Maximum power point in solar converter

[treez]

In solar converters , the maximum power point is simply referring to the current drawn from a solar panel, above which there is no more delivered power from the solar panel (to the load, eg battery) if the current is increased from that level.
Do you agree?

 

[asdf44]

Yes, It’s also the point where going down in current doesn’t deliver any more power either.

 

[treez]

Thanks, in that case, what we need to be able to do, is calculate the amount of power dissipated (wasted) in the solar panel itself by the current drawn from it.
..So, power = V*I…….

Well, we can measure the current…but how do we assess what the voltage is?....presumably the voltage is the nominal open circuit panel voltage minus the input voltage to the switching converter that’s downstream of it. Is this correct?

What is the open circuit panel voltage when the panel is under full sunlight?
Is this figure very different from the open circuit panel voltage when the panel is in dull daylight?

 

[D.A.(Tony)Stewart]

Although the PV is solar powered current source with a voltage limit, the MPT is when the PV impedance is matched by the load impedance for small incremental changes ranging from 82% down to 70% Voc.

It is also the Voc/Isc=Z point. Where Isc varies greatly with Solar input while Voc does not change as much.

 

[asdf44]

Thanks, in that case, what we need to be able to do, is calculate the amount of power dissipated (wasted) in the solar panel itself by the current drawn from it.
..So, power = V*I…….

Well, we can measure the current…but how do we assess what the voltage is?....presumably the voltage is the nominal open circuit panel voltage minus the input voltage to the switching converter that’s downstream of it. Is this correct?

What is the open circuit panel voltage when the panel is under full sunlight?
Is this figure very different from the open circuit panel voltage when the panel is in dull daylight?

I don't totally follow that logic and I'm not an expert but I understand that one way or another an MPPT controller would 'traverses' the load space, recording the watts delivered by the panel at different current draws (or better, perhaps the watts it's able to deliver) until it 'finds' the point where watts is highest. It then periodically has to deviate from this operating point to discover whether the peak has moved.

 

[Easy peasy]

Maximum power point tracking is just that, varying the drawn power until a peak is found and staying near the peak as the load and solar irradiance changes, a power converter (often a buck) is used to match the Z of the PV at peak power to the Z of the load...

 

[treez]

Thanks, it sounds to me that the best way to find the MPPT would be to have a buck converter with its feedback loop closed on its input current, then simply increment/decrement the current and keep doing that iteratively until the V*I product is greatest.
Then still keep doing it in case the MPPT changes.

Where "V" is the input voltage to the buck.
I am not sure if simple 8 bit PIC microcontrollers can actually do multiplication?

- - - Updated - - -

I was at a consultancy once and the boss said that generally, the MPPT is when the input to the buck is 80% of the panel o/c voltage....that if that is stuck to as a rule, then you'll always be very pretty near the MPPT......so in other words, close the loop on the input voltage to the buck...and make it 80% of o/c panel voltage....then you'll be as near as dammit to the MPPT.

 

[c_mitra]

Its complicated.

Just consider a resistive load; max power will be delivered to the load when the voltage drop (IR across the load) is half the open cell voltage.

But then we will get only 50% of the converted energy. Remaining 50% will be wasted in the panel itself. But this assumes the panel to be a voltage source with a series resistance.

But the solar panel is not a simple model and we need to have a graphical approach to select the operating point.

In your case the battery is not a simple load; it is more like a leaky capacitor that goes up in voltage with charging.

You have to optimize power transfer and the optimum point can be determined only by graphical analysis.

 

[betwixt]

I would disagree with the boss then. I have two PV systems here, both grid tied. One has a maximum output of about 1.6KW and the other about 400W and both have MPPT controllers. Aiming for about 80% voltage is probably right when the PV panels are in sunlight but when not fully insolated, for example in mornings, evenings and overcast days, the figure could be wildly different.

"perturb and observe" is the best method, actually modulating the load current to see how the power changes. In other words check the voltage at different load currents, multiply them and find the highest value. Do it in a loop so it is dynamically reactive to different loads and light conditions.

And yes, simple PIC are perfectly adequate. My small controller uses an 18F452, I've never opened the big controller but I imagine it is something similar.

Brian.

 

[D.A.(Tony)Stewart]

If you had a PD to detect ambient intensity, you could open loop get fairly close from 82%Voc +/-? bright to 72%Voc, at 10% Pmax

 

[FvM]

Two comments after lately reviewing this thread.

As far as I see, there are only two useful methods to perform MPPT, either perturb and observe or using a previously recorded set of curves that "knows" the MPPT for each irradiance level. Obviously perturb and observe has the advantage of not needing any prerequisites. A more sophisticated method would be an estimator that is continuously updated according to observed IV values and starts perturbation if required.

The previously mentioned impedance matching can be seen as an analogy, but doesn't consider the non-linear photo element IV characteristic. Respectively the MPP isn't at 50% open circuit voltage as it would be for a linear source impedance.

 

[kathmandu]

The amount of irradiance isn't everything about the PV output power. The ambient (panel) temperature it's a big factor, too.

Actually, I get more power from my off-grid PV system in the winter than in the summer time.

If you look at a PV datasheet, you'll see that the derating factor it's quite important. Well, we're talking about few (maybe ten) percent of power loss but the most important thing is the reduced maximum power voltage.

That's it, if you're doing your math taking into account 80% of the Voc (standard condition) you'll get a semnificative error.

Using the P&O method it's the best way to overcome this situation.

 

[kathmandu]

LE: "significant error".. lost in translation.

 

[D.A.(Tony)Stewart]

MPPT is all about dynamic impedance matching load to non-linear current source. However you do it, the optimum point is this result, which is consistent with max power theorem. Since PV’s have a negative impedance being sources and with a common PTC effect on the magnitude, we expect it to perform better in cooler weather.

So many newbies think they can put any old DCDC converter on a PV or direct to a battery and get good efficiency.

I modelled how the PV changes impedance with a photo diode to determine MPPT of V vs I and duplicated this with a nonlinear PD controlled Transistors to match the impedance between a battery or a large supercap and a PV source. I used Falstad sim but lost the model. It worked really well with no need for tracker but used precise Ri or ESR curves for each part

 

[Lutech]

Maximum power point (MPP) The point on a power (I-V) curve that has the highest value of the product of its corresponding voltage and current, or the highest power output. maximum power point tracker (MPPT) A device that continually finds the MPP of a solar panel or array.

 

[D.A.(Tony)Stewart]

Yes, of course, Max-Power-Point is max V*I product as a function of solar input, but few realize it is also where the interface equivalent impedance is conjugate-matched, no matter how it is regulated as long as it is. ( there are many ways )

Zs= ΔV/ ΔI is lowest or other words Ys= ΔI/ΔV has the fastest change in slope of the RED power curve corresponds to the peak of the BLUE curve.

You can model an active conjugate matched load using a linear PD sensor controlled LDO using a circuit with a Zener Op Amp and Nch FET so it cannot produce a load impedance lower than the source impedance.

Or in simple words demand only as much power as it can supply according to linear PD's (photodiodes or perhaps Light sensors ( never photoresistors... too innaccurate)

**broken link removed**

It is easily possible to generate an curve for the charger for the variation in battery State of Charge and other dynamic loads.

Thus one can design a simple light controlled MPT system to limit the demand current using this data and resulting models using inexpensive secondary solar sensors (PD's) to get an average.

Below the MPT load is an jagged curve but is actually a smooth quadratic curve just like zener diodes so it can be modelled with these with suitable gain and offset.

The tangent or maximum slope also reduces with sunlight but you can plot that in Excel for any PV array with a swept active load test with a at different sunlight levels.**broken link removed**