Duval Triangle 2 for compartment types, and Triangles2a to 2e for in-tank types.

-it can also be used for oils other than mineral oils, such asnatural esters (FR3, BioTemp), synthetic esters (Midel) andsilicone oils, using for example Duval Triangles 3.

DGA in LTCs at IEC/ IEEE:


62/65

6262

Duval Triangle 2 for compartment types Duval Triangles 2 for in-tank typesN1 (MR types M, D)N3 (MR types VR, VV)N4 (MR types R, V)N5 (MR types G, UZD)

Duval Triangles 3 for Non-Mineral oils


63/65

636363

Mineral oil

FR363

DGA in wind farm transformers at CIGRE


64/65

64646464

64

-Because they are usually Padmount transformers not

designed for that purpose, many tend to form lots of gases,as a result of:

-Corona PDs, because of poor oil impregnation.

-Stray gassing of oil, because of abnormal overheating.

Stray gassing of oil at CIGRE


65/65

65656565

Stray gassing of oil at CIGRE

-With mineral oil, H 2 at T200C.

-With vegetable oils (e.g.,FR3), H 2 at T200C.

65

Documents

Smart Grid Monitoring by Duval