limiting a grid-tie inverter

[timh2870]

Hi everyone! First time poster here. I'm trying to regulate a grid tie inverter for use with a battery, so as to run the house without feeding back to the grid. First off, I will not simply buy one of those Chinese Sun units, as they do not meet UL1741. I am currently working with a Craigslist find used Sunny Boy swr1800u. The current setup, which is for testing purposes only, is this; I have two car batteries running a 900va modified sine ups, which has a bridge voltage doubler plugged into it producing about 360 Volts DC. I then use a simple 555 pwm circuit driving two 600v, 6A mosfets in parallel, switching at just over 1khz, Minimum duty cycle is 9%. My problem is this; the mppt algorithm in the Sunny Boy will keep searching for more power, even at minimum duty cycle, the current keeps climbing, and eventually the 900va inverter calls it quits. The thing is, with a variac between the ups and the bridge rectifier, i can regulate the power output easily. Am I missing something fundamental here, or do I just need a tighter control loop ( take myself and the potentiometer out of the equation). I was going to build the final circuit around a tl494/ka7500. Ideas? Suggestions?

[Zero999]

You could design the DC-DC converter to limit the current and thus the power to a set point, with a maximum voltage too for safety purposes.

This isn't the correct way to do it.  You're using the wrong kind of inverter. A grid tie inverter designed to run off solar panels produces an output current, proportional to the input current generated by the solar panels. The output voltage just follows the grid. It won't work without an existing mains voltage to lock onto. Get a non-grid tie, pure sine wave inverter designed to run off batteries, not a solar panel.

[capt bullshot]

So, as far as I understood your setup, you use the MOSFETs to supply power from the rectified DC to the inverter. There's no inductor or whatever in between the MOSFETs and the inverters input.
This approach won't work anyway, since the inverter has a substantial input capacitor that just gets charged while your MOSFETs are turned on, and supplies power for the off period. So the PWM is just pointless here, only puts power losses on the MOSFETs.

What you need is some kind of current limiter to limit the power to the inverter. To make your setup work, you'll have to emulate the typical solar panels behaviour at various levels of sunlight.
Hard limiting voltage with a variable current limit may work, but may also not work, depending on the MPPT algorithm of the inverter. In theory you can limit the inverters power by limiting the current supplied to it, the MPPT will search for maximum power and hit the voltage limit. Depending on the MPPT, a hard voltage limit (as seen on your typical lab power supply) may disturb the MPPT, causing kind of oscillations or seeking cycles to occur. Better use some kind of soft limit, google for a "Solar Panel Simulator".
Next step would be to apply this solar panel simulation to your source that is  feeding the inverter.

Maybe you could contact a SMA representative and ask for the software and codes to change your inverters configuration to make it easier to fit your needs. These things often have a variety of parameters that one can tune, but some of them are secured by codes to keep the average consumer from fiddling with it. Maybe a parameter exists that allows you to set the max. output power of the device online through the inverters data interface.

[fourtytwo42]

The thing is, with a variac between the ups and the bridge rectifier, i can regulate the power output easily.

That is because the correct thing to do is to regulate the VOLTAGE NOT THE CURRENT! You must of course maintain the input voltage of the GTI above its cut-off then as you allow it to rise it will draw more current so as you can see regulating it's input voltage is a simple way of regulating the power it will use. So your PWM needs a simple output voltage feed back loop with the set-point controlled by your power control loop error. Thats How I do it and it works.

[Zero999]

The thing is, with a variac between the ups and the bridge rectifier, i can regulate the power output easily.

That is because the correct thing to do is to regulate the VOLTAGE NOT THE CURRENT! You must of course maintain the input voltage of the GTI above its cut-off then as you allow it to rise it will draw more current so as you can see regulating it's input voltage is a simple way of regulating the power it will use. So your PWM needs a simple output voltage feed back loop with the set-point controlled by your power control loop error. Thats How I do it and it works.

How do you strop the MPPT controller from trying to suck more current until it reaches its maximum power rating? A MPPT inverter connected to a constant voltage source will do exactly that. The maximum power point of a constant voltage source, is the point when the load impedance equals the impedance of the source, which will be very low for a battery or DC:DC converter, so the MPPT controller will simply reduce its impedance to the minimum possible without overheating.
https://en.wikipedia.org/wiki/Maximum_power_point_tracking

A solar panel behaves like lots of current sources, each with a massive diode connected in parallel,  in series. The whole thing can be roughly modelled as one current source will a string of diodes in parallel. The maximum peak power point is when the open circuit voltage is low enough so the diodes don't conduct, but not too low, otherwise the voltage across the load will drop, leading to less power draw.

Then there's the issue of operating the grid tie inverter off grid. If it doesn't have an island mode, then I don't see how this can be done. Another inverter would be needed to give it a voltage and frequency to lock on to and there would need to be something to dissipate any overproduction, otherwise the inverter will simply shut down.

[timh2870]

The whole point of this is to stay on the grid. The average load in my house is around 500 watts, but can peak over 13kw with the dryer, stove, toaster, and coffee pot on. an 1800 watt inverter could cover the whole load of the house about 95% of the time. If i wanted to go off grid I would need a split phase inverter of at least 8kw, and have to teach the wife how to live on a power budget. such an inverter would have an idle power draw over 100 watts and would spend most of its time under utilized, and wasting battery power.

Now back to the Sunny Boy. It has a lower cut off somewhere around 150 volts, and full wave rectified modified sine is somewhere around 140, so i don't think that inverter would work for testing purposes. I have another dc-dc that can do 180 volts, but it isn't ready yet. I am fully aware that there are input capacitors in the Sunny Boy, hence why I don't have any on the fets. I'm going to try adding an inductor on the high side to make a sort of lc filter. If the input caps are 1000uF and I can find a 220uH choke that "should" be 3db down at 330hz, making the input look more like a current source...hopefully. Will post results this evening.

[Zero999]

Oh I misunderstood. I thought you mean you wanted to go off grid, when you said you didn't want to feed back into the grid. You'll need to measure the power going into the house to make sure it isn't back-feeding into the grid.

Why have a DC:DC converter? Can't you use enough solar panels and batteries in series that it becomes unnecessary?

[timh2870]

That's exactly how I have it set up. Solar panels connect to charge controllers (el-cheapo aliexpress units) which charge my 7s lithium ion battery. The battery is currently 140p salvaged cells and about 6.3 kwhr, with enough cells tested and sorted to bring me up to 12ish kwhr and 240p. If i ran a higher battery voltage (above about 60 volts or 16s) i would need a different class of charge controller, different bms, etc. keeping things at 28 volts keeps it simple, as charge controllers, ac chargers, inverters, battery managers, etc. are all easy to find and inexpensive. I need to boost the voltage because this is the inverter i have, and I don't plan on buying more. I tried the enphase m190, but had bad luck with a pile of used units. they need at least 32 volts to start up, but will run on 22 - they're just painfully slow to boot and ramp up to full power.

I currently have a pair of CTs on the incoming power lines, and watch the current waveforms on the scope. As long as A-B=0, my non-net-metering power meter reads zero. I plan on using an arduino to control the limiter, but haven't gotten to that part yet. I suppose the ultimate goal is to produce something that can take an off the shelf, surplus, used, etc grid tie inverter and make it usable for someone else in my situation.

[timh2870]

Success! Adding an inductor from the parts bin (16awg, ferrite core with about 450uH) did the trick. Touchy, but I can regulate down to about 100 watts and keep it there or go up to about 700 when the ups starts to complain, then dial it back. About 70 volts dropped on the inductor at 600 watts.  Thanks everyone for your help. Now I just need to work on better control circuitry and a more powerful boost converter.