My understanding of island condition is that part of the low and mid voltage AC grids become isolated from the rest of the national AC grid. The isolated part may have PV systems providing power. As long as there is a balance between produced and consumed power within the island, and the capacitive and inductive load in the island has a resonance frequency close to the normal AC grid frequency, the inverter will not detect that it is part of an isolated island and will continue to produce power. At least this is what the power companies suspect and fear.
At least some inverters have built in methods for detecting the island condition.
My questions are:
How do you calculate the probability for ending up with a stable island condition within the isolated part? It seems to me that there are so many varying parameters like e.g. variations in produced and consumed energy. I guess that the size of the capacitive and inductive loads within the island will have an impact, the effect on frequency by adding or removing a capacitive or inductive load should have less impact the higher the total values are within the island.
What methods are common for island condition detection within PV inverters?
How do you calculate the time for detecting the island condition for a given detection mechanism?
The classical way of detecting the island condition was to observe the value of the frequency in the system. As soon the frequency departs more than a tolerance stated by the regional electric system authority, the participants should recognize the island condition and stick to the agreed protocol, depending to the technology of each participant. Of course, in some regions more sophisticated algorithms are in effect, such as observing the time-derivative of the frequency.
The problem introduced by renewables is that there is no power-frequency governors and the classical criteria are becoming obsolete at fast pace
In 400V low voltage voltage grids the presence of a neutral can be detected.
If the neutral deviates too much from the average of the three line voltages.
Or if one of the individual line to neutral voltages is not within resonable limits:
+10%, then it is also clear that island operation occurs.
These systems do not work for medium voltage as the neutral is not relevant there.
In general, if the island operation is not detected but all voltages are OK and also the frequency, the user will have no problem. Converters do not have a high current if the grid appears again with a wrong phase (they should be made to withstand such conditions).
The grid operators should short circuit the grid before start working on it.
However, grid operators make a lot rules which do not take converters into account.
It is a way for them to reduce too easy grid injection by making the islanding detection expensive and complicated.
Thank you for your replies. I agree with Prof. Van den Bossche, it seems that the grid operators tries to add unnecessary costs to the PV installations by demanding expensive equipment when it all comes down to them following their own rules of safety when working on the grid.
However, I would need a better understanding of the mechanisms and calculation methods in order to respond to the grid operator and have a fair chance of reasoning with them.
The inverters we use have limiting values on frequency and voltage deviations according to the standard of the country where it is installed so a variation in voltage between two phases or between one phase and the neutral would be detected and lead to shutdown if it exceeds the limiting value.
Thanks Mr. Kjell Klasson,
Gradually things get more clear to me. Older technologies such as grids and roads are politically influenced and the decision making is not always pure technical. In the first years the vision was to throw out the renewables when the grid frequency changes. The did not want them ("not wanting" and "vision" are non-technical, they are intrigues). After a big grid shut down in Germany with the open bridge on the Elbe and high voltage line for a sail boat in Germany, the technical aspects gained importance again, and rules got different. But in fact one should simply not throw out renewables on frequency changes, that item should be removed. Throwing out renewables is de-stabilizing the grid on disturbances.
Grid owners use the feature when a grid is in island for repair and runs on a diesel to set the diesel at 47 Hz or 53 Hz and hence stop all renewables. All watches show a wrong time. But this problem could be solved in a different way. So they deviate in fact the grid frequency on purpose.
The grid exploitation can simply short circuit the grid to avoid that PV starts generating when works have to be performed.
Some converter manufacturers did like the frequency rules as it is some import limitation (again non-technical), and impose an accreditation by the grid owner (against the European rules: it should only be CE). The precise frequency limits avoid also that cheaper and more reliable non-processor based controls would be used. PV converters have a lifetime of about 10years, not much more, partly due to the use of "performing" processors, necessary for the rules. Other constraints are the high frequency stability of converters, this is not an easy aspect, but technical.
The converter should be damping for all harmonics (positive real part of impedance)
However, on large voltage changes one should shut down the converters: Above 10% too much voltage is too much for other consumers and one should also not feed in a short circuit, let say 50 to 70% of voltage. This is a part of the enable of the PV converter circuit.
On the other side, the distribution companies should install or use the tap changers to try to regulate the voltage as far as possible. On the other hand, PV owners do not imagine now that they will not be able to generate all the time. But it will be like this. Storing electricity is about 5 times more expensive than PV.
I think the following paper can resolve some of the questions
https://pcmp.springeropen.com/track/pdf/10.1186/s41601-017-0075-8?site=pcmp.springeropen.com
- Badges
- Science topic