Designing a Grid Tie Inverter with a PFC chip to control it?

[treez]

Hello,
Regarding Grid Tie Inverter...

Why are there no dedicated control IC’s for it like there are for eg PFC?
Surely the grid tie inverter control algorithm is very closely allied to the algorithm for PFC?

Surely I can do a GTI with a PFC chip (eg LT1248) as follows…

A PFC controller such as LT1248, when acting as a PFC, takes in the following signal inputs…
a)…..Mains voltage level (obviously a changing signal, at 50Hz)
b)…The current in its sense resistor
c)…The Vout of the PFC.

…Now, I can do exactly the same as this and get a grid tie inverter(?)……all I have to do is exchange the Vout signal for the Vin of the GTI.
……So basically, I place a DCDC converter upstream of the GTI, which feeds power at a constant rate into the input capacitor of the GTI, I then get the overall GTI control loop to regulate the GTI’s own input voltage to 400V.
The GTI will then surely just shovel current into the mains at such a rate that keeps its own input voltage at 400V?

 

[BradtheRad]

This grid-tied inverter design was posted a while back.

https://www.edaboard.com/threads/307293/



The control signal comes entirely from the grid waveform. Although I'm not familiar with designs used in the real world, it's hard to imagine anything more straightforward than this.

Something that would make a difference is the volt level you send onto the grid. Should it be 400 or 500 or 600V? I think it should be whichever delivers maximum power. An MPPT type of circuit would be handy here, for trying different V and A levels. I don't suppose it could change the transformer step-up ratio, but there ought to be a way.

 

[schmitt trigger]

Brad; I had seen this grid tie inverter somewhere else. They claim 80% efficiency, which I find very hard to believe, as the Mosfets are operated in class C (and and off course, the transformer losses).

 

[Easy peasy]

There are many ways to feed the power into the AC line for grid tied purposes, one does not have to use a proper inverter structure that allows reactive power flow, hysteretic current control is favoured by many as it removes the need for synchronisation, also for full bridge output one pair can be 50Hz (60Hz) and the other HF giving reduced losses.
There are so many requirements for control (e.g. test for mains present) that an ASIC would soon be outdated compared to a uP of similar cost (excluding SW).

 

[BradtheRad]

Brad; I had seen this grid tie inverter somewhere else. They claim 80% efficiency, which I find very hard to believe, as the Mosfets are operated in class C (and and off course, the transformer losses).

Running a simple simulation, I think there may also be a power factor problem.

The control signal may need to be customized as to phase angle, etc.

Or, perhaps a SPWM signal, which could improve efficiency.

 

[mtwieg]

There are so many requirements for control (e.g. test for mains present) that an ASIC would soon be outdated compared to a uP of similar cost (excluding SW).

This is the real issue right here. In principle a PFC controller could be made to operate in both directions without a lot of effort, but a grid tie inverter is much more than that. It must have many advanced features, such as anti-islanding protection. A microcontroller is basically needed for any practical design.

 

[treez]

This is the real issue right here. In principle a PFC controller could be made to operate in both directions without a lot of effort, but a grid tie inverter is much more than that. It must have many advanced features, such as anti-islanding protection. A microcontroller is basically needed for any practical design.

I wonder if a compromise could be done by having a pfc chip in reverse, and also having a micro to shut it down if the grid goes down etc etc ( I appreciate that this is needed since otherwise lineworkers may get electrocuted

 

[treez]

Is a grid tie inverter typically a very transient load, or does it behave like a constant, unchanging load to the dcdc stage that supplies it.?

 

[FvM]

What do you think? The inverter is expected to source constant real power to the grid, according either to a given setpoint or the power available from the local source, e.g. solar panel or a load operating in recuperation (e.g. a braking motor drive). In the latter case, power variations are respective more dynamical.

Grid low voltage events can enforce a sudden reduction in power delivery. Newer regulations require an inverter to source reactant power during the low voltage event (low voltage ride through capability).

 

[treez]

our GTI would be supplying power to the grid from a car battery (EV), so there would be no transients there, however, sudden grid low voltage events would be presumably seldom, and so could be dealt with by a comparator type fast shut down and delayed restart. What would make the GTI a highly transient load would be if very little power was being drawn from the grid, and so the setpoint power delivery rate had to keep rapidly changing as people turn their kettles and showers on and off.

Our bidirectional, 7kw symmetrical LLC converter (that supplies the GTI) is going to struggle with transients.

 

[FvM]

A grid tied inverter would usually not react on small voltage changes caused by local load variations. I presume, the background of your design is the bidirectional electric car battery charger, as dicussed in your previous thread https://www.edaboard.com/threads/342560/

But this device won't derive it's power setpoint from a local voltage measurement (except for the discussed specific low voltage ride through operation). It's either controlled by data communication with the utility company, or the power bilance of the company's grid cnnection.

 

[treez]

Thanks, in that case, the setpoint could change suddenly, but the output of the GTI doesn't have to change suddenly, as long as it gets to the setpoint in say 20 seconds.....so therefore, no quick transient dynamic response needed here...just a nice slow microcontroller increment/decrement feedback loop for our bidirectional, symmetrical 7kw LLC converter

 

[treez]

Can you have a Grid Tie Inverter that is an Active PFC converter aswell?...ie GTI when passing power to the mains, and acts as PFC when passing power from the mains? (Just one circuit for both functions, as you know, obviously not simultaneously)

 

[Easy peasy]

Any proper inverter can correct for lagging/leading power factor, as well as harmonics if designed/programmed to do so, however some grid tied inverters are not designed to handle reactive power and can only inject real power against the mains voltage, there is some scope for modulating this current to compensate for harmonics present, but not in a closed loop form as you would have to measure the external load currents to do this. Phase lead/lag correction requires a lot of current capability and again extra external current measurement would be needed...

 

[treez]

Thanks its very interesting indeed, did enjoy reading, great food for thought, we aren't interested in correcting the grid in the way that you describe. We are interested in a single circuit that is an active PFC in one direction, and a grid tie inverter in the other direction. Is this possible?

 

[FvM]

The questioned function is known as "active front end" (AFE) and has been often mentioned in previous discussions, e.g. https://www.edaboard.com/threads/342558/

Good you start to review state-of-the-art.

 

[Easy peasy]

As per FvM, the answer is yes.