Microinverters and Anti-Islanding - how does it work? Technically?

lunarpower

How do microinverters actually accomplish anti-islanding? At a technical, component-and-control level?

For example, once they are up and running, they are producing 120/240 VAC, so how would they know if the grid was not producing 120/240 VAC (since they would be sensing the result of their own generation)?

Are they disconnecting periodically in order to sample the grid voltage and frequency? (Which presumably results in some loss of energy sent to the grid during normal operation; if so, how much gets lost?)

Since there are typically a number of microinverters working at the same time, then how do they all coordinate to sample the grid at the same time? If one was out of sync, then the anti-islanding wouldn't work, and you wouldn't know, and that would be a hidden hazard.

Anyone know?

SolarPowered

Re: Microinverters and Anti-Islanding - how does it work? Technically?

When the 60Hz signal is cut from the grid the inverters stop producing power.
The frequency between any grid inverter has to intercept 60Hz in frequency in order to initiate operation. In Europe it's a 50Hz frequency.

Micro inverters operate between 95~96.5% efficiency. String inverters are beginning to surpass 97.5% efficiency however dirty panels in one string spoil the whole bunch.

BB.

Re: Microinverters and Anti-Islanding - how does it work? Technically?

From an engineering, it is a pretty interesting question.

Most of us are very familiar with "Voltage Sources"... A 12 volt car batter is a voltage source. It will try its darndest to maintain 12 VDC (sourcing current, sinking current, etc.).

They utility grid is also a "Voltage Source". They Alternator (AC Generator) outputs a sine wave form that goes from -339 volts to +339 volts (and back) 60x a second (this is for a ~240 VAC RMS utility connetion). If there is a load, the utility will supply enough current to keep the voltage sine wave within its specifications. Turns out, that the utility "grid" is a bi-directional source--It will output current/energy to drive a load, but you can also send the current "backwards" and virtually "charge" the utility grid.

You can take a generator (or a huge generator for a hospital) and synchronize their output voltages together, throw a switch, and throttle up diesel engine and the alternator will push current out to the grid (and local loads at the hospital). If the "grid" goes away (voltage goes out of spec., frequency goes out of spec., current goes out of spec., etc.), diesel genset will simply disconnect from the grid and keep supplying power locally.

Turns out, from an engineering point of view, there is a simpler way of "Grid Tie" connection solar arrays so the utility grid if you make some assumptions.

First is--Imagine you have a 24 volt lead acid battery and a 12 volt lead acid battery and you connect them together--Which ever battery "is bigger", that will be the final bus voltage (and you will smoke some wire/pop a fuse, etc.). That is what can happen when you have a diesel genset and the utility power if they are not synced and they are connected together. The voltage wave forms are not "aligned" and you will have "short circuit" current flow (or even worse).

Instead, if you drive the voltage source from a current source--Things work out really well. For example, if I connect a 24 volt solar panel (current source--The output current will be 8 amps from 0 to 30+ volts). The battery as the voltage source regulates the bus voltage, and the current source simply "injects" current into the battery bus. And the "charge controller" can limit charging current when the battery is full.

Similar with a GT Inverter--They are (usually) current sources. They will output voltage according to the equation I=V/R -- The GT inverter simply "follows" the AC sine wave as it swings from -339 to zero volts to +339 volts... Where the average current is the available power from the solar array. There is no "frequency syncing" or locking of frequency involved. The GT inverter simply measures the Voltage, Frequency, and possibly the "quality" of the sine wave--If all of the parameters meet spec. for 5+ minutes, it then turns on its current output. For all intent, the GT Inverter is "recharging" the AC grid and sharing the local AC loads--Just like a car's generator/alternator + Battery do... It just does in in the "AC domain" instead of DC.

Note, if the AC power "goes away", the inverter would not output any current. If the frequency "goes away", the unit will shut down within a couple cycles. If the frequency goes out by more than 0.5 to 1.0 Hz from standard, if the voltage goes outside of ~210-264 VAC RMS, the GT inverter will shut down.

An off grid inverter must generate a stable AC voltage wave form and supply current as needed... Therefore, they are AC voltage sources and cannot (unless designed to) connect with other Inverters (or generator, utility grid) as they will be "out of phase" and something will smoke.

There are other ways of doing this--But this is the simplest and probably most common for GT Solar Inverters).

-Bill

SolarPowered

Re: Microinverters and Anti-Islanding - how does it work? Technically?

Once current becomes less than or equal to that of the micro inverter at 60hz, then there is no longer a level of resistance for micro inverter to sample at a rate greater than that to the micro inverter output. There fore it no longer will produce power to make islanding occur.

Voltage doesn't matter. Simply because depending at what location your home is at from the closest or nearest point to grid. Voltage per pole can range from 105V to 120V. Grid current should always be greater than the current produced by inverter.

lunarpower

Re: Microinverters and Anti-Islanding - how does it work? Technically?

Thanks for the responses, SP, BB.

BB. wrote: »

...
Similar with a GT Inverter--They are (usually) current sources. They will output voltage according to the equation I=V/R -- The GT inverter simply "follows" the AC sine wave as it swings from -339 to zero volts to +339 volts... Where the average current is the available power from the solar array. There is no "frequency syncing" or locking of frequency involved. The GT inverter simply measures the Voltage, Frequency, and possibly the "quality" of the sine wave--If all of the parameters meet spec. for 5+ minutes, it then turns on its current output. For all intent, the GT Inverter is "recharging" the AC grid and sharing the local AC loads--Just like a car's generator/alternator + Battery do... It just does in in the "AC domain" instead of DC.

Note, if the AC power "goes away", the inverter would not output any current. If the frequency "goes away", the unit will shut down within a couple cycles. If the frequency goes out by more than 0.5 to 1.0 Hz from standard, if the voltage goes outside of ~210-264 VAC RMS, the GT inverter will shut down.
...

Hmm, current source. Oh, OK, the light bulb turned on!

So then, in terms of ideal circuit theory, for worker safety on the Utility Transmission Line, it sounds like you actually would Not need an anti-islanding off switch, because without an AC source to "follow", you'd have DC, if anything, and that would stop at the distribution transformer. But you would still need anti-islanding for any work done on the service drop or transformer.

Conductor protection would then be handled by the DC ratings of whatever fuses or CB's are between the solar panels, inverter, and distribution transformer secondary winding.

Then this also means that absent any weird interactions, you could have more than one microinverter type on your system. Have one set of panels controlled by microinverters from brand X, and another set of panels controlled by brand Y. Head-to-head reliability testing!

Anyone see any problems with my (late night) reasoning?

NorthGuy

Re: Microinverters and Anti-Islanding - how does it work? Technically?

I think it's more complicated than this. The voltage follow the sine wave, so it rises from 0 ro 339, then falls back again and goes negative. You don't get a single voltage, but rather the whole set of voltages from 0 to 339. If GT inverter has 3kW array and voltage is only 10V, it would be pushing out 300A. That would produce a really bad power factor. If it doesn't push this current, it has to store the energy somewhere and then push it out when voltage is higher. It also needs to watch for the polarity because it changes 120 times per second and it needs to push out current of the correct polarity. So, it has to be some fair amout of voltage monitoring and some mechanisms for time-shifting the energy to produce good power factor.

When grid power goes off, the voltage will drop because the inverter(s) cannot power the whole grid. Inverter may detect this and shut off.

Cariboocoot

Re: Microinverters and Anti-Islanding - how does it work? Technically?

You don't actually need to know how it works, only that it does.
You are correct the that grid-tie inverters sample the AC they are connected to to be sure the power is within spec for both Voltage (RMS) and frequency. If it is found to be out of spec, they shut down. This effect can be demonstrated 'locally' by AC coupling to a suitable off-grid inverter such as the SMA Sunny Island which will purposefully shift its output power frequency to 'drop' the GTI when its power is not required.

Regarding your other inquiry, yes you can use multiple brands/sizes/types of GT inverter on the same system as each will detect the power supplied from the grid and react accordingly. It is not like off-grid inverter 'stacking' where the inverters must communicate with each other to make sure the output is the same waveform for all.

If you really want to understand how it works you're going to have to study electrical engineering for about six years.

jnh

Re: Microinverters and Anti-Islanding - how does it work? Technically?

Inverter manufacturers each have slightly different methods for island detection. The interconnect safety requirements (UL 1741, IEEE 1547, parallel EU standards, etc.) dont mandate any particular technique, so they're free to go with anything that can be demonstrated to reliably pass all the tests. The more difficult condition to detect would be when local loads (on your side of a line break) happen to exactly match the wattage output of the inverter(s), allowing energy supply and demand to potentially remain in balance for a while-- very unlikely to persist for long in the real world, but safety shouldn't depend on luck, so they have to design for this contingency anyway.

Here is Enphase's patent that goes into some detail about the method used in their microinverters:

http://www.google.com/patents/US7899632


<< Method and apparatus for anti-islanding of distributed power generation systems
US 7899632 B2

Abstract
A method and apparatus for anti-islanding of distributed power generation systems having an inverter comprising a phase locked loop (PLL), a phase shift generator for injecting a phase shift into the PLL during at least one sample period, and a phase error signature monitor for monitoring at least one phase error response of the PLL during the at least one sample period.
...
The phase error signature monitor:
- integrates the at least one phase error response over at least one integration period to obtain at least one integrated phase error response;
- compares the at least one integrated phase error response to at least one threshold; and
flags the at least one sample period as indicative of a possible islanding state when the at least one integrated phase error response satisfies the at least one threshold.
...
declares the islanding state when a first number of the at least one sample period within a second number of the at least one sample period are flagged as indicative of the possible islanding state.
...
the phase error signature monitor comprises at least one resettable integrator, a comparator, at least one sampler, a subtractor, and an islanding decision controller.
...
the phase shift is of a magnitude of 50 microseconds over a single cycle of a frequency of a commercial power grid.
...
phase error response comprises a first phase error response and a second phase error response, wherein the first phase error response occurs prior to an injected phase shift and the second phase error response occurs after the injected phase shift.
...
the phase error signature monitor:
integrates the first phase error response over a first integration period to obtain a first integrated phase response;
integrates the second phase error response over a second integration period to obtain a second integrated phase error response;
compares a difference between the first integrated phase error response and the second integrated phase error response to a threshold
...
a controller coupled to the plurality of inverters, wherein the controller communicates a message to the plurality of inverters that causes the phase shift generator of each of the inverters of the plurality of inverters to inject the phase shift simultaneously into the plurality of inverters
...
In one embodiment, where the connected grid operates at a frequency of 60 Hz, a phase shift of magnitude 50 microseconds over a period of one cycle (i.e., 16.7 milliseconds) is injected at 0.5 second intervals; such an injected phase shift represents a phase shift of one degree and causes an insignificant distortion to the current injected into the grid and/or load. Alternative embodiments may utilize different phase shift magnitudes, durations, and/or injection intervals. While the micro-inverter 102 remains connected to the utility grid, the DPLL 304 produces a certain phase error response as a result of the injected phase shift. If the micro-inverter 102 becomes disconnected from the grid, the DPLL 304 produces a different phase error response as a result of the injected phase shift.
...
The DPLL 304 comprises a phase detector 402, an adder 404, a Proportional Integral Derivative (PID) controller 406, an adder 408, and a numerically controlled oscillator (NCO) >>

It goes on with more details, albeit with a lot of repetition and legalese. The Google patent link above includes some helpful block diagrams.

I think certain other designs may inject a high-frequency pulse near the zero-crossings, rather than a small phase error, and measure the load/grid response to that.

lunarpower

Re: Microinverters and Anti-Islanding - how does it work? Technically?

Thanks for the further responses, N, C, jnh.

Regarding need to know, no, strictly speaking, you don't need to know the How; but I figure the more you know, the less likely you are to end up in this fellow's predicament (and many thanks to him for sharing it publicly). (Of course, the counterargument would be a little knowledge is a dangerous thing ....)

Thanks for the patent link. I hadn't thought to look for that. Looks like they only monitor frequency -- I guess that should work -- no grid, no AC; and you wouldn't have to turn it off to sample. Well, time to do some reading.

Cariboocoot

Re: Microinverters and Anti-Islanding - how does it work? Technically?

That fellow's predicament had nothing to do with GTI & anti-islanding; it was fake panels passed off as the real thing if I recall correctly. Not much you can do about people putting UL stickers on equipment that isn't really certified.

jaggedben

Re: Microinverters and Anti-Islanding - how does it work? Technically?

lunarpower wrote: »

Thanks for the patent link. I hadn't thought to look for that. Looks like they only monitor frequency -- I guess that should work -- no grid, no AC; and you wouldn't have to turn it off to sample. Well, time to do some reading.

It's possible that the anti-islanding function may only rely on frequency, but it's also the case that the inverters monitor voltage, and are required to shut down in over voltage or under voltage situations. This may be an essentially independent process, but I don't know. I do know that most inverters monitor both phase-phase voltage and phase-neutral voltage, but that the latter is not required; an inverter can be setup without a neutral and a few GT products are able to do this.

SolarPowered

Re: Microinverters and Anti-Islanding - how does it work? Technically?

jnh wrote: »

The more difficult condition to detect would be when local loads (on your side of a line break) happen to exactly match the wattage output of the inverter(s), allowing energy supply and demand to potentially remain in balance for a while-- very unlikely to persist for long in the real world, but safety shouldn't depend on luck, so they have to design for this contingency anyway.

That only happens for short periods literally under a few seconds.
The interruption from lack of resistance from the line side will abruptly and momentarily change not only frequency, the whole wave pattern changes and the nature of its wave will conduct wattage almost like an algorithm gone bad, due to lack of resistance from the line side, as it becomes "unbalanced" . Grid tied Pure sine wave inverters are always in a process of sampling frequency from a true sine wave induction motor, once that process and level of resistance is missing, 2 or more inverters are constantly trying to sample a pure sine wave form that will not directly match that of an induction motor. The sampling then becomes more redundant to error every fraction of a second. In every home there are appliances from fridges, to clocks, pool pumps, heaters etc. Since there is a demand for a power source from such inverters, hypothetically not one inverter will equal another or be able to sync up the frequencies simultaneously as the inverters are also bleeding wattage out through the line side to the transformer. Hence the fact grid power must be greater than or equal to inverter for operation, in any case grid power will always be greater than inverter potential, unless its a solar farm, or PG&E tells you, that you can't build the magnitude of solar you would desire, because the potential wattage would be greater than what the transformer can handle at line side.

Luigi chinè

Re: Microinverters and Anti-Islanding - how does it work? Technically?

lunarpower wrote: »

How do microinverters actually accomplish anti-islanding? At a technical, component-and-control level?
Anyone know?

The anti-islandid is done at control level, the reference frequency current is calculated by measuring the voltage grid, every about 1.5second the inverter tries to force a current with a different frequency far away from the tollerace of a nominal grid. If you are in island the grid line is a Resistive load, becasuese you are exactly in resonance, and you force a current with a different frequency also the voltage across the grid impedance will have the same frequency imposed, if you are not in island if you force the current with a certain frequency the voltage grid can not change the freqiency, staying at 60 HZ. The grig impedance is very low during noral operatio and you can not force any changing. So you take the system working at a different frequency, that you are still monitoring.
In island the impedace is purely resistive so voltage and current must have the same frequency for ohm's law.