High DC-to-AC Ratios

[andrew.tkelly]

I've heard it said in various videos and white papers that if you have a high DC-to-AC ratio, you'll be able to capture more energy than if the ratio was closer to unity. I could see how this would be the case in a string w/o MLPE. You would have higher voltages at earlier and later times during the day --higher "shoulders" on the graph of power vs. time.

However, would this be true with MLPE, specifically SolarEdge DC-optimizers? Since all the panels act as if they were in parallel, the start-up voltage would be the same regardless, no?

I'm trying to see the benefits of SolarEdge's new H-series having the ability to take on a 1.5 DC-to-AC ratio. Is more energy able to be captured? Something to do with the financial impact?

Regards,
Andy

 

[Besoeker]

I've heard it said in various videos and white papers that if you have a high DC-to-AC ratio, you'll be able to capture more energy than if the ratio was closer to unity.

I don't see how.
The incident solar energy is what it is. Conversion efficiency is down to the solar cells.

 

[jaggedben]

The 'shoulder' of the production curve (and the cloudy day production) will be higher when you add module power, even if the amount of time the inverter 'clips' is greater, thus you always get some more energy by adding a module. You don't get more energy from upsizing inverters unless you already have added module capacity. It's more or less completely independent of MLPE considerations, except that if you're using micro-inverters your consideration is just one-module-to-one-inverter.

The main advantage of higher DC-to-AC ratios is to avoid service and other electrical upgrades. Say the customer needs a 5.5kW system for their energy needs but only has a 100A panel. Using a 3.8kW inverter costs them a few hundred dollars in 'clipped' energy production over the life of the system, but avoids $3000 to upgrade the service panel. The extra 1.5kW pays for itself slightly slower than the rest of the system, but not much. I won't recommend that to someone with a 200A service where I save $50 on an inverter but the customer loses, say, $700 over the life of the system. But to avoid having to deal with a service upgrade that we make no money on and the customer doesn't otherwise need, I certainly would.

Similar considerations also affect commercial and utility scale systems, where there also may be additional considerations. For example, there may be more of an upper limit on available inverters, and grid operators may be happier with a system that operates at a fixed output for a good part of the day. Also adding some strings late in the process doesn't necessitate a comprehensive redesign of the AC side.

Back when modules cost a lot more than inverters this oversizing made less sense. Now that they cost about the same, it's cheap to add more modules.

 

[andrew.tkelly]

The 'shoulder' of the production curve (and the cloudy day production) will be higher when you add module power, even if the amount of time the inverter 'clips' is greater, thus you always get some more energy by adding a module. You don't get more energy from upsizing inverters unless you already have added module capacity.

Interesting! Let me see if I understand: So the energy gain in the 'shoulders' tends to be greater than what will get clipped in the 'peak' of the energy curve? And this is because the duration of time spent in the shoulders is much longer than the time spent in the peak?

-Andy

 

[jaggedben]

Interesting! Let me see if I understand: So the energy gain in the 'shoulders' tends to be greater than what will get clipped in the 'peak' of the energy curve? And this is because the duration of time spent in the shoulders is much longer than the time spent in the peak?

-Andy

I didn't say the energy gain on the shoulders is greater than the potential amount clipped in the peak. I just said that you make gains on the shoulders whenever you add more module capacity. The relative gains diminish as you add more modules. The first few modules above the inverter's rating will barely get clipped at all, whereas the higher you go the more of the additional potential energy will get clipped.

Bottom line: the cost-effectiveness of oversizing depends more fundamentally on the price of the module than the DC-to-AC ratio. If I could hypothetically keep adding modules forever for no additional money I'd always get a little more on the shoulder and it would always be worth it.

Various softwares will simulated clipping for you, including SolarEdge's. I tend to see no more than 1-2% clipped annual energy with DC-to-AC up to 125%. Over that amount and the clipped energy percent increases faster.

Another thing: you get less clipping if you have east-west arrays. With a 45 degree roof facing due east and west you could perhaps load up an HD Wave to its max and get no clipping at all.

 

[ggunn]

I don't see how.
The incident solar energy is what it is. Conversion efficiency is down to the solar cells.

Yes, but the orientation of the modules and location of the array is important as well. It is not uncommon to use a DC:AC ratio of 1.5 or even more in northern latitudes with low tilt and/or off azimuth PV arrays. The modules in a system like that will never see conditions even close to STC. Even in more ideal settings driving up the ratio makes sense if the added production on the shoulders of the power production curve more than makes up for the loss to clipping on the peak.

You won't see this with PVWatts but a more advanced (read: expensive) software program like PVsyst will show it.

 

[ggunn]

I've heard it said in various videos and white papers that if you have a high DC-to-AC ratio, you'll be able to capture more energy than if the ratio was closer to unity. I could see how this would be the case in a string w/o MLPE. You would have higher voltages at earlier and later times during the day --higher "shoulders" on the graph of power vs. time.

However, would this be true with MLPE, specifically SolarEdge DC-optimizers? Since all the panels act as if they were in parallel, the start-up voltage would be the same regardless, no?

I'm trying to see the benefits of SolarEdge's new H-series having the ability to take on a 1.5 DC-to-AC ratio. Is more energy able to be captured? Something to do with the financial impact?

Regards,
Andy

It's not a voltage thing, it's a current thing, and current depends on the amount of insolation (sunlight) that a module is being exposed to. If an array is seeing STC conditions (they almost never do, but anyway), at that point in time, a DC:AC ratio any greater than 1.0 will cause the inverter to clip. Some clipping is not necessarily a bad thing, though.

How much more than 1.0 you should "overload" an inverter for maximum production is very array location and orientation dependent.

 

[GoldDigger]

It's not a voltage thing, it's a current thing, and current depends on the amount of insolation (sunlight) that a module is being exposed to. If an array is seeing STC conditions (they almost never do, but anyway), at that point in time, a DC:AC ratio any greater than 1.0 will cause the inverter to clip. Some clipping is not necessarily a bad thing, though.

How much more than 1.0 you should "overload" an inverter for maximum production is very array location and orientation dependent.

And the relative value of peak and shoulder watts will strongly depend on your specific Time Of Use rates from POCO if you are working under a TOU tariff.
For pure off gridders who need to keep their batteries charged but do not care much when, the value of over paneling in the summer is that you are better off in the winter when both the number of sun hours and the efficiency of the panels may be lower (if you do not change the elevation angle of the panels seasonally.)

 

[ggunn]

And the relative value of peak and shoulder watts will strongly depend on you specific Time Of Use rates from POCO if you are working under a TOU tariff.
For pure off gridders who need to keep their batteries charged but do not care much when, the value of over paneling in the summer is that you are better off in the winter when both the number of sun hours and the efficiency of the panels may be lower (if you do not change the elevation angle of the panels seasonally.)

That, too. The bottom line is that there is no magic number for a DC:AC ratio.

 

[Electric-Light]

Utility AC is a sine-wave. What we call 120v means it has an RMS voltage of 120v which means a resistive load like a heating element will draw the same average power as 120v DC.

At the very top of waveform is 170v. (1.41 times). Add some voltage for voltage drop and you see why DC input has to be at least 1.5 times the AC RMS voltage you're trying to create to draw out the arches without using a boost converter. (which is part of MPPT).

I believe this is where 1.5x comes from.

 

[ggunn]

Utility AC is a sine-wave. What we call 120v means it has an RMS voltage of 120v which means a resistive load like a heating element will draw the same average power as 120v DC.

At the very top of waveform is 170v. (1.41 times). Add some voltage for voltage drop and you see why DC input has to be at least 1.5 times the AC RMS voltage you're trying to create to draw out the arches without using a boost converter. (which is part of MPPT).

I believe this is where 1.5x comes from.

The DC:AC ratio has nothing to do with how much voltage you need to generate line AC voltage. An SMA inverter (see data sheet http://files.sma.de/dl/27676/SB3.0-7.7-US-DUS170619W.pdf) can make 240V AC from as little as 100V DC.

 

[Electric-Light]

I see what it is now. Never mind on what I said. he is talking about balancing out sum of rated kW of solar panel whose minute by minute kW output is weather dependent vs rated kW of inverter.(and its service factor) and possible implications of on-premise equipment limitations and possible contractual limits to allowable export.

Not directly relevant but something I want to explore is possible productive valuable way to utilize the surplus production on-site.