Control chips for 50Hz Mains inverters?

[cupoftea]

Hi,
Why do we not get Controller chips for 50Hz Mains inverters?
I mean, surely if we can have controller chips for PFC, then we can have them for inverters?

Isn’t all that is needed is to sense the rectified and divided down Vout, and then compare it to a sinusoidal reference, via an error amplifier, and then control it like that, so that it gives a sinusoidal output voltage regardless of loading? (when i say rectified, i mean only the bit fed to the controller is rectified (for obvious reasons) , as you know, the output is AC voltage)

 

[Easy peasy]

always best to design your own ...

 

[cupoftea]

Hi,
Would you agree that a single phase mains inverter (sinusoidal output) is an extremely simple bit of hardware to design?...

Lets face it, it connects to the grid voltage, so that essentially defines its output voltage.
Surely all that is needed, is simply to have a DC bus which is higher in voltage than mains peak, and then you simply use sinusoidal PWM and an H bridge, to shovel an AC voltage out towards the grid?……there will obviously be an inductor between the inverter output and the grid wires…and that will limit the di/dt of the current going into the mains. Also, obviously current limiting would be needed.
What is the most complicated part of it? Its difficult to see where the feedback loop is needed…..there is obviously no control over the output current….just output current limit…..because it’s the load seen by the grid which determines what current is pulled from the inverter?
The bit of control that I can see, is that at any instant in time, the inverter’s output must be higher than the grid voltage, otherwise current wont be able to flow out of the inverter. (when I say “inverter’s output voltage, I mean the voltage on the non-grid side of the inverter’s output inductor.)

Hi again...
Can you buy single phase Inverter power modules for converting DC to UK/EU/USA/AUS mains AC?...
Eg like you can buy Vicor modules for DCDC, can you buy the same kind of thing, but an inverter?

 

[FvM]

Control principle of GTI is sinusoidal current source in phase with grid voltage. In practical designs, a voltage feedforward path helps to reduce current distortion.

Several protection circuits (overcurrent, over- and undervoltage, anti-islanding) are required for working GTI, also IEEE 1547 compliance.

All-in-all, I won't consider it extremely simple bit of hardware.

 

[D.A.(Tony)Stewart]

400-kW Grid-Connected PV Farm - MATLAB & Simulink

This example shows an average model of a small PV farm (400 kW) connected to a 25-kV grid using two-stage converter.

www.mathworks.com

Single-Phase, 240 Vrms, 3500 W Transformerless Grid-Connected PV Array - MATLAB & Simulink

This example shows the operation of a typical transformerless photovoltaic (PV) residential system connected to the electrical utility grid.

www.mathworks.com

 

[cupoftea]

Control principle of GTI is sinusoidal current source in phase with grid voltage. In practical designs, a voltage feedforward path helps to reduce current distortion

Thanks, but isnt it the loads on a network that determine the power factor.?.....i mean, as you know, if its a non-PFC'd load, then its not going to draw sinusoidal current just because there's a GTI connected in somewhere?.....so i am kind of investigatively saying, why would the non-PFC'd load draw a sinusoidal current just because the GTI has the feature of being a "sinusoidal current source in phase with grid voltage"

In fact, whether or not a GTI is connected in, the load will draw what it draws...in whatever waveshape...as long as the GTI doesnt distort the mains voltage?

 

[FvM]

It's the grid impedance that's the GTI sees, local loads are only a small part of it. As long as the grid impedance for fundamental frequency is low enough, it's always possible to inject a pure sinusoidal current, even if the voltage is distorted.

Low current distortion is a regulation requirement for GTI:

(Quoted from R. Teodorescu, M. Liserre and P. Rodríguez, Grid Converters for Photovoltaic and Wind Power Systems, Wiley 2011)

 

[cupoftea]

Thanks, so basically, a GTI is pretty much like a PFC stage, but putting the sinusoidal current into the mains instead?

And if a GTI is basically a PFC stage in reverse, then why is it said to be so complicated?

I must admit, it does appear to raise questions...for example, if the power stations arent set up to try and push sinusoidal current into the mains, then why are the GTI's having to?
I mean, power stations simply put a sinusoidal voltage onto the grid wires.

 

[FvM]

An inverter can be operated as active compensator to absorb harmonic currents of non-linear loads. This is particularly useful if a plant is billed for harmonic currents pushed to the the grid, or was imposed with a harmonic current limit value. This function is however beyond the scope of IEEE 1547 because it has no benefit for normal GTI users, just causes additional inverter losses and in so far reduces the solar power efficiency. Technically you need to relate harmonic currents to voltage harmonics to determine if the inverter is absorbing (wanted) or generating (unwanted) grid distortion.

 

[BradtheRad]

sinusoidal PWM and an H bridge

Somehow the controller must detect the amount of power coming from the PV panels. Its aim is to maintain the DC bus at 350 or 400 V through each entire 20mSec cycle. It must pace the amount of current it sends out. It must avoid sending too much current during the first few mSec of a cycle, lest the DC bus is unable to sustain for the remainder of the cycle.

Hence the controller needs to be smart enough to initially send very short PWM pulses, then lengthen the duty cycle gradually, until it senses DC bus voltage starting to droop.

It must monitor energy transfer continually, because if clouds move in, less power is available from the PV panels. Then the controller needs to be smart enough to shorten the PWM pulses.

 

[cupoftea]

It must monitor energy transfer continually, because if clouds move in, less power is available from the PV panels. Then the controller needs to be smart enough to shorten the PWM pulses.

Thanks, yes...that would, as you know, be the job of the converter which connects to the panels (MPPT)..........the output of that converter would go to another converter which would produce the 400VDC.

Also, reference GTI's......they have no current control...they cant, because the current drawn from them depends on the loading of the phase........they have no voltage control...the voltage at their output is determined by the mains.
The GTI's only have current and voltage limiting.

The attached LTspice sim shows an open loop GTI doing a pretty good job with no control whatsoever......i admit that a little bit of extra control is needed to get that current sinusoid and in phase with the voltage...but as you can see.....the jobs almost done here, and thats open loop!

This is why i say the GTI is nowhere near as difficult as its made out.

 

[KlausST]

Hi,

Also, reference GTI's......they have no current control...they cant, because the current drawn from them depends on the loading of the phase.

The current is determined by the loading:
This should be true for "non_grid_tied_inverters".
For GTIs I expect they need current control, not only for overall power control, but at least for wave shape (sine) control.

As said: not knowledge, just my expectation.

Klaus

 

[FvM]

Which "reference GTIs" are you referring to? As you guess correctly, your simple example can't work in real live without additional control loops. Presently it depends on internal timing "magically" phase locked to the grid and a grid voltage adjusted empirically to fit the modulated 400V. Not perfectly, resulting in PF of 0.78 according to LTSpice fourier transform:

Presuming a phase-locked reference generator, the control has two degrees of freedom, current phase and magnitude. Phase will be zeroed for maximal PF and current has to be controlled according to the intended power delivery, usually the generated PV output.

That's all well manageable, just a little bit more complex than a PFC rectifier. IEEE 1547 protection functions need to be added. No problem to put the control into a state-of-the-art microcontroller , e.g. STM32.

--- Updated Oct 4, 2021 ---

The current is determined by the loading:
This should be true for "non_grid_tied_inverters".
For GTIs I expect they need current control, not only for overall power control, but at least for wave shape (sine) control.

As said: not knowledge, just my expectation.

Absolutely correct. Similar to a PFC rectifier controller, the overall power control can be established through DC link voltage control. But due to the inherent 100 Hz DC ripple of a single phase inverter, DC link voltage control must be slow, otherwise it would generate current harmonics. Respectively a feedforward path from the power source will achieve faster reaction on varying power delivery.

 

[cupoftea]

Thanks, Actually i think im incorrect...the following is the very thorough GTI design guide that i got recomended 5 years back...so there is actually loads of stuff out there on the web on how to design these things....the TI one even writes the control software for you...
https://www.ti.com/lit/ug/tidub21d/tidub21d.pdf?ts=1633332343091&ref_url=https%3A%2F%2Fwww.google.com%2F

 

[Easy peasy]

Many GTI's are current control from a ref sine wave, that goes from zero to full ( = 32A pk or whatever ) these GTI's need mains present to shove current into - and were/are never designed to be stand alone ( they can't )

More modern ones with more control allow a stand alone mode ( but not on the mains obviously )

 

[BradtheRad]

H-bridge with 120VAC mains in roles both as load and bias control. It works in simulation although it's hard to say if this design is practical. Power is wasted through several components. Perhaps sine PWM can be added.

The scope traces show mains current in opposite phase to mains voltage, which indicates 'my' power system sending current onto the grid.

 

[cupoftea]

Many GTI's are current control from a ref sine wave, that goes from zero to full ( = 32A pk or whatever ) these GTI's need mains present to shove current into - and were/are never designed to be stand alone ( they can't )

Thanks thats interesting. Supposing we were to produce a sinusoidal AC current in phase with the mains voltage, then that goes a long way to being a grid tied inverter i believe? (excepting all the protections and shutdowns etc). Reading the following makes you wonder what the overall control strategy is...

https://www.ti.com/lit/ug/tiduay6e/tiduay6e.pdf?ts=1633885097383

...its hard ot find reference to a reference sine....which did appear odd to me.

 

[cupoftea]

Many GTI's are current control from a ref sine wave, that goes from zero to full ( = 32A pk or whatever ) these GTI's need mains present to shove current into - and were/are never designed to be stand alone ( they can't )

...thanks for this, i see what you mean, the mains is a very stiff source to drive into......very low impedance....so really, it makes it seem like a GTI is actually an easier job than a non-GTI?...since with a GTI, you can just set your error amplifier up to provide a sinusoidal source, and shovel it out into the mains?

One particular difference i see between GTI's and non-GTI's , is that GTI's require an output LCL filter, but non-GTI's require a LC output filter. Would you agree?

 

[cupoftea]

Do you think the attached would be OK to take forward as a Grid Tied inverter?
It puts a sinusoidal current into the mains. Is mostly analog...would just need some software to generate the sinusoidal reference and do all the protections needed for a GTI.
(LTspice and pdf schem attached)