Why design a system with any clipping of PV production?

[JoeNorm]

It is pretty common to oversize the PV array in relation to the inverter capacity. This often results in some amount of "clipping" at the highest output days.

I don't understand why you'd want to clip at all when the inverter size can be chosen to handle the highest output of the PV array. Maybe I remember reading that an inverter is most efficient running at full capacity(could this be true?). If this is the case i can see the argument for high DC/AC ratio.
I know there are at least a couple experienced PV designers here, maybe they can weigh in on this? Thanks

 

[wwhitney]

If inverters were free (or just much much cheaper than panels), you'd be absolutely correct.

If panels were free (or just much much cheaper than inverters), then the opposite would be true--you'd want to flood your expensive inverter with as many panels as possible so it's producing at full capacity for nearly the whole time the sun is up.

Since in reality neither is true, the economic tradeoff point will include some clipping.

You can also think of it this way--you have a system designed that will have some clipping (maybe DC/AC ratio = 1.3), and you want to spend $500 more for more production. You could spend it on upsizing the inverter--the energy gain will be the clipping avoided.

Or you could spend the $500 on some more panels for the existing inverter. Those panels will have cause even more clipping--i.e. the additional energy generation will be less than if you were adding panels to an inverter without any clipping. But that additional energy generation may still be more than the first case of upsizing the inverter.

So you keep adding panels to the inverter with more and more clipping until you reach the point that upsizing the inverter is actually more cost effective than just adding panels to the inverter you have.

Now if you hit a limit on the number of panels you can use (i.e. roof space), and you still want to spend more money for more production, then you might upsize the inverter even though adding panels would be more cost effective if it were possible.


Cheers, Wayne

 

[pv_n00b]

It's because it increases energy production for a given inverter size. If you look at the production curve over a day it's a bell shape. If you have a PV array sized to not clip, you will extract all the energy under the curve. Now increase the PV array size for the same inverter and you will find that the area under the curve that the inverter can extract power increases, the shoulders get wider even though some of the peak production is clipped.
There is a point where increasing the PV array size will result in the clipping losses exceeding the area increase of the shoulders. That's the ultimate financial limit. There are technical limits on the DC/AC ratio too.

 

[jaggedben]

Another thing is that everything on the AC side of the inverter output has to to be sized to that output. Bigger wires, panels etc. And especially if that triggers upgrades to existing equipment according to article 705 or on the utility side, then the cost of those upgrades is typically more than the value of clipped energy.

 

[jaggedben]

Also note that in the real world clipping is likely to be negligible or even non-existent with a DC to AC ratio below about 1.2, or with certain designs such as east and west facing panels. So actually matching inverter nameplate to panel nameplate is not necessary to avoid clipping most of the time. Nor is it worth spending much time analyzing for resi installs with moderate DC-to-AC ratio.

 

[BandGap1.1eV]

Net metering and utility scope also come into play. If there is only 500 kVA of capacity left on a feeder but your site can support much more, then you're going to saturate that inverter to make as many kWh as possible. Similar holds true on the residential side with net metering.

 

[solarken]

If inverters were free (or just much much cheaper than panels), you'd be absolutely correct.

If panels were free (or just much much cheaper than inverters), then the opposite would be true--you'd want to flood your expensive inverter with as many panels as possible so it's producing at full capacity for nearly the whole time the sun is up.

Since in reality neither is true, the economic tradeoff point will include some clipping.

You can also think of it this way--you have a system designed that will have some clipping (maybe DC/AC ratio = 1.3), and you want to spend $500 more for more production. You could spend it on upsizing the inverter--the energy gain will be the clipping avoided.

Or you could spend the $500 on some more panels for the existing inverter. Those panels will have cause even more clipping--i.e. the additional energy generation will be less than if you were adding panels to an inverter without any clipping. But that additional energy generation may still be more than the first case of upsizing the inverter.

So you keep adding panels to the inverter with more and more clipping until you reach the point that upsizing the inverter is actually more cost effective than just adding panels to the inverter you have.

Now if you hit a limit on the number of panels you can use (i.e. roof space), and you still want to spend more money for more production, then you might upsize the inverter even though adding panels would be more cost effective if it were possible.


Cheers, Wayne

Very well said!

 

[solarken]

It is pretty common to oversize the PV array in relation to the inverter capacity. This often results in some amount of "clipping" at the highest output days.

I don't understand why you'd want to clip at all when the inverter size can be chosen to handle the highest output of the PV array. Maybe I remember reading that an inverter is most efficient running at full capacity(could this be true?). If this is the case i can see the argument for high DC/AC ratio.
I know there are at least a couple experienced PV designers here, maybe they can weigh in on this? Thanks

Some excellent explanations supporting the fact that good design almost always includes tradeoffs. I'll add that the amount of DC array to AC inverter ratio that is acceptable can vary widely based on the details of the location and site. Due South arrays at an optimum tilt for the latitude would ideally have lower ratio than an array split between East And West. A site with shading would ideally have a higher ratio than one with no shading. Colorado systems would ideally have a significantly lower ratio than a system in Seattle. 1.2 to 1.3 is just a rough estimate. Experienced designers may choose ratios from 1.1 to 1.5 or even higher because that is the best fit for a specific site.

 

[pv_n00b]

Some DC/AC ratios can get pretty high, 3 is not that uncommon and I remember seeing a 5 one time but have no idea what that was doing for them. Inverter manufacturers usually have an upper limit that you have to get their signoff on if you want to go higher and that signoff is usually not hard to get.
Remember that the DC rating used is the STC DC output of the module, which is unlikely to happen in the field. So the base 1.2 ratio is an adjustment from STC to real life assuming the PV module is actually only going to produce 80% of the STC output in the field.

 

[ggunn]

Some DC/AC ratios can get pretty high, 3 is not that uncommon and I remember seeing a 5 one time but have no idea what that was doing for them.

3.0 and 5.0, or 1.3 and 1.5? I assume you mean the latter; I cannot imagine a 5.0 DC:AC.

 

[SunNinja]

It's because it increases energy production for a given inverter size. If you look at the production curve over a day it's a bell shape. If you have a PV array sized to not clip, you will extract all the energy under the curve. Now increase the PV array size for the same inverter and you will find that the area under the curve that the inverter can extract power increases, the shoulders get wider even though some of the peak production is clipped.
There is a point where increasing the PV array size will result in the clipping losses exceeding the area increase of the shoulders. That's the ultimate financial limit. There are technical limits on the DC/AC ratio too.

I had the exact same question as the poster, and I could not believe it until I ran the calculations for myself. Can't find the doc that explained this bell curve in technical terms with calcs so I could plug in the numbers myself now, but here's a related white paper.
chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://knowledge-center.solaredge.com/sites/kc/files/inverter_dc_oversizing_guide.pdf

 

[ggunn]

I had the exact same question as the poster, and I could not believe it until I ran the calculations for myself. Can't find the doc that explained this bell curve in technical terms with calcs so I could plug in the numbers myself now, but here's a related white paper.
chrome-extension://efaidnbmnnnibpcajpcglclefindmkaj/https://knowledge-center.solaredge.com/sites/kc/files/inverter_dc_oversizing_guide.pdf

Yes. We had a customer who complained about some minor clipping he saw in his eGauge waveform in the middle of some clear cool days. When we explained to him that the only ways to make his system stop doing that would be to reduce the size of his array (which would reduce his overall energy production) or to replace his inverter with a larger one (which would cost him more money than the clipping losses would over the life of his system), he decided to live with it.

 

[pv_n00b]

Yes. We had a customer who complained about some minor clipping he saw in his eGauge waveform in the middle of some clear cool days. When we explained to him that the only ways to make his system stop doing that would be to reduce the size of his array (which would reduce his overall energy production) or to replace his inverter with a larger one (which would cost him more money than the clipping losses would over the life of his system), he decided to live with it.

The age-old conundrum. Give a client access to too much data they have not been trained to interpret properly and you get calls about the system not working when it is working as designed. Don't give them enough data and they complain about that. Luckily after 3 months or so most clients don't even look at it anymore. I have gone into so many commercial sites that are more than a few years old and no one knows how to access the production data anymore, or that accessing it was even possible.

 

[ggunn]

The age-old conundrum. Give a client access to too much data they have not been trained to interpret properly and you get calls about the system not working when it is working as designed. Don't give them enough data and they complain about that. Luckily after 3 months or so most clients don't even look at it anymore. I have gone into so many commercial sites that are more than a few years old and no one knows how to access the production data anymore, or that accessing it was even possible.

We have been called in to look at existing systems that had not been operational for a very long time and no one had even noticed.

 

[ggunn]

The age-old conundrum. Give a client access to too much data they have not been trained to interpret properly and you get calls about the system not working when it is working as designed. Don't give them enough data and they complain about that. Luckily after 3 months or so most clients don't even look at it anymore. I have gone into so many commercial sites that are more than a few years old and no one knows how to access the production data anymore, or that accessing it was even possible.

Even though the production of that customer's system substantially exceeded the estimate provided to him in the sales process, he wanted us to replace his inverter with a larger one for free. We declined.

 

[wwhitney]

I think it's worth commenting that residential PV marketing is not always clear on the idea of separate AC and DC ratings of an installation, and that the DC/AC ratio may be bigger than one. So if all a person ever hears is that a system is a 10 kW system, they would reasonably expect it to be able to generate 10 kW under some circumstances. If that's a DC rating and the system has 7.6 kW inverter, obviously that's never going to happen.

Not a comment directed to anybody in particular, just an industry wide observation.

Cheers, Wayne

 

[jaggedben]

Often contracts and marketing materials state it obliquely in a way the average residential client is not going to notice. Like stating the system size as DC, and giving an somewhat inscrutable inverter model number.

But yes, I know from experience it's much more awkward to explain to a customer that you saved them $5000 on a service panel upgrade when it's past tense as opposed giving them the chance to decide beforehand.

 

[pv_n00b]

On top of that having to explain to people the difference between kW peak production and annual kWhr production.