Micro Inverters on Ground Mounted Array

Status
Not open for further replies.
I am far from an expert on this topic, researching it myself for a potential DIY.

The question you need to ask is how much differential shading the system will experience. If all the panels are experiencing the same sun exposure than maximum power tracking at the string level is cheaper and more efficient.

But if some panels will be in the shade and some in full sun for significant portions of the day, then panel level optimization makes sense.

Jon
 
Typically ground mounts are sited to not be shaded. They also do not have to comply with Rapid Shutdown requirements. Therefore string inverters are cheaper and IMHO, a better product in the end.
 
A couple things. As a solar installer with many systems in place, I would never install strings of solar panels without module level control...ever...and its not about code requirements. I want to have eyes on my system and see the modules individually in the monitoring system. I don't want to have to worry about a difficult to find weakness in the series chain whether it be large bird poop on a single module or something more significant especially with a ground mount system. They are more sensitive to wind vibrations and wind storms. Microinverters can be purchased that handle multiple modules but have individual channels per panel. AP Systems QS 1 handles 4 modules. $271 per unit. Should last 25 years unlike a string inverter that burns out in 10 years. Also, all those Tesla fires on Walmart buildings were systems without module level control. You could also try optimizers with string inverters. But, microinverters don't cost much more than a line inverter in the long run. Yet, its worth thinking about clipping. Not sure about pairing 350 watt panels with microinverters that only let 300 watts through. So panel output matters. Optimizers might be worth considering (you can get 400 watt optimizers). If it were me, I would probably get more cheaper lower wattage panels and the QS 1 microinverters.
 
As mentioned, I am researching this for myself as well.

I am looking at an elevated ground mount system which will see full clear sky from morning until about 3 pm, and then shade will move across the array from one side to the other over the course of 1 hour.

Would it make more sense to use panel level optimization, or simply use string inverters and perhaps arrange the strings to match the shading.

Thanks
Jon
 
winnie said:
perhaps arrange the strings to match the shading.
Click to expand...
Would like to know if I have the following correct.

With a rectangular array of panels, and in the afternoon a sharp N-S shading line moving across the array from west to east, then I understand the relevant issues with a string inverter are:

1) It will be marginally more efficient to use "portrait" orientation. A panel in this orientation has 6 columns of cells and uses 1 bypass diode per 2 columns. So as part of the western column get shaded, the first two columns will drop out, leaving the panel at 2/3 power output, etc. While in landscape orientation, with 6 rows, when the shade first hits the panel the whole panel will drop out.

2) If you have one series string, the power output will step down in increments per the above until the number of unshaded (one-third) panels is too low to meet the string inverter's minimum DC input. No significant efficiency gain for optimizers/microinverters.

3) If you have multiple series strings, each one on a separate MPPT input to the inverter, then it's the same situation as (2), just in parallel. With respect to the minimum DC input level, I think it might be a little more efficient to spread the shading evenly across all strings, rather than the opposite, but I'm not certain. [The idea being that by the time all the strings are shaded enough to cause them all to drop out, the sun angle of incidence will be lower, so the power lost is less. Of course, if late afternoon power is more valuable, doing the opposite would be better.]

4) If you have multiple series strings and need to parallel two of them on the same MPPT input, it's important they have identical shading vs time profiles. Paralleling a string with all panels on the west edge of the array with a string with all panels on the east edge of the array would perform poorly (haven't thought out the exact behavior, but the shaded string would preclude the insolated string from producing.)

Cheers, Wayne
 
wwhitney said:
Would like to know if I have the following correct.

With a rectangular array of panels, and in the afternoon a sharp N-S shading line moving across the array from west to east, then I understand the relevant issues with a string inverter are:

1) It will be marginally more efficient to use "portrait" orientation. A panel in this orientation has 6 columns of cells and uses 1 bypass diode per 2 columns. So as part of the western column get shaded, the first two columns will drop out, leaving the panel at 2/3 power output, etc. While in landscape orientation, with 6 rows, when the shade first hits the panel the whole panel will drop out.

2) If you have one series string, the power output will step down in increments per the above until the number of unshaded (one-third) panels is too low to meet the string inverter's minimum DC input. No significant efficiency gain for optimizers/microinverters.

3) If you have multiple series strings, each one on a separate MPPT input to the inverter, then it's the same situation as (2), just in parallel. With respect to the minimum DC input level, I think it might be a little more efficient to spread the shading evenly across all strings, rather than the opposite, but I'm not certain. [The idea being that by the time all the strings are shaded enough to cause them all to drop out, the sun angle of incidence will be lower, so the power lost is less. Of course, if late afternoon power is more valuable, doing the opposite would be better.]

4) If you have multiple series strings and need to parallel two of them on the same MPPT input, it's important they have identical shading vs time profiles. Paralleling a string with all panels on the west edge of the array with a string with all panels on the east edge of the array would perform poorly (haven't thought out the exact behavior, but the shaded string would preclude the insolated string from producing.)

Cheers, Wayne
Click to expand...
Portrait orientation will cost you more in racking costs for ground mount.
 
So is that a yes, my analysis about how a hard N-S shadow line moving west to east across a small array affects string inverters is correct? No real upside to microinverters/optimizers, assuming you properly divide the array into strings for the string inverter?

Cheers, Wayne
 
wwhitney said:
So is that a yes, my analysis about how a hard N-S shadow line moving west to east across a small array affects string inverters is correct? No real upside to microinverters/optimizers, assuming you properly divide the array into strings for the string inverter?

Cheers, Wayne
Click to expand...
I obsessed about this for a while when stringing my array. In my case, I have some early morning and late day shading and it's not a clean north south shading line on the array, plus it's different for the morning shading and the evening shading. There was no perfect way to do it. If you have a clear view of each horizon, then I think yes stringing in north-south lines would be the way to go. Either way, I think you are looking at quite small gains.
 
electrofelon said:
If you have a clear view of each horizon, then I think yes stringing in north-south lines would be the way to go.
Click to expand...
If by that you mean some strings would have more of the western panels, and some would have more of the eastern panels, I would think that would only make sense if each string is on a separate input on the inverter. Otherwise, if you need to parallel strings on one input, you'd want equivalent shading on each paralleled string, so for an E-W moving shadow line, you'd do the opposite.

Cheers, Wayne
 
One of the things that continually surprises me is how inexpensive the panels have gotten compared to the 'balance of system'.

It is often a win to simply throw more panels at a less efficient orientation if it gets you cheaper racking.

Jon
 
Its almost always better to run long distances with 600Vdc than it is 240Vac. If you don't have the ability to put a string inverter outside, you'll need to add a rapid shutdown disconnect to the DC feeders entering the house.
 
wwhitney said:
If by that you mean some strings would have more of the western panels, and some would have more of the eastern panels, I would think that would only make sense if each string is on a separate input on the inverter. Otherwise, if you need to parallel strings on one input, you'd want equivalent shading on each paralleled string, so for an E-W moving shadow line, you'd do the opposite.

Cheers, Wayne
Click to expand...
Yes assume there is no paralleling of strings. Paralleling is actually rarely done in residential systems. I guess the exception would be if one used one of the large fronius inverters (15kw) that only have two MPPT's. If using an SMA, their largest is 7.7 kw and has three. Of course this is assuming the philosophy that if you have the MPPTs then you might as well use them.
 
BandGap1.1eV said:
Its almost always better to run long distances with 600Vdc than it is 240Vac. If you don't have the ability to put a string inverter outside, you'll need to add a rapid shutdown disconnect to the DC feeders entering the house.
Click to expand...
Actually I really haven't found that to be the case. I've done three systems in the 700-800 foot range recently and running strings back wasn't worth it over just running large aluminum conductors and bringing a 230.40 exception 3 service out to the array. Even priced it with fronius inverters and their 1000v strings. Still not worth it.
 
If you want module level monitoring then microinverters or DC optimizers are great. The other advantages that MLPE provides will mostly be lost on a well designed ground mount system. At 15kW this is a large residential or small commercial PV system so it can go either way. The module level monitoring can be really helpful in maintaining the health of the PV array of this size so I would not discount it.
 
Status
Not open for further replies.
Top