Electrical – how to design a microgrid

Your inverter can't handle more than 15V, but the panel may put out a higher voltage in bright sunlight. Therefore you must find a way to drop the voltage. You want a cheap solution, but does it have to be lossless?

Since you are using the power directly from the solar panel with no storage, it doesn't matter if dropping the voltage results in a power loss (since you won't be using that extra power anyway). What does matter is that the regulator doesn't introduce extra loss when the panel is producing the minimum voltage required to power the inverter.

A series or shunt linear regulator that simply wastes the excess power produced by the panel could still have 100% efficiency when it is really needed. The obvious solution is to simply put a high power 15V Zener diode across the panel, but how much power will it have to handle and can it be done 'inexpensively'?

Here's the IV curves for a typical '12V' 130W solar panel. First thing to note is that it puts out 22V open circuit (at 25°c) and about 17V at the maximum power point.

Your panel might actually put out 12V at maximum power and 17V open circuit, though that would be an unusual spec for a '12V' panel. But let's assume it is correct and your panel produces an IV curve similar to the green line on the graph.

Your panel can produce 400W in full sunlight, but your inverter is only drawing about 190W (150W at 80% efficiency). So the Zener will have to absorb about half the power produced by the panel, ie. ~200W. That's going need several large transistors on big heatsinks.

The other alternative is a series regulator. This drops the excess voltage at the current being drawn by the inverter (~190W/15V = 13A). A good low resistance FET could pass this current with virtually no loss at lower panel voltage, and only has to dissipate about 26W in full sunlight (assuming the panel puts out 17V at 13A). This circuit should be considerably cheaper to make than the 200W shunt regulator.

Or you could just use an off-the-shelf switching regulator designed to deliver 12V at 16A or higher. This may have a slightly higher minimum voltage drop, but will the difference be significant?

The panel may only have to produce a fraction of a volt more to compensate for voltage drop in the regulator. But light intensity and temperature variations have a much greater effect on panel output. The difference between the panel producing just enough power to run the inverter and not enough power is so small that it will hardly be noticed, so 99% of the time the extra loss in the regulator will be nothing to worry about.

(1) The inverter output will be exactly the line voltage, having raised the (local) line voltage according to the source impedance of the supply line and the current developed As the inverter is connected to the line, its output is the (locally measured) line voltage by definition. The "meter spins backwards" when generation exceeds use, because it senses the current direction. Often there would be separate meters for consumption and generation.

^{simulate this circuit – Schematic created using CircuitLab}

This is a residential distribution network, your house is No.2, the third from the pole transformer and the only one with solar power. Currently it's in shadow, so the line voltage reduces slightly further from the pole, and the last house (No.3) sees 109V.

When the sun comes out, the line voltage at the pole remains 115V, but your local line voltage also becomes 115V.

No. 1 is now fed from both sides and it's line will rise from 113V to 114V, and No.3 is still 2V below your line voltage, now seeing 113V.

Weather Line No.1 No.2 No.3
Cloudy 115V 113V 111V 109V
Sunny 115V 114V 115V 113V

(2) Where are you and what connection regime are you in? If you're in the UK and conforming to the G83/2 standard for a single phase residential connection, the answer is up to 16A. (Higher output is allowed on a separate standard, G59/2 for 3 phase connections). Any excess power is simply not drawn from the panel (assuming this is a grid tie system with no local storage)

None of this applies outside the UK.

(3) Pretty much nothing visible, maybe a very slightly higher voltage and marginally cleaner waveshape (assuming your inverter is legal and conforms to the applicable standards, including harmonic distortion.

The G83/2 specification illustrates the requirements for protection against overvoltage, out of tolerance frequency and waveform shape - and the all-important consideration of disconnecting immediately if the grid fails, to prevent electrocuting the engineers trying to repair it!

Here's an article on the rules in California to get you started on the US standards - which may well vary from State to State.

Grid regulation is a topic in itself - it isn't automatic, someone monitors it and adds or removes generation capacity to keep its parameters (voltage and frequency) within limits.

One characteristic of traditional (spinning metal) generators is that as you load them, they slow down a little, which reduces their output voltage and their contribution to the grid - transferring their load to others, which slow down in turn - arriving at a concensus on the actual mains frequency. You can watch this process in the UK in real time here. At the moment, it's reading 50.007Hz, so there's no need for additional capacity, but if it falls to 49.9Hz, phone calls will be made and some other power source will be turned on...

The importance of this for grid tie inverters is the way they can affect demand and power flows in the short term in unpredictable ways, taht the grid isn't currently designed for.