Temperature Considerations in PV Design

[wwhitney]

No it doesn’t.

Can you clarify what you mean? If you have a current (y-axis) vs voltage (x-axis) curve, isn't the x-intercept (y=0, zero current) by definition the open circuit voltage?

Cheers, Wayne

 

[tallgirl]

Can you clarify what you mean? If you have a current (y-axis) vs voltage (x-axis) curve, isn't the x-intercept (y=0, zero current) by definition the open circuit voltage?

Cheers, Wayne

The open circuit voltage is always the voltage when there is zero current flowing outside the circuit.

PV junctions are constant voltage variable current devices where the open circuit voltage is determined by the forward voltage drop of the junction, and the current is proportional to the irradiance.

 

[ggunn]

The open circuit voltage is always the voltage when there is zero current flowing outside the circuit.

PV junctions are constant voltage variable current devices where the open circuit voltage is determined by the forward voltage drop of the junction, and the current is proportional to the irradiance.

I think what he was referring to is the family of curves in post #47 where the x axis intercepts (I = 0) are a little different for each irradiance curve.

 

[wwhitney]

The open circuit voltage is always the voltage when there is zero current flowing outside the circuit.

OK, good. So based on the changing x-intercepts on the graphs in post #47, we can see how the OCV changes with irradiance. Not a large effect, but still discernable. The OCV is about 40V at 200W/m^2 irradiance, and about 42V at 1000W/m^2.

PV junctions are constant voltage variable current devices where the open circuit voltage is determined by the forward voltage drop of the junction, and the current is proportional to the irradiance.

Hmm, judging from the curves in post #47, for a fixed irradiance, seems like they are behaving like constant current sources for voltages up to about 80% of OCV.

I assume a graph like the curves in post #47 is made (at least conceptually) by putting a variable resistor across the PV panel, and for each value of resistance, noting the voltage across the resistance and computing the resulting current. Then the resulting (V,I) pairs are plotted.

Cheers, Wayne

 

[winnie]

No it doesn’t.

Honest.

We can sit down with some PV silicon and do the experiments.

I'm always happy to learn from experiment. I could try a small 12V panel changing irradiance by pointing it in different directions. I don't have a good controlled setup, however.

 

[electrofelon]

No it doesn’t.

Honest.

We can sit down with some PV silicon and do the experiments.

Ok challenge accepted. I just went out and did this. Ive got a Trina 280 module, brand new, OCV listed as 39V. Conditions here are cloudy with light rain on and off, but no rain during this experiment, and about 70 degrees. OCV with panel laying on ground is about 34.9V. Tilt panel up to vertical and I get 33.9V. I repeated several times and confirmed the approx 1 volt spread. then tested panel inside my building but with the big doors open, got 30V. So yes large changes in voltage. Now remember that of course i terms of a PV system, those values would be multiplied by (roughly) 12 for 600V strings and 18 for 1000V strings.

 

[electrofelon]

IT would be nice if I had an irradiance meter, but I dont. I will continue to get values. I will try to get one early tomorrow b4 the sun directly hits the panels, and compare to solar noon (its going to be partly sunny tomorrow).

 

[tallgirl]

IT would be nice if I had an irradiance meter, but I dont. I will continue to get values. I will try to get one early tomorrow b4 the sun directly hits the panels, and compare to solar noon (its going to be partly sunny tomorrow).

When I was still doing PV work, we used PV silicon for irradiance sensors.

You also have to correct for dV/dT, which has a large affect between shade and full sunshine. Recall that voltage declines with increased junction temperature, and junction temperature increases with irradiance. You should be able to find the temperature coefficient for the panel, then using a contactless thermometer do the voltage correction and you should get approximately the same values.

 

[tallgirl]

OK, good. So based on the changing x-intercepts on the graphs in post #47, we can see how the OCV changes with irradiance. Not a large effect, but still discernable. The OCV is about 40V at 200W/m^2 irradiance, and about 42V at 1000W/m^2.

That doesn’t mean they aren’t a “constant voltage” power supply because the current changes far more with irradiance than the voltage does. “Constant voltage” doesn’t mean it’s highly regulated, like a lab-grade power supply. “Constant current” would mean the voltage changes from 0 to some maximum as irradiance changes from 0 to some maximum.

 

[wwhitney]

That doesn’t mean they aren’t a “constant voltage” power supply because the current changes far more with irradiance than the voltage does. “Constant voltage” doesn’t mean it’s highly regulated, like a lab-grade power supply. “Constant current” would mean the voltage changes from 0 to some maximum as irradiance changes from 0 to some maximum.

You are speaking about how the supply changes with irradiance. I was speaking about how it behaves for any given fixed irradiance.

I guess at this point I'm unclear on what statement of winnie's you are disagreeing with.

Cheers, Wayne

 

[wwhitney]

Going back to the spec sheet cited in Post #47, it says for the 400W binned panel:

Voc at STC (1000 W/m^2 irradiance, 25C cell temperature) = 41.2V
Voc at NOCT (800 W/m^2 irradiance, 43C cell temperature) = 38.8V
Temperature Coefficient of Voc = - 0.25%/C

So for an 18C temperature increase, Voc would change -4.5%, and at 1000 W/m^2 irradiance and 43C cell temperature that gives 41.2 * 95.5% = 39.3V. Which means that the change in Voc at 43C cell temperature as the irradiance goes from 1000 W/m^2 to 800 W/m^2 is 0.55V. Or at 25C cell temperature it would be 0.57V.

Cheers, Wayne

 

[ggunn]

But for most PV systems (all I have ever designed, anyway), it's a distinction without a difference. Module Vmax is Voc @ STC, corrected for lowest expected temperature

 

[wwhitney]

But for most PV systems (all I have ever designed, anyway), it's a distinction without a difference. Module Vmax is Voc @ STC, corrected for lowest expected temperature

If the spec sheet gave a Voc coefficient for irradiance (which is obviously non-zero based on posts #47 and #71), then you could take your hourly historical temperature data, and calculate for each hour the maximum clear sky irradiance (based on sun location and array orientation), and then come up with a maximum Voc for each hour. Then the largest of those Voc values would be lower than assuming maximum irradiance and minimum temperature can occur simultaneously.

That would presumably be more trouble to do by hand than it's worth in almost all cases. But would be easy enough to program into software.

Cheers, Wayne

 

[ggunn]

If the spec sheet gave a Voc coefficient for irradiance...

As my dad was fond of saying, "If a frog had wings..."

 

[electrofelon]

just did another test. Conditions are about 75% sunny, 61 degrees. under a stretch of blue sky, with the module temperature stabilized, measured 34.5 V. A small cloud passed over, and voltage dropped to about 32.5. the cloud was not very dense, it was right on the edge of being too bright to look at the sun behind that cloud. Voltage popped back to 34.5 immediately after the cloud.

 

[ggunn]

just did another test. Conditions are about 75% sunny, 61 degrees. under a stretch of blue sky, with the module temperature stabilized, measured 34.5 V. A small cloud passed over, and voltage dropped to about 32.5. the cloud was not very dense, it was right on the edge of being too bright to look at the sun behind that cloud. Voltage popped back to 34.5 immediately after the cloud.

Does the meter pass enough current to move the curve off the intercept?

 

[electrofelon]

Does the meter pass enough current to move the curve off the intercept?

I did think about that, but its just a DMM so I imagine the impedance is very high and has negligible effect.

 

[jaggedben]

I did think about that, but its just a DMM so I imagine the impedance is very high and has negligible effect.

Yeah the IV curves in the graph show a drop of two volts when amps are flowing, not milliamps.

 

[electrofelon]

Got some good values today. Conditions are about 80% sunny and the clouds are thick and fast moving, so I was able to get good solid readings of clear vs behind a cloud. Time was about an hour before solar noon. Got about 2.5-3 volts difference. I wanted to get some early measurements but I didnt get up early enough .

 

[ggunn]

Got some good values today. Conditions are about 80% sunny and the clouds are thick and fast moving, so I was able to get good solid readings of clear vs behind a cloud. Time was about an hour before solar noon. Got about 2.5-3 volts difference. I wanted to get some early measurements but I didnt get up early enough .

Isn't this all irrelevant to PV system design? I know, I know; when has relevance mattered to us?