GRID VTAGE | Georg Wirth | Gerd Becker | Robert Pardatscher | … - GRID... · 2018-10-16 · worst...
1
Andreas Spring 1 | Georg Wirth 1 | Gerd Becker 1 | Robert Pardatscher 2 | Rolf Witzmann 2 GRID VOLTAGE INFLUENCES Most of the installed PV inverters do have a power factor specification of one. Hence, these PV inverters should feed-in no reactive but only active power. Various observations in low voltage grids show a dispersion of the active and apparent power feed-in and thus a reactive power flow that has an impact on the grid voltage. The investigations are based on the analysis of data from a very detailed measure- ment campaign in a distribution grid with high PV penetration in southern Germany. Every single inverter with a power factor specifi- cation of one shows a reactive power flow. The dependency of the reactive power flow on the active power can be described by a best-fit poly- nomial of the third degree. Fig. 1 shows these polynomials for several measured inverters. A negative reactive power stands for an inductive reactive power consume. The unintended re- active power depends therefore on the utilization rate of the inverter. Most of the installed inverters are inductive in full-load operation (Fig. 2). Fig. 1: Unintended reactive power of PV inverters in dependency on the active power. MOTIVATION (1) University of Applied Sciences Munich Department of Electrical Engineering 80335 Munich · Germany Phone +49(0)89/1265-3483 [email protected] (2) Technische Universität München Department of Electrical Engineering and Information Technology Institute of Power Transmission Systems 80333 Munich 1 2 3 4 5 6 7 8 9 10 11 12 13 −1.0 0.0 1.0 2.0 Inverter Reactive Power in kVAr Q max Q mean Q at P max For each inverter the reactive power best-fit poly- nomial of the first degree for the day with the highest active power feed-in in 2011 of a com- plete low voltage grid is compiled. Fig. 3 com- municates this polynomial. The size of the marker represents the quotient of the reactive power in full load operation of each specified inverter and the sum of all reactive powers in full load opera- tion (left) as well as the maximum reactive power flow of the specified inverter (right). Fig. 2: Unintended reactive power of PV inverters in full-load operation (green). Red and Blue indicate the maximum and medium reactive power of the inverters. Inverters with a maximum reactive power in the capacitive range are described by circles; in the inductive range a rectangle is used. A low voltage grid with seven measured PV sys- tems is displayed. There is one 20 kW system with 4.4 kVAr inductive reactive power in near full load operation. The mean reactive power of all seven systems is also inductive and amounts to 339 VAr. The exemplary low voltage grid has an inductive medium reactive power flow in full load opera- tion. This means the grid voltage should be de- creased due to the unintended reactive power flow. On the other hand there are some inverters that feed-in inductive reactive power in full load operation and boost the grid voltage. Hence, there must be areas in the grid with increased and areas with decreased grid voltages due to the unintended reactive feed-in. Fig. 3: Reactive power behavior of an inverter-fleet of a low voltage grid. In the following the variation of the grid voltage in one exemplary low voltage grid (U N = 400 V), during full load operation with and without the unintended reactive power flow will be presented. Therefore the active and reactive power flow of the day with the highest active feed-in in the year 2011 was chosen. The input parameters of the simulation are the active and reactive power flow of all measured PV systems. Out of the measured active power a normalized average value for the active feed-in is calculated. This normalized profile multiplied with the rated power of the remaining PV systems is the input data of all non-measured PV systems. The reactive power of the non-measured PV systems is calculated in three different ways: The first way (C1-Mean) is an average reactive power out of all measured reactive power flows. In the second calculation (C2-Ind) the course of the PV system with the maximum inductive power flow is divided by the rated active power of this system. This course is then multiplied by the rated power of the non-measured systems and finally the specific system assigned. The third way (C3-Cap) is similar to the second by using the maximum capacitive power. Fig. 4 shows the relative grid voltage on all nodes for the timestamp with the highest active power by neglecting the unintended reactive power flow of PV inverters with a power factor specification of one. Fig. 5 displays the results of the grid voltage on all nodes with and without reactive power feed-in for the calculation way C1-Mean. Therefore the difference between the two calculations is exposed. The absolute values of voltage changes due to unintended reactive power are marginal. The simulation ways C2-Ind and C3-Cap are worst case scenarios. The voltages on all nodes of the distribution grid for the scenario C2-Ind and C3-Cap are shown in Fig. 6 and 7. By apply- ing the unintended reactive power flow according to scenario C2-Ind the maximum and also the medium grid voltages are reduced. By applying the reactive power according to scenario C3-Cap only a slight increase of the grid voltage occurs. Fig. 4: Relative Grid voltage in full-load operation by neglecting unintended reactive power flow of PV inverters. Fig. 5: Comparison of the grid voltage in full load operation of the PV systems with and without the unintended reactive power flow of PV inverters for scenario C1-Mean. Fig. 6: Voltages on all grid nodes for simulation scenario C2-Ind. Fig. 7: Voltages on all grid nodes for simulation scenario C3-Cap. Despite a power factor specification of one, PV inverters supply and/or consume reactive power. Under normal operating conditions this reactive power is not relevant for grid operation and grid planning. The results of this study show an inductive behavior of the inverters with the highest deviation from the power factor specifi- cation in full load operation. This means that CONCLUSION even if a larger number of inverters – that do not fulfill their power factor specifications – is installed, no grid disturbance with voltages that are too high will be expected. The grid voltage, rather, will be decreased. Nevertheless, if voltage or overload problems are not explainable by the active power flow, they can be a result of the unintended reactive power flow and their influence on grid voltage and grid loss. Those grid regions have to be analyzed by applying the unintended reactive power flow of PV inverters with a power factor specification of one. This approach can help network operators locate and understand the reasons of grid problems. from REACTIVE POWER FLOW of PHOTOVOLTAIC INVERTERS with a POWER FACTOR SPECIFICATION of ONE REACTIVE POWER BEHAVIOR of the INVERTER-FLEET in a LOW VOLTAGE GRID −0.2 0.0 0.2 −5.0 −4.0 −3.0 −2.0 −1.0 0.0 1.0 2.0 Q Mean = −339 VAr Q P Max = −4.41 kVAr at P Max = 16.45 kW Gradient in kVAr/kW Reactive Power in Full Load Operation in kVAr −0.4 −0.2 0 0.2 −1.0 −0.5 0.0 0.5 1.0 (Q P−Max / Q Total ) Mean = 14.3 % (Q P−Max / Q Total ) Max = 58.3 % (Q P−Max / Q Total ) Min = 1.32 % Gradient in kVAr/kW Reactive Power during Times of no Active Feed−in in kVAr 114 House Connections (HC) − 700 kW of installed PV power − 30 PV systems − 7 measured PV systems SIMULATION of the INFLUENCES on the GRID VOLTAGE Capacitive Consumer Inductive Consumer Capacitive Consumer Inductive Consumer
Transcript of GRID VTAGE | Georg Wirth | Gerd Becker | Robert Pardatscher | … - GRID... · 2018-10-16 · worst...
Andreas Spring1 | Georg Wirth1 | Gerd Becker1 | Robert Pardatscher2 | Rolf Witzmann2Grid voltaGe influences
Most of the installed PV inverters do have a power factor specification of one. Hence, these PV inverters should feed-in no reactive but only active power. Various observations in low voltage grids show a dispersion of the active and apparent power feed-in and thus a reactive power flow that has an impact on the grid voltage. The investigations are based on the analysis of data from a very detailed measure-ment campaign in a distribution grid with high PV penetration in southern Germany.
Every single inverter with a power factor specifi-cation of one shows a reactive power flow. The dependency of the reactive power flow on the active power can be described by a best-fit poly-nomial of the third degree. Fig. 1 shows these polynomials for several measured inverters. A negative reactive power stands for an inductive reactive power consume. The unintended re-active power depends therefore on the utilization rate of the inverter. Most of the installed inverters are inductive in full-load operation (Fig. 2).Fig. 1: Unintended reactive power of PV inverters in dependency
on the active power.
motivation
(1) University of Applied Sciences Munich Department of Electrical Engineering 80335 Munich · Germany Phone +49(0)89/1265-3483 [email protected]
(2) Technische Universität München Department of Electrical Engineering and Information Technology
Institute of Power Transmission Systems 80333 Munich
1 2 3 4 5 6 7 8 9 10 11 12 13−1.0
0.0
1.0
2.0
Inverter
Rea
ctiv
e Po
wer
in k
VAr
QmaxQmeanQat P
max
For each inverter the reactive power best-fit poly-nomial of the first degree for the day with the highest active power feed-in in 2011 of a com-plete low voltage grid is compiled. Fig. 3 com-municates this polynomial. The size of the marker represents the quotient of the reactive power in full load operation of each specified inverter and the sum of all reactive powers in full load opera-tion (left) as well as the maximum reactive power flow of the specified inverter (right).
Fig. 2: Unintended reactive power of PV inverters in full-load operation (green). Red and Blue indicate the maximum and medium reactive power of the inverters.
Inverters with a maximum reactive power in the capacitive range are described by circles; in the inductive range a rectangle is used.
A low voltage grid with seven measured PV sys-tems is displayed. There is one 20 kW system with 4.4 kVAr inductive reactive power in near full load operation. The mean reactive power of all seven systems is also inductive and amounts to 339 VAr.
The exemplary low voltage grid has an inductive medium reactive power flow in full load opera-tion. This means the grid voltage should be de-creased due to the unintended reactive power flow. On the other hand there are some inverters that feed-in inductive reactive power in full load operation and boost the grid voltage. Hence, there must be areas in the grid with increased and areas with decreased grid voltages due to the unintended reactive feed-in.
Fig. 3: Reactive power behavior of an inverter-fleet of a low voltage grid.
In the following the variation of the grid voltage in one exemplary low voltage grid (UN = 400 V), during full load operation with and without the unintended reactive power flow will be presented. Therefore the active and reactive power flow of the day with the highest active feed-in in the year 2011 was chosen. The input parameters of the simulation are the active and reactive power flow of all measured PV systems. Out of the measured active power a normalized average value for the active feed-in is calculated. This normalized profile multiplied with the rated
power of the remaining PV systems is the input data of all non-measured PV systems. The reactive power of the non-measured PV systems is calculated in three different ways: The first way (C1-mean) is an average reactive power out of all measured reactive power flows. In the second calculation (C2-ind) the course of the PV system with the maximum inductive power flow is divided by the rated active power of this system. This course is then multiplied by the rated power of the non-measured systems and finally the specific system assigned. The third way (C3-Cap)
is similar to the second by using the maximum capacitive power.
Fig. 4 shows the relative grid voltage on all nodes for the timestamp with the highest active power by neglecting the unintended reactive power flow of PV inverters with a power factor specification of one. Fig. 5 displays the results of the grid voltage on all nodes with and without reactive power feed-in for the calculation way C1-mean. Therefore the difference between the two calculations is exposed. The absolute values of
voltage changes due to unintended reactive power are marginal.
The simulation ways C2-ind and C3-Cap are worst case scenarios. The voltages on all nodes of the distribution grid for the scenario C2-ind and C3-Cap are shown in Fig. 6 and 7. By apply-ing the unintended reactive power flow according to scenario C2-ind the maximum and also the medium grid voltages are reduced. By applying the reactive power according to scenario C3-Cap only a slight increase of the grid voltage occurs.
Fig. 4: Relative Grid voltage in full-load operation by neglecting unintended reactive power flow of PV inverters.
Fig. 5: Comparison of the grid voltage in full load operation of the PV systems with and without the unintended reactive power flow of PV inverters for scenario C1-mean.
Fig. 6: Voltages on all grid nodes for simulation scenario C2-ind.
Fig. 7: Voltages on all grid nodes for simulation scenario C3-Cap.
Despite a power factor specification of one, PV inverters supply and/or consume reactive power. Under normal operating conditions this reactive power is not relevant for grid operation and grid planning. The results of this study show an inductive behavior of the inverters with the highest deviation from the power factor specifi-cation in full load operation. This means that
ConClusion
even if a larger number of inverters – that do not fulfill their power factor specifications – is installed, no grid disturbance with voltages that are too high will be expected. The grid voltage, rather, will be decreased.
Nevertheless, if voltage or overload problems are not explainable by the active power flow, they
can be a result of the unintended reactive power flow and their influence on grid voltage and grid loss. Those grid regions have to be analyzed by applying the unintended reactive power flow of PV inverters with a power factor specification of one. This approach can help network operators locate and understand the reasons of grid problems.
from REaCtivE PoWER FloW of PHotovoltaiC invERtERs with a PoWER FaCtoR sPECiFiCation of onE
REaCtivE PoWER BEHavioR of the invERtER-FlEEt in a loW voltagE gRid
−0.2 0.0 0.2−5.0
−4.0
−3.0
−2.0
−1.0
0.0
1.0
2.0
QMean = −339 VAr
QPMax
= −4.41 kVAr
at PMax = 16.45 kW
Gradient in kVAr/kW
Rea
ctiv
e Po
wer
inFu
ll Lo
ad O
pera
tion
in k
VAr
−0.4 −0.2 0 0.2−1.0
−0.5
0.0
0.5
1.0(QP−Max / QTotal)Mean = 14.3 %
(QP−Max / QTotal)Max = 58.3 %
(QP−Max / QTotal)Min = 1.32 %
Gradient in kVAr/kW
Rea
ctiv
e Po
wer
dur
ing
Tim
es
of n
o Ac
tive
Feed
−in
in k
VAr
114 House Connections (HC) − 700 kW of installed PV power − 30 PV systems − 7 measured PV systems
simulation of the inFluEnCEs on the gRid voltagE
Capacitive Consumer
Inductive Consumer
Capacitive Consumer
Inductive Consumer
