Power Transformer Load Loss Measurement Gert Rietveld Ernest Houtzager Milos Acanski Dennis Hoogenboom Enrico Mohns Henrik Badura Ilija Pecelj ELPOW workshop, 31 Aug 2016
Power Transformer Load Loss Measurement
Gert RietveldErnest HoutzagerMilos AcanskiDennis HoogenboomEnrico MohnsHenrik BaduraIlija Pecelj
ELPOW workshop, 31 Aug 2016
• Power transformers - losses
• Why loss measurements?
• Transformer loss measurement systems
• ELPOW: TLMS calibration systems
• Summary
Outline
2
Power transformers are used toscale voltages in the grid
High Voltage low losses
Power Transformers
3
Loss mechanisms:1. No load losses (Zb = ), caused by iron core, continuous2. Load losses (Zb = 0 ), copper loss, depends on current3. Stray losses depend on design
Power transformer losses
4
1
2
Why loss measurements (1)
• Cost of losses in power transformers are comparable to the productcost TCO (total cost ownership)
• Environmental impact is significant
Cost of (no-)load loss equals product costs
Consequence: regulations - EcoDesign Directive:“improve environmental performance of energy-
related products through better design”
Why loss measurements (2)
6
Saving potential in use phase through more efficient designsestimated as 16 TWh/year ( 17 % of present losses)3.7 Mt of CO2 emissions (50 % of total Danish electricity consumption)
Consequence: more requirements in standardsIEC 60076-19 (uncertainty calculation), and upcoming IEC 60076-20:
Why loss measurements (3)
7
Tier 1: 1 July 2015Tier 2: 1 July 2021
99.1
99.2
99.3
99.4
99.5
99.6
99.7
99.8
99.9
0 20 40 60 80 100 120
Min
imum
PEI
[%]
Rated Power [MVA]
Liquid immersed - Tier 1Liquid immersed - Tier 2Dry-Type - Tier 1Dry-Type - Tier 2
Consequence: customers (utilities) put fines on losses in excess ofguaranteed maximum losses
Example calculation for a large transformer (100 MVA)• Guaranteed maximum losses: Pgar = 500 kW (0.5 % loss)• Fine for excess losses: 10.000 €/kW
Why loss measurements (4)
8
Reliable loss measurement with low uncertaintygives confidence and less discussions
3 % measurement uncertainty (150 ppm) corresponds to500 kW · 3 % · 10.000 €/kW = 150.000 €
(manufacturer and customer have to decide who pays for uncertainty)
Prob
abili
ty
Loss
Uncertainty matters! (compliance)
9
Actual loss
Ecodesign limit
Test resultfails to detectcompliance
Accurate test
Inaccurate test
Ecodesign allows no tolerance: if measured loss is above the limit,the transformer is not in conformity
High accuracy =low risk of incorrectdecisions
TLMS typical measurement range: 0 – 100 kV, 0 – 2 kAUncertainty: 3 % - 5 % (IEEE C57.12.00, IEC 60076-19, Ecodesign)
Traceable to international measurement standards
Transformer Loss Measurement
Power supply P = V · I · cos
Power: P = U · I · cos with 90
P = U · I · cos (90 - ) = U · I · with = (90 - ) 0
Measurement challenge
11
phase accuracy, not amplitude accuracy required!!
U and I are large numbers 1 % uncertainty is very easyis a very small number:for PF=0.01, = 0.57 1 % uncertainty is a big challenge!
(6 m , 0.34 min, 100 µrad)
TLMS consists of high-quality components
1 % losses
Only a calibrated TLMS system, traceable to national standards, givesproven, reliable quality in load loss measurements
TLMS calibrations
12
Two approaches in TLMS calibrations• Component calibration (< 0.5 %; at PF=0.01 0.2 min, 3 m )
Easier to perform, larger overall system uncertainty• System calibration (< 3 %; at PF=0.01 1 min, 17 m )
Difficult to perform, low uncertainty, all effects included
Increased measurement challenge: TUR = 3 – 10
Reference measurement accurate to < 0.15 min (50 µrad, 3 m )
Aim: system calibration of TLMS up to 100 kV and 2 kA, with anuncertainty of better than 50 µW/VA
ELPOW aim for TLMS calibration
13
Two approaches:
• Simultaneous generation of voltage V and current I
• Phase lock current I to externally applied voltage V
• Realise feedback loop
LMS for testing power trafos
Generating part Measuring part
• “voltage and current up to 150 kV and at least 2 kA…”• “uncertainties better than 50 ppm…”
Generation: phantom power
30 kVA0 V - 250 V
10 kVAR1: 0 A – 50 AR2: 0 A – 25 A
To do:• Stability transconductance amplifier must be optimized• Assessment of the phantom power source
2 channel source• 16 bit DACs• 10 VPK• phase lock• external synchronization
or 15 Hz … 60 Hz intern DACor 50 Hz line
Measuring system
Voltage TransformerClass 0.02
Current TransformerClass 0.002
Power ComparatorClass 0.01
U, I, , P, Q, S
I = 2 kA
U = 150 kV
LabView Software
To do overview:1. Optimization of the software
• timing to read the comparator• automatic correction of the transformer error
2. Uncertainty calculation3. Assessment of the system
Challenge: lock I to Vwithin 0.3 m (5 µrad)
TLMS system calibration
17
V I
DSP is a key element2nd reference CT + RD22 watt meter as check
Actual VSL implementation
18
Control (DSP)
Power amplifier (G)
Transformer /current generator
Power reference (RD22 watt meter)
CTs(3-stage compensated)
VT (CC-based capacitive divider)
Generation:• Rohrer wideband amplifier (20 A, 150 V) + step-up transformer
Measurement:• VT: HV capacitor (100 pF, 100 kV) + CC-based LV divider
– Uncertainty: 15 – 25 ppm (phase / magnitude)• CT: 3-stage wideband; errors < 1 ppm (uncertainty 5 ppm)
– Wideband 5 A shunt for conversion to voltage• Power: two 24-bit ADCs (NI)Expected overall phase uncertainty: < 25 ppm (25 µrad)
Verification: RD22 power meter (< 10 ppm) + CT + VT
Components of VSL system
19
The control loop has 2 main parts:
First part: high speed. Second part: high accuracy ( feedback)(NRC analogue system needs accurate 90 reference signal)
Results:• Extensive testing with different voltage signals; different phase
shifting blocks studied• Low noise: 4 – 5 µrad; agreement with power meter < 10 µrad
Control loop
20
ADC
90°
V [n] DAC
I
Q
I [n-m]
Phaseshifting
Z-n
• Much more noise than in lab (30 µrad vs 5 µrad)• Agreement VSL LL and RD22 reference: within 15 ppm• Agreement with NRC better than 25 ppm
First on-site trial
21
Power transformer losses are significant– Economically: TCO and fines– Environment: CO2 emissions
Requirements from EU regulations,IEC / IEEE on both the actual lossesand measurement accuracy
Conclusion
22
Transformer Loss Measurement Systems need calibration– Proven, validated accuracy < 3 % @ PF = 0.01– ELPOW: System calibration < 0.5 % @ PF = 0.01
– simultaneous generation– current phase locking
(promising first results < 0.3 %!)