Harmonic levels in LV networks and methodology for assessing the impact of new equipment

Transcription

1 1 Harmonic levels in LV networks and methodology for assessing the impact of new equipment Panel on New Harmonic Sources in Modern Buildings: Characterization and Modeling 14PESGM1020 Jan Meyer Technische Universitaet Dresden, Germany

2 2 Agenda Background Characterization and impact of new equipment Compact fluorescent lamps (CFL) Electrical vehicle chargers (EVC) Impact of circuit topopolgy changes Survey of harmonic levels in public LV grids Framework of the survey Some analysis results (3 rd harmonic, 5 th harmonic) Historical trend

3 3 Types of Distortion Below 2 khz Harmonics (Linked to power frequency) Above 2 khz (up to 150 khz) HF emission/supraharmonics (Independent of power frequency) Compact fluorescent lamp (P < 25 W) Compact fluorescent lamp (P > 25 W) i(t) i(t) t Devices with nonlinear i/u characteristic Simple electronic devices (line-commutated, e.g. no PFC) Mains signalling t Energyefficient circuit topologies, like active PFC Narrowband Power Line Communication

4 4 i(t) / A Harmonic phase angle and phase diversity Time domain (Halfwave) t / ms Im(I 5 )/A Frequency domain (5. Harmonics) (5) I Re(I 5 )/A (5) I 2 (5) I ges Cancellation effect: (5) I1 (5) I2 = 0,172 A = 0,135 A Phase diversity index: (5) k = 1 I + (5) (5) 1 2 p (5) (5) I1 + I2 0,131A = 1 = 0,57 0,307 A I I = 0,131A (5) ges Computer power supply (P > 75 W) and notebook charger (P < 75 W) Different circuit topologies have different behavior (passive power factor correction ppfc, no power factor correction npfc) Do prevailing phase angles exist for electronic mass-market equipment?

5 5 Agenda Background Characterization and impact of new equipment Compact fluorescent lamps (CFL) Electrical vehicle chargers (EVC) Impact of circuit topopolgy changes Survey of harmonic levels in public LV grids Framework of the survey Some analysis results (3 rd harmonic, 5 th harmonic) Historical trend

6 6 Characterisation of equipment Manufacturers prefer specific circuit topologies as consequence of given standards (e.g ) Different topologies have different prevailing harmonic phase angles Effective cancellation for 5 th harmonic, but less efficient for 3 rd harmonic) Efficiency of cancellation depends on mix of equipment (circuit topologies) Is it possible to identify dominating technologies by grid measurments? 5 th current harmonic 228 electronic mass-market devices CFL k = 0,78 (5) p k = 0,44 (3) p ppfc Lighting 25 W Lighting > 25 W Other equipment 25 W Other equipment > 75 W 30 npfc 330

7 7 Equipment Harmonic Database More than 10 labs from all over the world More than 500 different devices Web-based platform for exchanging measurements of harmonic emission of single phase equipment

8 8 8 6 Field study - customer terminal (CFL) Waveform CFLs OFF CFLs ON th harmonic current i(t) / A 4 2 I imag / ma Winkel / I real / ma ON/OFF comparison for switching CFLs and ISLs (incandescent lamps): Similar magnitudes ( 210mA), but different phase angles (-40 -> 148 ) Identification of connected equipment based on harmonic phase angle seems to be possible

9 9 Field study - LV feeder (EV charging) Measurement up to 150 khz Measurement up to 2.5 khz Transformer Junction box Feeder end 10 single family houses 10 electrical vehicles (EV) of 4 different brands Coordinated ON/OFF switching of all EVs in one phase

10 10 LV feeder study Individual current harmonics 5 th harmonic Electrical vehicles 5 th harmonic Sum of households Type A Type B Type C Type D Low emission of households in 4 th quadrant Similar 5 th harmonic phase angle for different EV types in the 3 rd quadrant (About 90 phase difference to typical household emission) Phase angle change from 4th to 3rd quadrant is expected, when cars are connected

11 11 LV feeder study Total harmonic current 5 th harmonic at junction box 3 rd harmonic at junction box Switching ON Switching OFF th harmonic current shifts phase (considerable diversity between households and EVs) 3 rd harmonic current does not shift phase (almost no diversity between households and EVs)

12 Impact of technology changes 350 5th harmonic current Cancellation effect 175 CFL 0 (5) k p I real / ma Most effective cancellation for CFLs in combination with passive PFC equipment Virtually no cancellation effect for 5 th harmonic by active PFC equipment No PFC (past) Passive PFC (today) Active PFC (future)

13 13 Agenda Background Characterization and impact of new equipment Compact fluorescent lamps (CFL) Electrical vehicle chargers (EVC) Survey of harmonic levels in public LV grids Framework of the survey Some analysis results (3 rd harmonic, 5 th harmonic)

14 14 Total harmonic emission of a LV grid I imag / A L1 L2 L3 5th current harmonic of a residential grid CFL ppfc npfc I real / A Distinctive clouds represent a nonrandom behavior with a prevailing direction Location of prevailing direction may indicate dominating equipment mix (this case: npfc and ppfc devices) Increasing number of CFLs or EVs could slightly reduce the 5 th current harmonic Balance between the different types of equipment is essential Is this result representative for other LV grids?

15 15 Improvement of vizualisation 3th current harmonic of a residential grid Color-dependent densitiy charts (2-D histogram) Improvement for explorative analysis (more information is revealed) BUT, processing of many sites requires further reduction of data amount and respective assessment indizes

16 16 Assessment indices (2 stage procedure) (A) Variation ratio (diversity index) : VR ( h) = 1 n i= 1 n i= 1 I I ( h) i ( h) i Measure for level of diversity (significance of prevailance) Opposite to prevailing ratio High value -> no prevailing direction Low value -> prevailing direction (B) Prevailing vector I V (only useful in case of low diversity): Calculation of vector sum n ( h) ( h) ( h) ( h) VEC = i = VEC VEC i= 1 I I I φ Magnitude Angle I θ = n ( h) i= 1 V = φ ( h) ( h) V VEC ( h) ( I ) 2 i n

17 17 Range of diversity index 150 Low diversity Medium diversity High diversity IV = A θ = 43 V VF = IV = A θ = 84 V VF = IV = A θ = 19 V VF = 0.8 Very low diversity (VL) VF 0,05 Low diversity (L) 0,05 < VF 0,11 Medium diversity (M) 0,11 < VF 0,2 High diversity (H) VF > 0,2 Prevailing vector meaningful

18 18 Survey on harmonic behavior of public LV grids Systematic survey in Germany More than 30 network operators More than 150 sites Comparability Duration: 14 days Interval: 1 minute Location: MS/NS-substation (LV busbar) Same type of monitors Classification of sites acc. to: Consumer topology (4 classes) Generation topology (4 classes) Grid topology (3 classes) Headoffice of network operators

19 19 Network classification schema Classification based on topology classes and priorisation according to the expected impact on harmonic levels High priority: Consumer topology A1: Single family houses (small, medium, large) A2: Multi family houses (small, medium, large) A3: Shopping A4: Office A5: Special consumer Medium priority: Generation topology E1: No generation E2: Generation smaller than 10% S r E3: Generation between 10% S r and 50% S r E4: Generation larger than 50% S r Low priority: Network topology N1: Total length of lines smaller 800m N2: Total length of lines between 800m and 1600m N3: Total length of lines larger than 1600m 96 possible combinations Preference on grids with pure consumer topologies A1 to A4

20 20 Coverage of consumer topologies Distribution of sites for consumer topology Distribution of customer number per grid for SFH c.d.f. A1: SFH A2: MFH A3: Shopping A4: Office Number Major focus on grids with residential customers Good representation of different grid sizes

21 21 Variation ratio of all sites 100% 80% 60% 40% 20% 0% 3rd harmonic A1 A2 A3 A4 H WM L VL Most residential grids show clear prevailing phase angles. Higher probability of larger variations for shopping and office 100% 5th harmonic 80% 60% 40% 20% 0% A1 A2 A3 A4 H WM L VL

22 22 Prevailing 3 rd harmonic vectors 3 rd harmonic (SFH) 3 rd harmonic (Shopping) 3 rd harmonic (Offices) Residential: Small variation Office and Shopping: Higher variation between sites Virtually no impact of generation and network topology observed No prevailing phase angles in 1 st quadrant

23 23 Prevailing 5 th harmonic vector 5 th harmonic (SFH) 5 th harmonic (Shopping) 5 th harmonic (Offices) Residential: Small variation; no dominance of CFLs Office and Shopping: Higher variation; Dominating number of measurements in 2 nd quadrant Virtually no impact of generation and network topology observed

24 24 Some further data analytics (2) 40 Span vs. average current 95-%-percentile vs. average current in A 20 in A 20 I S 5 15 I I V in A A1 A2 A3 A I V in A Span vs. average significantly higher for shopping (1,5) compared to residential topologies (0,6) -> no distinctive evening peak for 5 th harmonic current in many residential grids

25 25 I imag / A rd harmonic current passive PFC no PFC CFL Morning Evening I real / A Historical data I imag / A 5th harmonic current I real / A Measurement on Saturdays in 1999 and 2010 for 2 load states: No changes in consumer or network topology (350 residential customers in multi family houses) Magnitude: Decrease of 5th harmonic, but increase of 3rd harmonic Phase angle: Shift indicates increased number of passive PFC equipment (technology change) Morning Evening CFL passive PFC no PFC

26 26 Conclusions Present management for 5th harmonic seems to be more effective than for 3rd harmonic Different circuit topologies and their share have to be considered in modelling studies Emission of equipment considerably changes with changing voltage distortion, especially for active PFC equipment and inverters Every participation in PANDA project is highly appreciated (check out the website and send us an ) Measurements of prevailing harmonic phase angle in other countries are very welcome (contribution to IEC SC77A WG1)

27 27 Thank you for your attention! Contact details: Jan Meyer Technische Universität Dresden Institute of Electrical Power Systems and High Voltage Engenieering Dresden tel fax