Author(s):	Verhoeven, Bas
Organization:	KEMA Nederland B.V. Utrechtseweg 310, P.O.box 9035 6800 ET, Arnhem, The Netherlands
Type:	Report IEA  PVPS Task 5
Reference:	IEA  PVPS T5-07: 2002
Date of issue:	September 2002
Details:	45 p.: figures, graphs, photos, tables; with references and appendices
Download:	Open or download this report as PDF document (1422 kB)
Available from:	Task 5 members or Operating Agent of Task 5
Abstract:	Many international forum discussions have been dealing with ‘Islanding’. Islanding is when a disconnected part of the power network is sustainable powered by the connected PV-systems or other embedded generators for a period of 5 or more seconds. A general conclusion of these discussions was that views on the subject are very polarized. On the one hand, the islanding phenomenon is considered such a rare or improbable event that it does not merit special consideration. On the other hand, the mere theoretical possibility of unintentional islanding, confirmed in laboratory experiments, is sufficient for individuals to have great concerns over the possibility of islanding. The reality probability lies somewhere between the two extremes. An important issue here is the lack of any real data on how often and for how long islanding can occur in practice and the associated risk of occurrence. An important observation in the discussion about islanding is that the discussion is based on “personal feelings” and/or “intuition”, which make the discussions even more difficult.
	Various theoretical studies have been made to determine how often islanding can occur in a real power network. Disadvantages of these theoretical studies are the assumptions made to simplify the analysis. The results of these studies are therefore often disregarded. In the Netherlands, an intensive study was made in an attempt to help the discussion and to provide real numbers on how often and for how long islanding can occur in a distribution network. This study is to measure the loading of a representative residential area together with the power produced in a PV-system. The measurements (active and reactive power) were taken every second for two years and stored in a computer for off-line analysis. The offline analysis is possible due to the direct correlation between the loading of the network and the power produced by the PV-system. This analysis result in actual figures, which predict precisely how often and for how long islanding can occur in the residential area studied. The main conclusions of this study are: The maximum PV-power in a power network for which balanced conditions never occur is approximately two to three times the minimum night load of the relevant power network. Balanced conditions and subsequently probability of islanding can not occur if PV-systems are installed on every house with a power rating of about 400 Wpeak or less. The penetration level of PV-systems does not significantly influence how often and for how long balanced conditions between the load and the PV-systems occurs. Balanced conditions between active and reactive load and the power generated by the PV-systems do occur very rarely for low, medium and even high penetration levels of PV-systems. The probability of a balanced condition does not depend on the number of houses connected to a feeder. The probability of occurrence of a balanced condition in a low voltage power network is well below 1E-6 to 1E-5. The probability of encountering an island is virtually zero! The overall conclusion of the work performed in this study is: Balanced conditions occur very rarely for low, medium and high penetration levels of PV-systems. The probability that balanced conditions are present in the power network and that the power network is disconnected at that exact time is virtually zero. Islanding is therefore not a technical barrier for the large-scale deployment of PV-system in residential areas.

[ Top ]