Voltage control with PV inverters in low voltage networks – In depth analysis of different concepts and parameterization criteria

 

Abstract

In some rural and sub-urban areas, the hosting capacity of low voltage networks is restricted by voltage limits. With local voltage control, photovoltaic generators can mitigatethe voltage rise partly and therefore increase the hosting capacity.This paper investigates the effectiveness and general performanceof different reactive and active power control concepts. It presentsthe findings of an extensive simulation-based investigation into theeffectiveness of voltage rise mitigation, additional reactive powerflows, network losses and power curtailment. The two mostcommon implementations of reactive power control have a similareffectiveness. The voltage rise can be compensated for by up to 25 % and more than 60 % for typical cable and overhead feedersrespectively. By additionally using active power curtailment of upto 3 % of the annual yield, the hosting capacity can be increasedby about 50 % and 90 % for the considered cable and overheadfeeder respectively (purely rural feeders).

EXISTING  SYSTEM:

In the recent years, exhaustive research results have beenpublished on how to mitigate the voltage rise caused by PVgeneration, with voltage rise being one of the most importantlimitations towards hosting capacity. While several local control concepts have been proposed,the most covered control modes are those currently required in grid codes: cos(P) (power factor as afunction of the injected power) and Q(U) (reactive power as afunction of the voltage at point of connection).  In , the impacts of the parameterization of local reactivepower control (Q(U)) on its performance have been investigated for a MV network located in Belgium. In, several localreactive power control methods for overvoltage prevention areinvestigated. In that paper, a new concept is proposed: a cos(P,U) control. Authors in propose a concept forensuring a more homogenous contribution of PV invertersconnected along a LV feeder. It consists of parameterizing thedead band of the Q(U) controller, depending on the networkimpedance at the point of connection. In, a concept forcurtailing the active power in an equitable manner viacentralized control or consensus control is proposed. In, a concept of offline coordination is proposed for sharing theactive power curtailment between generators connected to thesame feeder. It uses a similar approach to.

PROPOSED   SYSTEM:

While most of the literature related to Q(U) control focuseson steady-state considerations, recently some research has been devoted to studying the stability of Q(U) control. For this control type, it has been observed thatstability problems are not expected, even at high PVpenetration levels. Some of the publications also coverlaboratory tests and field tests , and in general confirmthe proper operation of the inverters even under realconditions. Finally, some papers propose the use of coordinated control,thereby combining central assets and controllers with PVinverters.  In , various strategies are proposed for increasing thehosting capacity of low voltage networks. In addition to activeand reactive power control by PV inverters, distributiontransformers with On-Load-Tap-Changing are considered.  In, different architectures for voltage control, whichinclude local, centralized and model-based control, arecompared. Coordinated control with adaptive zoning isproposed. In this concept, generators are grouped within zonesin order to limit the complexity and communication needs. In , a step model with increasing complexity isproposed to control the voltage in LV networks. The proposed concepts include coordinated control in which settings are sentto an on-load-tap-changer and to the PV inverters, depending onthe voltage measurements provided by smart meters. B.  ContributionA study of the literature in this field clearly shows the lack of general findings due to the diversity and complexity of thenetwork setups. Most existing studies are hard to generalise,due to their being case-studies with findings that are specificto that particular set of conditions, albeit having realisticresults. In light of this, this paper investigates the effectivenessand performance of different local voltage control concepts on a general basis. For this purpose, two generic feeders are used(cable and overhead line) which allows general conclusions tobe drawn, with all due precautions. The performance isevaluated on the basis of a comprehensive set of parameters:voltage level, feeder losses, reactive power consumption fromthe upstream network and power curtailment.

 

DISCUSSION AND CONCLUSION

This paper presented the results of comprehensive investigations on the performance and effectiveness ofcommon reactive power-based controls with optional activepower curtailment. These investigations tend to show a slightlybetter performance of the Q(U) control, as summarised below. The proposed Volt-Var Index (VVI) proved to be a suitableindicator for the comparison of different potential settings ofthe Q(U)-control (i.e. dead-band, droop value). These settingshave a large impact on the consumed reactive energy (factor sixbetween the least effective and the most effective Q(U) control) and on the effectiveness of the control (ability to compensate thevoltage rise caused by the infeed). A comparison between the cos(P) and Q(U) controls showsthat their general performance is comparable for an averageparametrization (VVI close to 5 – dead band until 1.05 p.u. andmaximum reactive power consumption beyond 1.07 p.u.).  On the one hand, their effectiveness is very similar: with aVVI-value of about 0.5, both controls lead to a reduction of theinitial voltage rise of 11.1 % by more than 4 percent (voltagerise compensation by about 40 % for the overhead line feeder).  The consumed reactive energy is also similar for both typesof control with a VVI-value around 0.5.  In terms of network losses, the Q(U) control parametrized at aVVI-value of around 0.5 leads to slightly higher losses than the cos(P) control (increase of 15 % vs. 10 %).In practice, a Q(U) control will however generally lead to a lower reactive energy consumption and to lower network lossessince it reacts on the actual voltage which is influenced by loadsconnected to the different voltage levels. Active power curtailment (P(U)) has been implementedand integrated in addition to the reactive power-based voltagecontrol, showing an interesting potential of increasing the hosting capacity by +50 % for the cable feeder (and 90 % forthe overhead line feeder) at the cost of a small reduction in the annual yield. This parametric study tends to show that the impact of thesettings of voltage controllers is limited, however, with a generally better performance of the Q(U) control, especially for very high penetration or for networks not yet needing this controls (low PV penetration). According to these investigations,the Q(U) curve should be parametrized to lie in the middle between the no-load voltage and the maximum allowed voltage (VVI=0.5), to ensure maximal effectiveness at minimal reactiveenergy consumption, network losses and curtailment.

 

REFERENCES

[1] G. Masson, S. Orlandi, and M. Rekinger, “Global market outlook forPhotovoltaics 2014-2018.” 2014.

[2] “EnergyMap – Auf dem Weg zu 100% EE,” 04-Feb-2015. [Online].Available: http://www.energymap.info/. [Accessed: 04-Feb-2015].

[3] Stetz, “Autonomous Voltage Control Strategies in Distribution Gridswith Photovoltaic Systems: Technical and Economic Assessment,”Kassel University press GmbH, 2014.

[4] G. Kerber, R. Witzmann, and H. Sappl, “Voltage limitation byautonomous reactive power control of grid connected photovoltaicinverters,” in Proc. Compatibility and Power Electronics, 2009.CPE’09., 2009, pp. 129–133.

[5] CENELEC, “Voltage characteristics of electricity supplied by publicelectricity networks.” 01-Mar-2011.

[6] M. Braun, T. Stetz, R. Bründlinger, C. Mayr, K. Ogimoto, H. Hatta,H. Kobayashi, B. Kroposki, B. Mather, M. Coddington, K. Lynn, G.Graditi, A. Woyte, and I. MacGill, “Is the distribution grid ready toaccept large-scale photovoltaic deployment? State of the art, progress,and future prospects,” Prog. Photovolt. Res. Appl., vol. 20, no. 6, pp.681–697, Sep. 2012.

[7] M. H. J. Bollen and A. Sannino, “Voltage Control With InverterBasedDistributedGeneration,”IEEETrans.PowerDeliv.,vol.20,no.1,pp.519–520,Jan.2005.

[8] VDE-FNN, “VDE-AR-N 4105:2011-08 Power generation systemsconnected to the low-voltage distribution network.” Aug-2011.

[9] “Generating Plants Connected to the Medium-Voltage Network,” Jun.2008.

[10] CEI, “CEI 0-21 – Reference technical rules for the connection ofactive and passive users to the LV electrical Utilities.” Dec-2013