Technical question regarding string inverters and watts per string

[bcroe]

Bruce, thanks for your comments...I've always enjoyed reading your posts here as you give super valuable information and insight. So do you think my inverter is clipping at 6001 watts or am I actually getting whatever it's reading (looks like 6150 is about the max) and I'm actually over 6150 at midday?

I think you're right that the whole system is hanging around over 5000 watts for a lot longer during the day since the extra panel went on. So perhaps I did get a 3 panel bump when I added one more panel to change the 3 strings from 9-9-8 to 9-9-9 ??

Your Fronius is doing what mine do, go 100W or a bit more output over the rating in clipping mode.
I would be inclined to believe the readout to nearly 2 decimal places, and nearly 3 incrementally.
Remember the output will be running around 96% of the input.

A check of the DC operating voltage would give a strong indication of 8 vs 9 panel string operation,
other factors similar and not in clipping. Bruce Roe

[MARKSDCA]

Since you are into clipping, the noontime improvement may be even greater than you can
measure. But the gain will carry into all day. Bruce Roe

Bruce, thanks for your comments...I've always enjoyed reading your posts here as you give super valuable information and insight. So do you think my inverter is clipping at 6001 watts or am I actually getting whatever it's reading (looks like 6150 is about the max) and I'm actually over 6150 at midday?

I think you're right that the whole system is hanging around over 5000 watts for a lot longer during the day since the extra panel went on. So perhaps I did get a 3 panel bump when I added one more panel to change the 3 strings from 9-9-8 to 9-9-9 ??

[bcroe]

Today things seemed to jump nicely over 6000 for most of the time between noon and 12:30 (see pic)

Not sure if this is due to atmospheric conditions or what.....[ATTACH]7838[/ATTACH]

Since you are into clipping, the noontime improvement may be even greater than you can
measure. But the gain will carry into all day. Bruce Roe

[MARKSDCA]

Today things seemed to jump nicely over 6000 for most of the time between noon and 12:30 (see pic)

Not sure if this is due to atmospheric conditions or what.....20151003_121115XX.jpg

[sensij]

I have an account but I have not input my data.

If your inverter has data push capability, you can set it up to upload the data automatically. With your model of inverter, that might require an add-on communication card though.

http://pvoutput.org/help.html#push-froniuspost

[MARKSDCA]

Are you on PVOutput.org? This is an interesting case, and I would enjoy reviewing the data.

I have an account but I have not input my data.

[sensij]

Are you on PVOutput.org? This is an interesting case, and I would enjoy reviewing the data.

[MARKSDCA]

Right now I would say that the watts I'm getting today reflect maybe a "one panel" increase.....I'm looking at 5800-5850 peak between noon and 1. I was truly hoping for an average over 6000 watts.

But it seems like the system was generating bigger wattage earlier in the day.....so without a software monitoring program I would say the jury is out. All I can do is look at total daily production and decide.....and this morning we had broken clouds for several hours in the AM so I will keep my eye on it.

20151002_121158XX.jpg

[inetdog]

Ok, let's look at it analytically. Start by assuming the equation for the IV curve can be approximated by the equations described here. Let's use panel data from this data sheet.

Also, we know that the voltage of the 8 panel string will be equal to the voltage of the 9 panel strings, so the voltage of each panel in the short string will be 1.125 * the voltage of a panel in the longer string.

A good MPPT will find the voltage that maximizes the power throughput. For example, assuming 800 W irradiance and the NOCT of 46 dec C, we know from the data sheet the maximum power point for a panel in these conditions is at V = 28.3 V and I = 7.12 A, giving a power of 201.3 W per panel. If we bump up the voltage by 12.5%, that is 31.8 V, and by IV calculation, that results in a current of 4.763 A, for a power of 151.5 W. If this case were true, it suggests that you would get [(18 * 201.3) + (8 * 151.5)] = 4835 W from the array.

Compared to the maximum possible power of 5234 W, if all 26 panels were at 201.3 W, that is an impairment of almost 8% by the short string, and as others speculated in the thread, equivalent to the maximum power of only 24 panels, instead of the 26 actually installed.

But wait! The MPPT shouldn't settle on whatever voltage maximizes the 9 panel strings, at the expense of severe underperformance in the 8 panel string. Instead, it will adjust the voltage until it maximizes the power output of the entire array. In this model, maximum power occurs at a string voltage of 237.6 V, with the panel voltage in the long strings at 26.4 V, and in the short string at 29.7 V. The current flowing in the long string is 7.430 A, giving a panel power of 196.2 W, while the current flowing in the shorter string is 6.613 A, giving a panel output of 196.4 W. The total combined array power at this maximum is 5105 W, only about 2.5% less than the 5234 W maximum. (It turns out the varying temperature or irradiance doesn't affect this much... the 2.5% is pretty consistent)

So... from this model, I would guess that adding the 9th panel to the short string would increase output by 3.8% (1/26th increase in array size) + 2.5% (by allowing a more optimal MPPT), or 6.4% total improvement. This is a bit more than I guessed above, but still close enough to the PVWatts data to be believable. I hope you let us know how it goes!

There are differences in MPPT algorithms as impemented between brands and maybe even models.
Some will scan the current and voltage over a wide range and find the highest of several local maxima.
Others will start essentially at Voc and scan down in voltage until they reach a maximum. They will not scan past that point to find what may be a higher maximum power at a lower voltage.
It is also theoretically possible for the MPPT scan to start at zero volts and scan up until it finds a local maximum (the 8 string point) and stop there, but I do not know of any MPPT devices that work that way.
Some may not "scan" at all but simply go a small amount in either direction from their current operating point and see whether the power increases. Those will end up moving bit by bit to the closest maximum point in he upward trending direction.

And of course some just sense Voc (by cutting off current flow momentarily) and the set the operating voltage to 80% of that. That would find the 9 string point only.

When SMA introduced their new highly touted super optimization (Opti-track?) to deal with partial shade, etc., all they seem to have done is change from a scan to first maximum algorithm to a wider range scan and selection of the highest peak.

[MARKSDCA]

Ok, let's look at it analytically. Start by assuming the equation for the IV curve can be approximated by the equations described here. Let's use panel data from this data sheet.

Also, we know that the voltage of the 8 panel string will be equal to the voltage of the 9 panel strings, so the voltage of each panel in the short string will be 1.125 * the voltage of a panel in the longer string.

A good MPPT will find the voltage that maximizes the power throughput. For example, assuming 800 W irradiance and the NOCT of 46 dec C, we know from the data sheet the maximum power point for a panel in these conditions is at V = 28.3 V and I = 7.12 A, giving a power of 201.3 W per panel. If we bump up the voltage by 12.5%, that is 31.8 V, and by IV calculation, that results in a current of 4.763 A, for a power of 151.5 W. If this case were true, it suggests that you would get [(18 * 201.3) + (8 * 151.5)] = 4835 W from the array.

Compared to the maximum possible power of 5234 W, if all 26 panels were at 201.3 W, that is an impairment of almost 8% by the short string, and as others speculated in the thread, equivalent to the maximum power of only 24 panels, instead of the 26 actually installed.

But wait! The MPPT shouldn't settle on whatever voltage maximizes the 9 panel strings, at the expense of severe underperformance in the 8 panel string. Instead, it will adjust the voltage until it maximizes the power output of the entire array. In this model, maximum power occurs at a string voltage of 237.6 V, with the panel voltage in the long strings at 26.4 V, and in the short string at 29.7 V. The current flowing in the long string is 7.430 A, giving a panel power of 196.2 W, while the current flowing in the shorter string is 6.613 A, giving a panel output of 196.4 W. The total combined array power at this maximum is 5105 W, only about 2.5% less than the 5234 W maximum. (It turns out the varying temperature or irradiance doesn't affect this much... the 2.5% is pretty consistent)

So... from this model, I would guess that adding the 9th panel to the short string would increase output by 3.8% (1/26th increase in array size) + 2.5% (by allowing a more optimal MPPT), or 6.4% total improvement. This is a bit more than I guessed above, but still close enough to the PVWatts data to be believable. I hope you let us know how it goes!

Wow! Fun with SCIENCE!! Thanks for the analysis!!

As of 8:49am I was hovering around 4200 watts before the scattered clouds rolled in. That's a pretty high number for this early in the day. Fingers crossed. More in a couple of hours.

[sensij]

Well that contradicts what Fronius Tech Support is telling me. I hope you're wrong!!

Ok, let's look at it analytically. Start by assuming the equation for the IV curve can be approximated by the equations described here. Let's use panel data from this data sheet.

Also, we know that the voltage of the 8 panel string will be equal to the voltage of the 9 panel strings, so the voltage of each panel in the short string will be 1.125 * the voltage of a panel in the longer string.

A good MPPT will find the voltage that maximizes the power throughput. For example, assuming 800 W irradiance and the NOCT of 46 dec C, we know from the data sheet the maximum power point for a panel in these conditions is at V = 28.3 V and I = 7.12 A, giving a power of 201.3 W per panel. If we bump up the voltage by 12.5%, that is 31.8 V, and by IV calculation, that results in a current of 4.763 A, for a power of 151.5 W. If this case were true, it suggests that you would get [(18 * 201.3) + (8 * 151.5)] = 4835 W from the array.

Compared to the maximum possible power of 5234 W, if all 26 panels were at 201.3 W, that is an impairment of almost 8% by the short string, and as others speculated in the thread, equivalent to the maximum power of only 24 panels, instead of the 26 actually installed.

But wait! The MPPT shouldn't settle on whatever voltage maximizes the 9 panel strings, at the expense of severe underperformance in the 8 panel string. Instead, it will adjust the voltage until it maximizes the power output of the entire array. In this model, maximum power occurs at a string voltage of 237.6 V, with the panel voltage in the long strings at 26.4 V, and in the short string at 29.7 V. The current flowing in the long string is 7.430 A, giving a panel power of 196.2 W, while the current flowing in the shorter string is 6.613 A, giving a panel output of 196.4 W. The total combined array power at this maximum is 5105 W, only about 2.5% less than the 5234 W maximum. (It turns out the varying temperature or irradiance doesn't affect this much... the 2.5% is pretty consistent)

So... from this model, I would guess that adding the 9th panel to the short string would increase output by 3.8% (1/26th increase in array size) + 2.5% (by allowing a more optimal MPPT), or 6.4% total improvement. This is a bit more than I guessed above, but still close enough to the PVWatts data to be believable. I hope you let us know how it goes!

[MARKSDCA]

OP needs to figure out how exactly the array was wired and add the additional panel onto the string of 8.

Done! But it was done at sunset. Results tomorrow!!

[MARKSDCA]

If your array is 20 deg tilt south facing azimuth, I model with PVWatts around 5500-5750 watt peak this time of year. It sounds like you are getting that, so the penalty for having a shorter string in your case is probably not much. I wouldn't expect to see much more than a 4% increase.

Well that contradicts what Fronius Tech Support is telling me. I hope you're wrong!!

[sensij]

I normally get 5500-5700 watts at peak hours so hoping to be able to report a nice jump in that number tomorrow. Assuming we have good San Diego sunshine at mid-day!!

If your array is 20 deg tilt south facing azimuth, I model with PVWatts around 5500-5750 watt peak this time of year. It sounds like you are getting that, so the penalty for having a shorter string in your case is probably not much. I wouldn't expect to see much more than a 4% increase.

[MARKSDCA]

UPDATE: My installer is coming in a couple of hours to turn my 9-9-8 array into 9-9-9. It will be interesting to see if the Fronius will actually give me a 3 panel boost after adding one panel. (the argument being that right now it thinks there are 3 strings of 8 rather than 9-9-8)

I normally get 5500-5700 watts at peak hours so hoping to be able to report a nice jump in that number tomorrow. Assuming we have good San Diego sunshine at mid-day!!