Presentation - Renewables Grid Initiative
Higher integration of PV systems into
existing low-voltage networks by
probabilistic planning
Ing. Walter Niederhuemer, LINZ STROM Netz GmbH
Linz, Austria
Roland Sperr MSc, LINZ STROM Netz GmbH
Linz, Austria
-‐ 1 -‐ Agenda
•
presentation LINZ STROM Netz GmbH
•
framework for connecting distributed generation
– general and economic frameworks
– technical frameworks
•
probabilistic planning approach
– planning approaches
– detailed probabilistic planning approach
– simplified probabilistic assessment
•
results from the field test area Prendt
– reactive power control CosPhi(P), Q(U)
– combined control Q(U) and P(U)
– not fed in energy
•
conclusions
-‐ 2 -‐ LINZ STROM Netz GmbH
• electricity grid for 82 communi5es (up to 110 kV) • about 240.000 metering points -‐ 3 -‐ Framework for connecting distributed
generation
•
the objectives of the European Union are
– to increase the energy efficiency
– to increase the supply from renewable energy sources
•
from an economic point of view of the customer, a feed-in
– of 100% power
– in 100% of the time (any time) is required
•
feeding into the low voltage network represents for the distribution system
operator (DSO) a major challenge
•
according DACHCZ assessment rules, all feeders should not raise the voltage
more than 3% in a LV-network
– network, consumers and feeders share the available voltage range of +/- 10% Un
(EN 50160: 100% of 10-min mean values have to be within +10%)
•
if specified voltage limits according to TOR D4 and DACHCZ are reached or
exceeded and to enable feeding into the low-voltage network and to guarantee
the voltage quality, it is necessary
– to invest into low-voltage networks
– to limit the installed feed-in
-‐ 4 -‐ Framework for connecting distributed
generation
different optima: supplier ßà DSO
supplier
• full feed-in at any time
• maximum energy production and thus to optimize the profit
• small or no grid connection costs
distribution system operator (DSO)
• efficient distribution network
• network costs as low as possible
to increase the total energy feeding
• to find a macroeconomic optimum supplier ßà DSO
• compromise between network investment and the volume supplied by the
individual PV-generation
-‐ 5 -‐ Planning approaches
•
conventional planning approach
– in the conventional planning, it is assumed that the maximum power is fed at the worst
operating condition in the distribution network
– calculation with the maximum possible output voltage at the MW/LV transformer (107%)
•
3% voltage lift is reserved for suppliers
– calculation of the voltage lift according to the formulas of DACHCZ and TOR D4
– only then is it possible to guarantee the 100% feed-in at any time
– networks with high integration of distributed generation, come quickly to the limits of the
permissible voltage raise
– in fact, show current network conditions that critical voltage level rarely occur
•
a higher integration of decentralized feeders would be possible
•
currently assessment according to TOR or DACHCZ not taking into account existing network resources
from the entire system
-‐ 6 -‐ Use of real voltage band
Minimale und Maximale Spannungen je Trafostation
Minimum and maximum voltages of (Mit Einspeisung MS-Netz)
transformers (incl. feed-‐in in MV) 10,00%
9,00%
Garantierte
Reserve für
Spannungsanhebung
dezentrale Einspeiser
guaranteed reserve for voltage liI durch
by decentralized feed-‐in im
in NS-Netz
LV (TOR lt.DTOR
4) D4
8,00%
7,00%
short-‐terme 5ll medium-‐term useable voltage band Spannungsabweichung von 400V
6,00%
Kurz- bis mittelfristig
Spannungsband
für dezentrale
for dnutzbares
ecentralized feed-‐in Einspeiser
5,00%
4,00%
3,00%
Minimale
undand maximale
Spannungen
Sekundärseite
der Trafostation
Minimum maximum output auf
voltages at transformers 2,00%
1,00%
0,00%
-1,00%
-2,00%
-3,00%
-4,00%
-5,00%
-6,00%
-7,00%
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
1600
1700
Trafostation
•
not all MV/LV transformers on the 400V-side are operated with U=107%
–
–
•
depending on the voltage level in the MV network
depending on the transformer ratio
in many local networks, there are therefore theoretically larger voltage reserves as 3% for the
feed-in
-‐ 7 -‐ Planning approaches
•
probabilistic planning approach
– this planning approach takes into account the static behavior of the parameters
•
voltage fluctuation at the MV/LV transformer
•
feed-in power
– the aim of the planning approach is, to increase the installable feed-in power and to increase
the amount of energy supplied
•
at low network costs and
•
low amount of not feed-in energy
– the goal is only achievable if it is possible for the DSO to control or cut off the feed-in power
for rare and short periods of time when needed
implementation of a probabilistic reduction factor “F”
d
ΔSa
SkV
Ψ
φ
F
rela've voltage change feed in power [kVA], for PV-‐system [kWp] short circuit power at connec'on point grid angle angle of apparent power probabilis'c reduc'on factor -‐ 8 -‐ Simplified assessment procedure
-‐ 9 -‐ Probabilistic assessment (methodology)
measured PV-Power (kW/kWp) und expectation of the power at noon
expecta5on value Erwartungswert
trend of PV-‐power evalua5on of the expected values via density es5ma5on power measured voltage at MV/LV transformer (June–September) and expected value of the voltage
Kernel (Triweight-‐Func5on) Erwartungswert
expecta5on value Source: hRp://de.wikipedia.org/wiki/Kernel-‐Regression -‐ 10 -‐ Probabilistic assessment (methodology)
maximum voltage 107%
incl. feed-in into MV
and- minimum load
electrical substation
(HV/MV)
voltage lift ULIFT from
variation of controller and
load change
variation of
controller
UHub
ULIFT
Spannungsband fürPV z.B 3%
voltage band for PV e.g. 3% Bezugspunkt 0% S0pannungsanhebung
reference point % voltage liI transformator oautput voltage liI e.g. Spannungshub m Trafo z.B 3%
3% E = ELIFT * EPV ΔU resul'ng voltage liF ΔUHub varia'on of transformer output voltage from the reference point ΔUPv voltage liF by PV E total expecta'on value of resul'ng voltage liF Ehub expecta'on value of transformator output voltage Epv expecta'on value of voltage liF by PV E UPV ULIFT U UPV cumulated expecta5on ΔU = ΔULIFT + ΔUPV § the voltage values ΔULIFT and ΔUPV are two independently
occurring values
§ total expected value is determined by multiplying the individual
expected values
ULIFT UPv
UPV
total voltage liI -‐ 11 -‐ Probabilistic assessment (methodology)
factor F is depending on ra5o of ULIFT/UPV, so by the rated voltage on the MV/LV transformer and the allowed voltage-‐raising by PV-‐ genera5on ULIFT > UPV ULIFT < UPV Chart for factor „F“ expecta5on of feed-‐in energy -‐ 12 -‐ Probabilistic assessment (methodology)
example
• maximum voltage in MV-grid 107%
• voltage lift (variation of controller) at electrical substation (HV/MV) and at transformer station (MV/LV) 2%
• allowable voltage band (planning value) for PV-feed-in in LV-grid 3%; expectation value of 95%
Chart for factor „F“ • 30 kWp PV-feed-in 1 with CosPhi = 0,95
at connection point with SkV= 570 kVA
ψ= 30°
• 30 kWp PV-feed in 2 with CosPhi = 0,95
at connection point with SkV= 1200 kVA
ψ= 40°
ULIFT < UPV • ULIFT/UPV = 2% / 3% = 0,67
• factor F = 0,6 (from line ULIFT/UPV = 0,6)
• d1 = 30 kWp / 570 kVA * Cos(30 + 18) *
0,6 = 2,11%
• d2 = 30 kWp / 1200 kVA * Cos(40 + 18)
* 0,6 = 0,79%
• d1+ d2 = 2,9%
<3% (allowable voltage band for PV)
à feed-in OK
ULIFT > UPV expecta5on of feed-‐in energy -‐ 13 -‐ Field test area Prendt
transformer
station
field test area PRENDT
754 m a.s.l.
-‐ 14 -‐ Field test area Prendt
in the framework of the research project "DG DemoNet smart LV grid“
the probabilistic planning approach has been tested
1113 m 2 70mm² non-‐isolatetd powerline and 95mm² isolated overhead powerline 67m 35 K
29 8,25kWp/2
PRENDT_1
30m 70 F
697m 95 B
6 58m 70 F
33m 70 F
5,1kWp/2
50m 70 F
4 14,54kWp/3
40m 70 F
3 93m 70 F
14 20kWp/3
13 10kWp/3
68m 70 F
10kWp/3
12 36 74m 70 F
3,57kWp/1
9 106m 70 F
8,25kWp/3
63m 70 F
10,2kWp/3
32 124m 70 F
61m 95 B
35m 95 B
48m 95 B
10,2kWp/3
7 75m 95 B
734 m 70mm² non-‐isolatetd powerline and 95mm² isolatetd powerline 9kWp/3
8 6kWp/2
branch „Prendt P1rendt_1 “ • Abzweig – 14,54 kWp Bestand • 14,54 kWp exis5ng – 29,13 kWp DG DemoNet • 29,13 kWp D
G DemoNet • Abzweig P2rendt_2 branch „Prendt “ – 11,34 kWp Bestand – 87,22kWp DG DemoNet • 11,34 kWp • 87,22 kWp DG DemoNet • 142,23 kWp PV-‐Leistung kWp DG DemoNet) 142,23 k(116,35 Wp PV-‐power (116,35 kWp DG DemoNet) -‐ 15 -‐ 31 35 3,5kWp/1
83m 70 F
30 5,28kWp/1
30m 95 B
6a 8,82kWp/3
83m 70 F
1 104m 95 B
Trafo Prendt
160 kVA
uk= 6,2%
7kWp/2
71m 95 B
2,52kWp/3
Results field test area Cosφ(P)
voltage chart Spannungsverlauf
260
cos phi
CosPhi=0,9 = 0,9 kap.
(over-‐exited) (übererregt)
250
Spannung [V]
240
0%
5%
230
50%
95%
220
cos phi(P) (P)
CosPhi Kennlinie
chart 0,5
1,0
P/PNenn
100%
210
200
00:00:00
03:00:00
06:00:00
12:00:00
15:00:00
18:00:00
21:00:00
00:00:00
Spannungsverlauf
voltage chart 260
cos phi
CosPhi=0,9 = 0,9 ind.
(under-‐exited) 09:00:00
250
(untererregt)
Spannung [V]
240
0%
5%
230
50%
95%
220
100%
210
200
00:00:00
conventional assessment with CosPhi=1 à ΔU= +9%
03:00:00
06:00:00
09:00:00
12:00:00
15:00:00
18:00:00
21:00:00
00:00:00
measured maximum voltages (100% und 95% quan5le) transformer station
Prendt 2
ΔU [%]
U Q95
101,7%
233,91 V
109,2%
251,16 V
7,5 %
U Q100
102,5%
235,75
110,5%
254,15 V
8,0 %
-‐ 16 -‐ frequency of occurence voltage Prendt 2 CosPhi (P) control Results field test area Q(U) + P(U)
voltage chart Spannungsverlauf
Q(U) + P(U) 260
250
Spannung [V]
240
P(U)
109% - 112%
0%
5%
230
50%
95%
220
100%
210
200
00:00:00
Q(U)
106% - 109%
03:00:00
06:00:00
09:00:00
12:00:00
15:00:00
18:00:00
21:00:00
00:00:00
Spannungsverlauf
voltage chart 260
250
Spannung [V]
240
0%
5%
230
50%
95%
220
100%
210
200
00:00:00
conventional assessment with Cosρ=1 -> ΔU= +9%
03:00:00
06:00:00
09:00:00
12:00:00
15:00:00
18:00:00
21:00:00
00:00:00
measured maximum voltages (100% und 95% quan5le) U Q95
U Q100
transformer station
101,7%
233,91 V
102,5%
235,75 V
Prendt 2
107,2%
246,56 V
108,2%
248,86 V
ΔU [%]
5,5 %
5,7 %
-‐ 17 -‐ frequency of occurence voltage Prendt 2 CosPhi (P) control Results field test area Q(U) + P(U)
In Prendt 2, a Fronius datalogger was installed by the customer.
This data logger recorded the control actions ('events' for 223 days) of the P(U)-Control in the following form:
- timestamp of the control action
- duration of the activation
- minimum and maximum voltage of the activated phase(s) (1s-value)
chart of frequency of occurance of events frequency of occurence -> 9771
Anzahl Events
number of events: 9771 ~73%
about 7mit
3% weiner
ith a dDauer
ura5on o0-5s
f 0-‐5s ΔP=0%
P(U)
109% - 112%
ΔP= -100%
-‐ 18 -‐ Results field test area
Energie
5091 kWh
Leistung
90%
6,034 kW
Ertragseinbuße
loss of yield [kWh]
Ertragseinbuße
loss of yield [%]
100%
50%
25%
100%
50%
25%
>0 s
133,6
66,8
33,4
2,62%
1,31%
0,66%
>5 s
104,1
52,0
26,0
2,04%
1,02%
0,51%
expected power reduc5on
In the power inverter recorded voltages (1 s values) show, that during the control actions the voltage values are
less than 109%.
If we now assume that such control interventions take a very short time (compensation of voltage spikes when
devices are powered on) and takes into account the control curve to estimate the power reduction, the result is
a yield loss of 0.176% at 15.83 h control intervention time.
not feed-in energy amount is very low
in the range between 0.18% - 1.31%
-‐ 19 -‐ Results field test area
Voltage liI, voltage band winnings voltage liI and voltage band winnings at Prendt 2 Voltage liI Voltage band winnings conven)onal (calculated) -‐ 20 -‐ Conclusions
•
•
•
•
•
probabilistic planning approach represents a very effective method for improved
evaluation of network capacity for decentralized PV feed-in
while conventional assessment always assumes worst-case assumptions, the
presented probabilistic planning approach takes into account the static behavior of
the voltage on the MV/LV transformer and the feed-in power
it is shown that the worst-case assumptions only occur with low probability
if it is possible for the distribution system operator (DSO) to cutt-off or control the
feed-in power (P(U)-control) of one or more feeder as needed for rare short periods
of time when the upper voltage limit is reached, an increase of installed PV feed-in is
possible
both the results of the probabilistic planning approach and the results of the field tests
show that a doubling of installed photovoltaic capacity in existing low voltage
networks is possible with a small amount of not feed-in energy
-‐ 21 -‐ Thank you for your attention!
Ing. Walter Niederhuemer
LINZ STROM Netz GmbH
Fichtenstraße 7, 4021 Linz
Austria
Roland Sperr MSc
LINZ STROM Netz GmbH
Fichtenstraße 7, 4021 Linz
Austria
Phone: +43 (0)732 3403 3182
Mail: [email protected]
Phone: +43 (0)732 3403 6491
Mail: [email protected]
-‐ 22 -‐
© Copyright 2024