Oversizing the system

[Ampster]

..........

The way I learned and use the term "specific output" with respect to a PV system is that it is the annual output (either modeled or actual) in kWh/yr. divided by the system size in STC kW. Using AC or inverter size instead of the array STC size reduces the usefulness of specific output as a design tool.

I agree that the DC to AC ratio using array size is the most commonly used ratio.
Until the STC Wattage of an array exceeds the Wattage of the inverter, the usefulness of using inverter size in a ratio is not meaningful.

However, in the context of the title of this thread, which is about oversizing, an additional analytical tool may be useful to some people. I am not suggesting using that ratio as a dasign tool, I view it as an analytical tool to understand performance. For example one poster here, has a system which has a very high DC to AC ratio which on the surface may appear suboptimal. That poster has a dufferent cost per installed Watt of panels and inverters. Therefore he may have a different financial tradeoff in terms of optimal mix of components.

That is the way I learned financial analysus and how to use financial analysis as tools. The concept is to apply them where they can be useful to the analyst in evaluating the performance of a system. That may explain why commercial solar installations may use a different ratio and why that ratio may be irrelevent to most users here.

[J.P.M.]

I agree that the DC to AC ratio using array size is the most commonly used ratio.

Until the STC Wattage of an array exceeds the Wattage of the inverter, the usefulness of using inverter size in a ratio is not meaningful.

However, in the context of the title of this thread, which is about oversizing, an additional analytical tool may be useful to some people. For example one poster here, has a system which has a very high DC to AC ratio which on the surface may appear suboptimal. That poster has a dufferent cost per installed Watt of panels and inverters. Therefore he may have a different financial tradeoff in terms of optimal mix of components.

That is the way I learned and use financial analysis tools. The concept is to apply them where they can be useful to the analyst in evaluating a design or the performance of a system. That may explain why commercial solar installations may use a different ratio and it may be irrelevent to most users here.

1.) I'm not sure I understand what you're getting at or how what you're talking about has anything to do with the ratio of an array's annual energy output to the STC power rating of the panels and so an array. Not trying to start an argument here but I just don't know what your trying to say when you write about DC to AC ratio as you do.

2.) If a system has an inverter whose max. output is substantially lower than the system's panels are capable of producing "much" of the time, so that the lifecycle cost of the clipped electricity over the lifecycle period under consideration is greater than the lifecycle savings from using a smaller (and assumed less costly) inverter, that may impact the cost effectiveness of the system in a negative way and needs to be looked at.

If you or anyone else is interested, and maybe by way of a brief and oversimplified example of how I used some of the financial analysis tools I learned and brought from project engineering days to implement looking at the cost effectiveness of a PV system with respect to # 2 above as to how I sized the inverter for my system:

I first got the array sized for the duty (the annual array output) that met my project goal which was to supply my home's annual electrical usage with a mix of PV and POCO supplied power that resulted in the lowest lifecycle cost using a 12 year lifecycle using what I believe were realistic but maybe slightly conservative assumptions using modeled array performance considering performance rolloff with time, POCO electricity price inflation rates, my future usage changes, taxes and other stuff.

The system size is 5.232 STC kW. The theoretical string inverter size I started with was equal to the array's max. modeled output.

If the assumptions I used for the lifecycle analyses were valid (while being aware that any output from such analyses are very sensitive to the parameters and assumptions made, many of which are a crap shoot), the array size can be considered optimal as long as the assumptions a not invalidated or the parameters of the design goals don't change too much. If I stick with that assumption, any system output reductions due to clipping from inverter size reductions below that theoretical and initial inverter size would decrease the life cycle cost effectiveness and so would need careful considerations with respect to any cost reductions realized from possible savings from using a smaller inverter.

Working from that theoretical size, I reduced that theoretical inverter size in 100 watt increments until the LCOE (the Levelized Cost Of Electricity) of the blended mix mix of power costs went up by $0.01.
Then, I looked around for a string inverter that actually existed and was available and also met my other (non financially related) design requirements, but was slightly larger than the size I'd come up with.
Then, I adjusted the lifecycle cost by the changes caused by the available inverter I'd chosen and iterated the whole design same goal of system optimization with respect to lowest LCOE and other design goals. Fortunately or not, the method's first iteration converged.

[Ampster]

1.) ......I just don't know what your trying to say when you write about DC to AC ratio as you do.

The topic of the thread was "Oversizing the System" and I am saying that there may be circumstances unique to a user such that the traditional methods of analysis could result in sub optimal performance. Clearly the majority of users are wise to use the traditional tools. The example I gave, was that of a familiar member here who has a very high DC to AC ratio and who has a system that has very good performance measured by very little clipping and a ratio of annual production to kW of installed inverter capacity, His system has multiple ground mounted arrays of various orientations giving him a wider window of solar energy collection. That system was not designed using the methodology you described above. It all depends on where you are standing.

2.) If a system has an inverter whose max. output is substantially lower than the system's panels are capable of producing "much" of the time, so that the lifecycle cost of the clipped electricity over the lifecycle period under consideration is greater than the lifecycle savings from using a smaller (and assumed less costly) inverter, that may impact the cost effectiveness of the system in a negative way and needs to be looked at..............

I have no dispute with how you sized your system years ago and I agree that a significant amount of clipping is not optimal. Since the cost of systems is now in the area or $2 per Watt I will probably oversize my next system because I put some value on having a hedge against increased cost of energy. I also have the expectation that there will be erosion of NEM benefits and to me that requires additional analysis and assumptions. . In the past I have had success shifting significant loads to off peak but based on rate trends in California, I may have to move to a model of more self consumption. That is a topic of another discussion, and one that gets more traction on threads discussing trends in California.

[J.P.M.]

The topic of the thread was "Oversizing the System" and I am saying that there may be circumstances unique to a user such that the traditional methods of analysis could result in sub optimal performance. Clearly the majority of users are wise to use the traditional tools. The example I gave, was that of a familiar member here who has a very high DC to AC ratio and who has a system that has very good performance measured by very little clipping and a ratio of annual production to kW of installed inverter capacity, His system has multiple ground mounted arrays of various orientations giving him a wider window of solar energy collection. That system was not designed using the methodology you described above. It all depends on where you are standing.

I have no dispute with how you sized your system years ago and I agree that a significant amount of clipping is not optimal. Since the cost of systems is now in the area or $2 per Watt I will probably oversize my next system because I put some value on having a hedge against increased cost of energy. I also have the expectation that there will be erosion of NEM benefits and to me that requires additional analysis and assumptions. . In the past I have had success shifting significant loads to off peak but based on rate trends in California, I may have to move to a model of more self consumption. That is a topic of another discussion, and one that gets more traction on threads discussing trends in California.

Ampster: There are a lot of reasons dealing with personal choice that can lead to oversizing a system - however the person paying for the system chooses to define "oversizing". While I glory in your, my and their freedom to do pretty much what we want - even stuff I may think is not wise - NOMB - my experience has been that most people's primary reason for getting residential PV is to lower their electric bills. So I'll confine my comments to cost effectiveness and ways I think some folks have gone off the rails and shot themselves in foot by thinking multiple off optimal array orientations when not mandated by design or site limitations will lead to greater cost effectiveness. IMO and experience, it probably won't.

If you are referencing Bruce's system(s), I have a lot of respect for what Bruce has done and what he probably knows. I'm pretty sure I couldn't do what he's done, and just as sure he knows, and is most likely very familiar with, a lot a lot of stuff I'm ignorant about. But that door swings both ways and from what I think I might know, I believe his methods are missing some information and that's caused him to design things in a way that are less than optimal from a cost perspective. What I'm trying to say is that if cost effectiveness is one of Bruce's design goals, I don't see where he's optimized that parameter by using multiple orientations. Actually, I think the opposite. (Bruce, if you're reading this no disrespect or rancor intended)

If cost effectiveness does have any bearing on the design process, Bruce's or anyone else's, I respectfully disagree with the idea that, as you seem to imply, a wider window of of solar energy collection (time) is an automatic advantage to increased annual system production - again, unless the design is limited by some POCO mandated, or some site/facility mandated limitations on size or array orientation that precludes the use of one optimal orientation with respect to how much power, but not energy (as I'll be quick to differentiate), a system can produce.

Unless site limitations or POCO mandates preclude using an orientation that realizes the largest annual output per installed STC kW, that orientation, and not multiple orientations that result in lower system annual output per installed STC kW will have the best chance of being more cost effective because it will produce more kWh/yr. of energy to the inverter(s) per installed STC array watt and so usually be more cost effective. I don't remember Bruce mentioning those types of mitigating conditions for his site, and if he did, my apologies for having forgotten it.

As I've written in the past, while Bruce's his way of doing PV may give him more energy collection farther away from solar noon than a single, probably mostly southerly array orientation, the sum of the areas under the power production curves for an equal STC size but optimally oriented single array will be greater than one at the same site but that uses multiple orientations. A longer collection day doesn't mean more energy collection. In most cases it won't.

Doing a single optimally oriented array usually means either more energy production for the same STC panel wattage, or a smaller STC array size for the same annual production than a multiple orientation system. Either way, the single, optimally oriented array option (that is, the orientation that yields the highest annual output of kWh per installed STC kW of panel) will have an easier path to being more cost effective.

I do note that a single, optimally oriented array with the same STC rating as the multiple orientation array will probably mean larger (and more costly) inverters to handle higher power levels, probably for an hour or 2 on either side of the daily time of minimum solar incidence angle on the array, and quite possibly heavier gage wiring mandated by more current flow at times for the single orientation array, but maybe not, particularly as the array size can be reduced, so current flow will be somewhat reduced. However, even with lower prices as are now the norm, the incremental cost (or savings ) per STC panel W of increasing (or decreasing) an array's STC size is greater than the per watt incremental savings gained from decreasing an an inverter's size that's enabled by the lower peak output rom the less optimally oriented array..

As for this issue of using inverter capacity rather than array STC wattage in the denominator when figuring annual specific production, while I understand some situations such as solar or wind commercial power plants use power output from inverters as the descriptor for nominal plant size/capacity, that's done for reasons having to do mostly with interfacing with the grid and its financial machinations, and/or for other reasons that are, in any case, mostly not applicable to residential PV. But using inverter size or max. output rather than the STC array size for calcing annual specific output for most residential PV applications can be confusing and/or misleading. As a matter of fact, using inverter capacity to describe specific system yield in terms of annual kWh production per "installed kW" is both pretty useless, and I'd suggest perhaps misleading as a measure of system annual production performance.

[Ampster]

Hopefully @Sami Najjar has the feedback he wanted in order to decide the appropriate balance of system components for his situation.

[J.P.M.]

Hopefully @Sami Najjar has the feedback he wanted in order to decide the appropriate balance of system components for his situation.

If he needs more info/opinion, I'm sure he knows where we all are.

[Sami Najjar]

Thank you guys I'm grateful