Go to the Northern Arizona Wind & Sun - Online Solar Store

Announcement

Collapse
No announcement yet.

Enphase Microinverters - Cost Effective?

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • Enphase Microinverters - Cost Effective?

    Hi All!
    Just became aware of these Enphase Microinverters (http://www.enphaseenergy.com/products/) and was wondering if anyone here has experience with them. since I have a very long cable run (~400ft (round trip) I was thinking that conversion to 240V AC at the PV source might be a good idea to minimize IR loss. What would be the advantage of this microinverter option vs. wiring up panels for 24V or 48V (more loss) and using a conventional MPPT controller and a standard pure sine wave inverter. Looks like you need one microinverter per PV panel (~$200 each) and for ~180W panel you can anticipate ~750ma @240V. If I am doing my math correct for a 15A AC branch circuit that would add ~$4000! (need ~15 inverters). What's the deal???

    Thanks for any comments/ thoughts!
    Best
    Wayne

  • #2

    Re: Enphase Microinverters - Cost Effective?

    Re: Enphase Microinverters - Cost Effective?

    Yep... You are looking at them correctly.

    You might call NAWS (our host)... They were supposed to have installed a large area on their roof some time this spring--last I heard.

    One thing I am not sure I like is their logging... They use the AC line for communications and you need to purchase a host to monitor them. There is nothing to look at to make sure they are all working correctly.

    $340 for the device--And, it may need internet access for full functionality (you access your data through their website). And there may be a yearly web fee too.

    If you had a large GT inverter--you could run a higher voltage DC (around 400-500 VDC) from the panels to the GT inverter and save copper losses that way.

    The best you could pick for a off-grid system would be ~100 VDC (140 VDC maximum working voltage)... Better than a lower voltage--but still not as nice as the high voltage GT inverter.

    Could also install the battery shed 400' away (with the inverter) and run 240 VAC to the home too.

    Lots of options--but none are good if you don't get sun from ~9am-3pm--or possibly 7am-noon or noon-7pm (arrange your arrays to point east or west).

    -Bill
    Last edited by BB.; May 24th, 2009, 16:15. Reason: fix spell'n and grammer.
    20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

    Comment


    • #3

      Re: Enphase Microinverters - Cost Effective?

      Re: Enphase Microinverters - Cost Effective?

      The Enphase converters produce 240 with a neutral so from an IR loss point of view, they are 120V. My 14 panel string produces about 350V nominally and that is #X the voltage giving a 3X IR advantage. For IR advantage, you want the lowest current and highest voltage. Running 120V requires 3 times the copper for the same IR loss at a equivalent current. That would suggest the Enphase approach is not good from an IR point of view for long distances from the panels to the main meter panel.

      On the other hand, it does have some large advantages that COULD outweigh the copper cost. Since the individual Enphase inverters maintain the optimum MPP per panel over all conditions and life and provide per panel performance monitoring, that can have a huge advantage if you have a large array with long strings and combined ( large series parallel array) and are concerned about shade ( trees, telephone poles, towers, TV antennas etc..). If you have a small array with no combining and short strings and NO shade issues and the environment is clean plus long panel to meter runs, the Enphase gain is minimal. Unless they get their price down to the $100-$125 range, there will always be trade offs. At $100 per inverter and good reliability and performance monitoring, its almost a NO BRAINER.

      National Semiconductor is also offering an MPP optimizer called the "Solar Magic" http://www.solarmagic.com/products which has the same MPP gains plus I believe the Performance monitoring. They too need to get their cost down to $100. Both approaches can squeeze more power out "harvesting" but efficiencies are critical. The aggregate efficiency of the MPP optimizer and the DC-AC inverter must be good = 95% TOGTHER!

      Tom

      Comment


      • #4

        Re: Enphase Microinverters - Cost Effective?

        Re: Enphase Microinverters - Cost Effective?

        As far as I know (which is not that much) wrt to the Enhase inverters--they are true 240 VAC inverters (or 208/230 VAC depending on the particular unit you purchase).

        I would guess the Neutral is part of the new code for GT inverters that not only require the inverter to monitor for "240 VAC 60Hz" across L1 and L2, they are also monitoring for 120 VAC across L1-Neutral and L2-Neutral so they can shutdown if there is a neutral "fault" (broken wire or earth ground to neutral common bus connection).

        My 4 year "old" 3 kW GT inverter (Xanrex) only has L1/L2/green wire for connections. The "new" GT inverters now have a neutral wire pulled to the inverter (for monitoring--no power transmission that I know of).

        Enphse does not (that I am aware of) send any current down the neutral tap.

        -Bill
        20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

        Comment


        • #5

          Re: Enphase Microinverters - Cost Effective?

          Re: Enphase Microinverters - Cost Effective?

          I'm not sure if I understand what you mean by "branch circuits" and "add $4000". As I understand the Enphase micro inverters, they are used one per panel optimizing that panels MPP. That means the panel will always produce close to its real power potential. That is their main claim to fame. It also allows monitoring on a per panel basis making trouble shooting EASY. If you array has say 12 panels in a string, all the Enphase inverters are connected to a string of 3 wires ( hot, neutral, hot) for a 240V circuit. For the normal DC string case, if you used 10AWG wire for the series string, you would also put all the Enphase inverters on that same string except it has 3 wires. The gauge would depend on the distance and current which is all 12 Enphase converters summed so its not 750ma but 12 x 750 ma or 9 Amps for 12 converters. That string would return to a breaker combiner at 15A. The other Enphase strings would hit other breakers and all the current combined with a summation breaker back into the main panel. If there were 6 strings of 9 AMPS each, then the output breaker would be 6x9AMPS x1.25 x 1.25 = 90AMP breaker.

          The Enphase inverters all hang on one string of panels where the size would be chosen by physical convenience or wire gauge and distance back to the panel.

          The main cost difference is that the price tag on the Enphase Inverters x the number used must be about the same as the one centralized Inverter. This of course is adjusted by the MPP gain the micro inverters give which can be used to offset the higher price. Or so the story goes.

          Comment


          • #6

            Re: Enphase Microinverters - Cost Effective?

            Re: Enphase Microinverters - Cost Effective?

            I am not convinced that on a good solar array (no shade problems, no "bad panels") that an Enphase would generate much more than a few percent of "additional" power over a central GT inverter installation.

            If there are multiple shade problems (shadows tracking across and array)--perhaps an argument could be made for an Enphase (assuming that the panels+obstructions cannot be move/fixed).

            Add in the issue of weather sealing 20+ times more boxes and mounting on a hot roof of your electronics--and additional work required for Removal and Replacement of failed units (vs one GT unit on the ground)--I would have to have a very special installation for an Enphase to make obvious sense for a large installation.

            By the way, have you found out if the Enphase monitoring hardware/software must connect to the Web and pay the yearly fee to access the data--or can you directly talk with the Data Unit with a local PC and bypass the web/internet subscription issues (more hardware and more costs).

            -Bill
            Last edited by BB.; May 24th, 2009, 16:18. Reason: fix spell'n
            20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

            Comment


            • #7

              Re: Enphase Microinverters - Cost Effective?

              Re: Enphase Microinverters - Cost Effective?

              bill,
              you bring up a good point on failures as those inverters will be out there more exposed to the elements and even more prone to lightning emp too. even barring that, eventually failures do occur in electronics and that would mean a trip to the roof and fixing or replacing the inverter. so is it written that an enphase inverter must be on the roof with the pv or can it too be on the ground? on a positive note for the enphase, the downtime and power lost for 1 enphase won't affect any other pvs with enphases than would be encountered with 1 main ground mounted gt inverter.
              NIEL

              Comment


              • #8

                Re: Enphase Microinverters - Cost Effective?

                Re: Enphase Microinverters - Cost Effective?

                I am not convinced that on a good solar array (no shade problems, no "bad panels") that an Enphase would generate much more than a few percent of "additional" power over a central GT inverter installation.
                It is true that shade problems and bad panels are two key cases where some form of MPP optimization strategy can make a 25%+ improvement. When you buy brand new panels that are identical and from the same manufacturers process run, their MPP VI curves are fairly well matched. That means they will all sit close to their maximum power point ( the knee of the VI curve ) and with a clean shade free array, the per panel inverter doesn't have much MPP loss to harvest the power from.

                That however isn't very realistic. I am an electrical engineer having founded 3 large systems startups and was the system architect for about 10 large systems in my years. This is analogous to having a car that has 300HP when you drive it off the lot but as the engine, fuel injectors, and all its systems slowly change over time and in 5 years you have only 200 HP. The only way to keep the engine at near its 300HP is to keep it maintained and in good tune. With a car that an be done but can be expensive to keep taking it to the shop.

                In a solar array that is hard wired in a series string and combined in parallel, as the panels age the VI curve maximum point slowly changes and not all exactly the same. In addiion, anything such as shade, partial or moving or variable clouds, some dirt on the glass obscuring part of a panel, or simply upcoming sunrise or sunset or the gradual temperaturs rise/fall across the panels differential all combine to cause the VI curve to change and or differ.

                If you have say brand new 200 watt panels then at the STC ( standard test conditions) they will all have their VI curve move together as the temperature and light conditions change. If the temperature of ALL the panels doesn't rise or fall together or stabilize at the same point ( due to say the roof material under them providing heating or cooling effects ) from sunrise to noon to sunset, then their VI curves and therefore their VI knee or MPP will all vary. Eventually those 200 watt panels might be producing 150 watts even at noon on a sunny day.

                Just like batteries in a flashlight, one with low current will drag down the other batteries from producing their maximum power. In a string, a panel that wants to operate at the lowest current will push all the other panels in that string down their VI curve away from their maximum power point where V x I = max. That one panel might be bad, it might have aged 1, 3, 5,10, 20 years differently, a passing cloud is impacting it, shade is passing it, dirt is obscuring some of its cells, the sun angle hits that panel last in the morning, its mounted on op of a different roofing material (metal, composite, slate etc..). Then the string that would like to produce the lowest voltage (sum of all its panels voltage given the same varied across string panel sets conditions) will drag down all the other strings to match. There can be only ONE current in a string and only ONE voltage once the strings are combined. That one panel that forced a string current and that one string that forced a string voltage drop both due other same varied conditions pushes many if not all panels away from that perfect MPP point where V x I = maximum.

                The VI=maximum equation multiplies V x I to get the maximum power. That means the maximum power is exponentially a function of the voltage and current. That means a 2% change in voltage or current pushes a 4 times ( 2 squared) power change. A 5% change in V or I becomes a 25% change in power output. That is nasty!

                The problem is very real and has been studied and understood for years. New panels under ideal conditions will not experience much loss. In the real world after 5-10 years of aging and typical cleanliness, clouds, temperatures, mounting arrangements, etc... the average of all installations will probably suffer a 20-25% loss. Changing a 7 year old problem panel with a new one where the original panel is out of manufacture or the process has evolved such that the VI curve is different will just compound the problem. When a panel is changed, the new panel MUST have the same VI curves as the original or there is another bummer. I'll guarantee that SHARP, Kyocera, etc.. will all discontinue and or move up their processes and after 5 or so years, you won't be able to just replace a panel without causing possibly BIG problems. A new set of panels probably came from the same line and are marvelously matched but after 5 years, its anybody's guess what you will get for a VI curve and MPP point curve.


                This a NOT a small problem. Now the idea of installing 56 little Enphase inverters under my solar array in the heat, cold, dust is not very appealing. Not being able to know what's wrong if their communication system has failed is a bummer. Crawling under my array to replace one of their inverters doesn't seem nice. The AC IR loss on my 175 feet of wire would no work. The Enphase solution to this very real problem is effective but not appealing. Reliability of 56 units under m array gives me a bad feeling.

                National Semiconductor is working on a MPP optimizer which effectively electronically adjusts the VI curve of each panel automatically is simpler but it too must sit under my hot array in the dust. It has an efficiency ( not yet disclosed) that must be multiplied by the central inverters efficiency which cold trash the whole idea if not well above 95%. It doesn't help to adjust the MPP to its maximum and then lose it in the optimizer and inverter efficiencies.

                My wish is that PV panel manufacturers would standardize on a communication protocol built into the panels( like all other industries through the IEEE or maybe Bluetooth orZigbee ) so at least users can determine if a panel is off its MPP, its V and I or has a problem easily. Then they could also help by standardizing MPP curve ratings and eventually build the National Semi "like" optimizer into the panel structure to take advantage of the packaging and cooling cost. Many IEEE and other papers have been written on MPP optimizers for many years.

                Technically designing a small converter that adaptively tracks the MPP and adjusts the voltage and current characteristics that is very efficient is feasible. It could support the communication protocol like Zigbee or Bluetooth and the problem would disappear into the panels electronics. Then the simple topology used today with a very efficient inverter on the large string/combined would be perfect and the 25% loss gone and problems automatically reported via a standard wireless link to any laptop within a few hundred feet.

                So I disagree its not a big problem but do agree the micro-inverter sounds good but in practice has issues I would avoid.

                T
                Last edited by BB.; May 24th, 2009, 9:10. Reason: Fix "quote" for "PHP" tags -Bill

                Comment


                • #9

                  Re: Enphase Microinverters - Cost Effective?

                  Re: Enphase Microinverters - Cost Effective?

                  Again, to a degree, I think this is a solution in search for a problem.

                  If you are adding large arrays--typically, the new array will be from the same mfg and all collected into one MPPT inverter--so that is not a problem.

                  On my simple 3.5 kW array--if I was to replace one bad panel--and say that new panel had 100% more available power (who knows why--better I*V curve, shifted IV curve, or whatever)--and in my "degraded" application it only produced ~the same power as the rest of the panels... Assume "replacement" panel is 2x the Watt/sq efficiency of the "existing" panels.

                  20 panels, 19 at "100%" + 1 running at 100% of array
                  20 panels, 19 at "100%" + 1 running at 200% of rest of array

                  "Old" single inverter output = 100% with central inverter
                  "New" distributed inverter = 21/20= 105% with distributed inverter

                  Even if I had one bad panel, and I needed to take a second panel (parallel string) out of service (to balance Vmp) because I could not afford to replace the panel--I would still only see a:

                  18/20=90%

                  A 10% reduction in output... I would have to loose more than 4 panels before I would see a 25% reduction in output (and a distributed inverter could not recover that 25% loss anyway).

                  So--replacement wise of a single panel--it would be hard to see a 25% improvement (a "New" distributed inverter output would only give me a 5% improvement). And, if the inverter is "matched" to the panels--the old panels are 175 watt peak, and the new "2x efficiency panel" would be 350 watts. Another problem--I would need to replace my distributed inverter too--The "old" inverter was a 200 watt device. I need a new 350 watt device to use the capabilities of the new (mythical) panel anyway.

                  Make the case that I have to replace 1/2 the array (hail damage and/or mfg defect which the mfg sent me new panels).

                  So, now I replace 10 of my panels which are back to 100% and the rest are at 90% because of aging... I am a smart guy--so I replace the panels in a balanced method (two strings, 1/2 of the panels go in each string to balance Vmp--or if I think Vmp's are the same, I put all panels in one string to allow Imp of new panels to exceed Imp of old).

                  10 old panels at 90% + 10 new panels 90% with central inverter = 90% of "new" output
                  10 old panels 90% + 10 new panels 100% (distributed inverter) = 95% of "all new" output

                  So--I got a 5% improvement with distributed inverters when replacing 1/2 my string (at best). And depending on what aged (Imp or Vmp)--then I can get 100% of the capability of both old+new panels. I only get 95% of capability if we assume that the panels aged in both Imp and Vmp--and the overall string is limited by the other panels' deficiencies.

                  But--to get this 5% improvement--I would have to have spent:

                  [Note: "Solartek" a few posts below said that my guess of $0.50 a watt installation cost was way too high--His estimates are about 2% of his labor costs. So, the price differences would be less between the two types of inverters--Bill B.]

                  Assume panels are $4 per watt. Assume Central Inverter is $0.50 per watt. And distributed inverters are $1.00 per watt +$0.50 per watt additional installation costs.

                  3,500 watt of panels * ($4+$0.50) = $15,750
                  3,500 watt of panels * ($4+$1.00+$0.50) = $19.250

                  $19,250/$15,250=126% price hit...

                  So--to get my 25% possible improvement in the future, I would have to take a 26% hit in pricing now.

                  Instead, I could increase my array by 26% and have more energy now, and for the next 25 years.

                  And that does not include the costs for monitoring/communications/internet equipment and web-subscription (if required) that comes "free" with my central inverter.

                  This does not even take into account the costs/issues involved with failures and failure rates of distributed vs centralized equipment and the costs for R&R (labor costs for replacing a single failed distributed inverter are probably the same, or slightly more, vs the average central inverter). Obviously, the part cost is less for the distributed inverter. But the central inverter is probably reparable (depot repair) vs the (most likely) sealed and smaller/less costly distributed inverter (makes sense to repair a large inverter--smaller inverter may make no sense to repair).

                  I am not saying that distributed inverters never make sense--it is just that they don't seem to make sense (at this time) for a "vanilla" installation. If the 25% "extra" power is important--then spend the 25% on more solar panels (and inverter capacity) now and get the extra power now.

                  I don't have the experience with old/new panels to know if aging is an issue or not regarding MPPT. The physics would seem to indicate that Imp and Vmp for Silicon panels between lots and manufacturers would be pretty stable (same materials, same physics, mass production of silicon ingots from generic vendors around the world).

                  Perhaps Solar Guppy (and others with experience) can expand on their experiences with old panels... I believe SG said before that he has not seen any measurable degrading of crystalline silicon panels over 10-20 years (excluding mfg and physical defects). And he has built a pair of "balanced" arrays for his GT inverter testing (down to 0.xx% matching of power output)--that could not be done if there was a constant shifting of I*V curves (and these were not small 1,000 watt arrays).

                  And even the vendors guarantee a minimum of 80% of panel output over 20-25 years (and folks expect 40 year panel life).

                  I just don't see the 25% distributed inverter improvement (as arrays may age) as physically possible within the standard lifetime of the arrays or that makes it worth spending 25% more for installation hardware to possibly prevent a non-problem in the first place.

                  -Bill

                  PS: If somebody has a large array and is concerned about finding a non-performing panel in the array... It would probably best addressed by purchasing a good quality DC current clamp and measuring the array's parallel string currents. A quick and easy test, and if done at the same time--sun/cloud/temperature conditions drop out because all panels are at the same test conditions.

                  If any one string does not match the rest--then can dig into the details and see what the problem maybe.
                  Last edited by BB.; May 26th, 2009, 17:17. Reason: Add note about installation price error.
                  20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

                  Comment


                  • #10

                    Re: Enphase Microinverters - Cost Effective?

                    Re: Enphase Microinverters - Cost Effective?

                    I second Bills thoughts, I think this is a solution in search for a problem.

                    I have about the same experience in number of systems, I have never once seen panels degrade for age if they are Si panels, now Asi is another beast and should be avoided at all costs. Si Panels are warranted for about 0.5%/year and in 8 years I have never been able to see any change in production.

                    I have seen a fair number of panel failures, its very obvious when a panel has issues, both physically and in the production. Most large systems are multi-string and multi inverter, its pretty straight forward to check production

                    Purely from a statistics view, using micro-inverters with life that might be measured in years ( yet unproven that they can last that long ) welded onto a panel that has a life measured in decades ( very proven ) is a very poor choice, that's a fact that no amount of PR or posting on message boards will overcome.

                    One point that most people don't realize or know about is the fill-factor of solar panels and just how wide the voltage variance can be for stringed panels and still get very close to the rated power ... if one panel is slightly higher current, it will run matched to the lower current one BUT will run at a slightly higher voltage, offsetting the loss to almost nothing.

                    For a large single string, there is zero advantage to micro-inverters, and shading or panel that fail the bypass diodes kick in and the string runs at a slight lower vmp ... the inverters already have mppt, so its all automatic.

                    Also, the reported efficiency's and hoped for reliability are very deceiving on the micro units as the reported numbers are for 25C, not the typical 65C that is the temperature on the back of a panel on the roof. Add in a typical 35C rise for an operating unit and the internals are running over 100C ... very close to the limits of the technology of the Fets and inductors

                    A properly designed central inverter system is lower cost, will have much longer life for the inverter(s) and when an electrical part fails be it a panel or inverter can be replaced individually.

                    In the case with micro inverters, UL requires the manufacture certify the solar-panel/inverter pair and require they be replaced as such, so now your making ALL the panels have the failure rate of the micro inverter, so instead of 25+ year life all the panels have 5+ year life before needed to be replaced ... pretty dumb idea

                    So in efficiency, they are not better, in cost, they lose, in allowing multi source for panels or electronics they don't win and any what ifs to show they are better are so hypothetical what-if that any central inverter system could perform they same if designed optimally.
                    Last edited by Solar Guppy; May 24th, 2009, 15:30.
                    www.solar-guppy.com
                    sigpic

                    Comment


                    • #11

                      Re: Enphase Microinverters - Cost Effective?

                      Re: Enphase Microinverters - Cost Effective?

                      Originally posted by Pgovetom View Post
                      The VI=maximum equation multiplies V x I to get the maximum power. That means the maximum power is exponentially a function of the voltage and current. That means a 2% change in voltage or current pushes a 4 times ( 2 squared) power change. A 5% change in V or I becomes a 25% change in power output. That is nasty!
                      That's not quite an accurate representation of the IV curve of a solar cell. P=VxI is a first order equation, so a 2% change in V or I is going to result in a 2% change in P. In any case, the VI relationship of a solar cell isn't linear -- it's the typical Shockley equation for a diode. Any aging-related factors are going to have insignificant impact on that function, and most quality modules are going to be fairly well matched to begin with. So I think it's a waste of money to spend $200 in an effort to ensure that the modules are MPPs are all matched as they age.

                      For arrays they see a lot of shade for much of the day, then individial tracking makes sense. But right now I think that the reliability, cost, and efficiency issues outweigh the potential benefits. In many installations, you may well end up with less power than you'd have otherwise. Someone correct me if I'm wrong, but I think I read that the push for individual module tracking comes from the tax credit end of things. Arrays that experience too much shade may not be eligible? Or something to that effect.

                      Marc
                      Rogue Power Technologies -- http://www.roguepowertech.com

                      Comment


                      • #12

                        Re: Enphase Microinverters - Cost Effective?

                        Re: Enphase Microinverters - Cost Effective?

                        Here is a data sheet for a crystalline silicon panel that includes a couple of graphs/curves that you don't see on most specifications (however, other Si products should behave pretty much identically).

                        For a panel with Vmp~35 volts--The 5% loss of power can be estimated from the graph to be:

                        +/- 2.5 volts at 1,000 Watts/sqmtr
                        +3/-8 volts at 600 Watts/sqmtr

                        So, if the panels are within +/- ~5%-10% of "ideal" Vmp--your output will be within 5% of optimum...

                        That is a pretty big variation in the IV curve that gives a relatively small Pmp change.

                        To get a 25% power improvement--the "micro inverter" would have to do better than:

                        Vmp= outside of 42-25 volts => ~25% power reduction (Vmp=35 volts)

                        Vmp does not (appear) to need to be to the nearest 1/10th of a volt... Large swings in "computed" Vmp do not make a very large change in the optimum Pmp output.

                        And if is one panel in a string of 10--then it would contribute only a 2.5% overall loss of power (assuming 25% out of spec.). Again, no way a distributed inverter can claim an overall 25% possible system improvement.

                        -Bill
                        20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

                        Comment


                        • #13

                          Re: Enphase Microinverters - Cost Effective?

                          Re: Enphase Microinverters - Cost Effective?

                          Again, BB is right on, I have patents in Mppt and understand the technology better than most. While in my work strive to get Mppt tracking to as close to 100% as possible, your measuring things in single watts or less when your withing 1 volt of the true vmp ... this is another way of expressing fill factor

                          The Mirco-Inverter technology is being pushed by a company that wants to make sales, they bet a big chunk of money on this, regardless if there is a problem to solved or not. I have no coin in this, but I see this as just a modern Trace Micro Sine ( same thing came out in 1998 ), which became the NTT unit and now EnPhase

                          The only thing I really could add to this is that very large systems that parallel 2 more more strings together for single channel mppt can get into less efficient harvest modes with panel shading or other non-optimum configurations. The solution is already here for that and its multi-channel input inverters, which I believe the Kaco brand does and I would expect this to become more main stream with time.

                          A more recent change is the solar panel Manufactures keeping making panels bigger as well, you can now get 5kW of panels into a single string and not have any shading or failure concerns.
                          www.solar-guppy.com
                          sigpic

                          Comment


                          • #14

                            Re: Enphase Microinverters - Cost Effective?

                            Re: Enphase Microinverters - Cost Effective?

                            It is an oversimplification to just call it a second order effect but when the VI, VP, IP, curves are examined and simulated in a string that is paralleled with other strings, the impact can be far worse than suggested here.

                            That's not quite an accurate representation of the IV curve of a solar cell. P=VxI is a first order equation, so a 2% change in V or I is going to result in a 2% change in P. In any case, the VI relationship of a solar cell isn't linear -- it's the typical Shockley equation for a diode. Any aging-related factors are going to have insignificant impact on that function, and most quality modules are going to be fairly well matched to begin with. So I think it's a waste of money to spend $200 in an effort to ensure that the modules are MPPs are all matched as they age.

                            For arrays they see a lot of shade for much of the day, then individial tracking makes sense. But right now I think that the reliability, cost, and efficiency issues outweigh the potential benefits. In many installations, you may well end up with less power than you'd have otherwise. Someone correct me if I'm wrong, but I think I read that the push for individual module tracking comes from the tax credit end of things. Arrays that experience too much shade may not be eligible? Or something to that effect.
                            Some useful information can be found in a paper by Act Solar who was purchased by National Semiconductor.

                            http://www.actsolar.com/files/tr1-ac...management.pdf

                            and the Solar Magic product they are pushing:

                            http://www.solarmagic.com/vcm/SolarM...c_Brochure.pdf

                            The IEEE is full of papers that have modeled and analyzed the real effects of a panel imbalanced and how it can push the other panels off their MPP and the effect isn't just one panel but its impact on a string or even parallel strings.

                            Shade is the worst problem but in the real world there are:

                            Shade
                            Clouds
                            Sunrise/sunset effects
                            Smog
                            Smoke
                            Failed panels
                            Partially failed panels
                            Dirty panels
                            Bird poop on panels
                            Dust
                            A replacement panel after the original process and panel model is discontinued with different curves
                            New technology add-ons or replacements 5-10 years down stream

                            I'm not promoting Enphase as putting 200watt DC-AC converters under panel arrays is a nasty reliability, access, lost communication sounds like a nightmare. In addition they have to try and match all panels power, MPP and voltage levels which is probably why they have 3 models at 3 power levels and their list of compatible panels is lousy so far. I can see problems with running that many parallel inverters at 120/N/120 on 3 wires back over long wires ( like mine) and reliability of one taking them all out.

                            I prefer the CHEAP MPP matching panel converter that only dissipates the difference and not the total power with a wireless Zigbee or modem performance monitoring link ONLY.

                            Others like National are promoting very simple MPP tracking high efficiency DC-DC converters with a simple communication protocols to find problems but that too has reliability and other issues. If they could build a MPP tracking DC-DC at 98% per panels ( or built into panels) for $50 each that were bullet proof, that would do the trick and leave the centralized DC-AC inverter alone. Just optimize each panel and get it to report any problems CHEAP would be nice.

                            Another interesting set of slides

                            http://www.univ-lehavre.fr/recherche...pe/dedonck.pdf

                            Comment


                            • #15

                              Re: Enphase Microinverters - Cost Effective?

                              Re: Enphase Microinverters - Cost Effective?

                              From the link you provided comparing Module Integrated Converters "MIC" and a central inverter, 112x 150 watt panels:

                              http://www.univ-lehavre.fr/recherche...pe/dedonck.pdf

                              Strings with Modules unsorted (from Page 21 of presentation):

                              Central Inverter / Dist. Inverter = 99.467%

                              Strings with Modules sorted for Imp current:

                              Central Inverter / Dist. Inverter = 99.868%

                              So--we are left with ~0.5% maximum power difference between central inverter and a MIC type string... Not a 25% improvement on typical string (no shading).

                              The same presentation offers a design with MIC that is "shade tolerant"--but, if you follow Solar Guppy's suggestion of large panels on a single string inverter--you would get the same results (assuming Vmp with shading is >Vmin for the inverter). For my 10 panel strings, that would be Vmp 350v - 200 v = 150 volts of "head room".

                              Divide the 150 volt/Vmp 35 volts = 4+ panels worth (realistically, probably closer to 3 panels / 30% shade tolerance) of shading (out of ten)... I would be hard pressed to justify an array that spent the day in 40% shade...

                              From the rest of the Solar Magic links, they are mixing managing megawatt systems (fault isolation, "large scale" array in field differences using very large central converters) with problems of "small" home systems (chimney, trees).

                              In general--taking a large system and doubling the components and electrical connections (1 solar magic per panel? Plus a blocking diode per string)--usually does not make for a "simpler to mange" / more cost effective system.

                              Just measure the current per string (shunt, dc current probe, "blocking diode" voltage drop, etc.)--and you got fault isolation down to a dozen or so panels.

                              From the first link:

                              http://www.actsolar.com/files/tr1-ac...management.pdf

                              However, when the performance of an entire PV system is analyzed in year 20, itís common to see less than 50% of the original power output for the system as a whole[2]. This is a result of primary losses in a few excessively worn components combined with the fact that accumulated variations across all of the system components has a magnifed penalty on system output due to secondary losses at the string level and tertiary losses at the array level. Not only is power lost from worn or broken components, but substantial power is lost from the remainder of a PV systemís components since many can no longer achieve maximum power.

                              Footnote: [2] Bridges National Monument Analysis of a 20 year old solar system paper
                              Note that the "common to see...50%" degradation is based on one installation. And that the is the "...result of primary losses in a few excessively worn components..."

                              From earlier in the paper:

                              Analyses of older PV systems have consistently shown that individual solar panels have low degradation rates and a long lifespan. In year 20 most of the solar cells still produce 90% of the power they did in year one and will likely remain viable for another 20 years[1]; 59% of the modules showed a variation less than 10%, 35% showed variation of 10-20% and only 6% showed variation greater than 20%.

                              [1]
                              Study of a 20-year-old PV plant (MTBF project) - SUPSI
                              Again, a study of one (panel types? Mfg? The 6% >20% under spec... Was there visible damage?

                              In the end--it sounds like the "major" faults typically result in "gross" failures (broken panel, bad wiring, ground faults) etc... The panels (in general) resulted in some power loss (but would the Solar Magic module have "fixed" any of those issues?)--the the other problems would have been found/isolated to the string level with standard metering (or optional instrumentation).

                              Again--there is very little to go on how the "solar magic" works or what it does (little boost converter? with communications?). And is it worth $x per watt to install them? And will they last the 20-40 years that panels will (when exposed to thermal cycling in extreme heat/cold)?

                              I don't know--but I would want somebody to prove it to me first (take one of those megawatt plants and install the Solar Magic--without any other "fixes/repairs") and show me the results.

                              -Bill
                              20x BP 4175B panels (replacement) + Xantrex GT 3.3 inverter for 3kW Grid Tied system + Honda eu2000i Inverter/Generator for emergency backup.

                              Comment

                              Working...
                              X