MPPT and MPP problem Questions

I am curious if anyone is aware of good solutions to the MPP problem.

I see it having two basic forks:

1) Some mechanism to balance the differences in VI MPP characteristics of any one panel whether it be due to manufacturing, aging, shadows, dusty panel, dirty area of a panel, failure, wire resistance, partial fault etc... That is cost/power/efficiency effective and reliable since overall cost/power performance is the goal.

2) Identifying the panel(s) which need attention ( e.g. cleaning, problem correction or replacement) EASILY so the cost, effort or lost full power time is minized.

The brute force approach is taken by Enphase Energy by isolating every solar panel with its own DC to AC inverter which accomplishes both goals. It can track the MPP for it's panel plus it can report if the panel has a problem. The downside is cost and reliability. I have 56 210W panels in my system and I would need 56 210 watt inverters. My 2 SB6000 inverters cost about $7200.

Just to match the material cost of my 2 inverters, the Enphase approach leaves them a $129 budget per inverter to match my 2 SMA units. Their approach has grounding and and potential redundancy advantages but installing those inverters on every panel wouldn't be simple either. Since the temperature under the panels can vary from -40F to probably near 200F depending on whether its Arizona or Montana or night versus day or summer versus winter, power inverters that operate over such hostile temperature ranges is DIFFICULT to do inexpensively. Then the inverters must match all popular panels ranging from 150W to 225W and voltages from the low 20's to 60 volts. Then they must adaptively track the MPP and produce a tracking phase locked 60HZ 120/240V with a very reliable output.

The Enphase also provides a very nice per panel performance monitoring available via web access that covers my second fork.

Ok..My question:

Are there any other real products out their that tackle this problem in other ways?

Are there MPP optimizers that are put across each panel but are just DC-DC tracking adjusters that are cheap and simpler and efficient? If the efficiency is say 9% and then the main inverter is 95%, that in tandem almost ruins whatever gain is achieved.

Are there any per panel performance monitoring ( VI monitoring) available?

Are there any new PV cell designs coming out where this MPP problem is improved?

Are there any Panels with there own performance monitoring?

Any input appreciated!

Tom

Comments

  • lorelec
    lorelec Solar Expert Posts: 200 ✭✭
    Re: MPPT and MPP problem Questions

    National Semiconductor has been touting their "Solarmagic" optimizers for some time now, but as far as I know have yet to release them to production. They have supposedly been demonstrating them at several trade shows and have issued a number press releases, but so far there is no hard data that is easily accessible to the public. I've heard through the grapevine that they intend on retailing them for $200 apiece (and you'd need one for each PV module), and guaranteeing them for 20 years. They have been very secretive about the design details, but I suspect all they do is gradually route power around a PV that is producing less than its optimal output -- so it doesn't bring down the rest of the string. With a 20-year warranty and given the hot environments they would be expected to operate in, I can't imagine that there's too much to them. Certainly no electrolytic capacitors. They also require a blocking diode to be used with each string, so that's going to bring down efficiency there some, too (and increase cost).

    Marc
  • boB
    boB Solar Expert Posts: 1,030 ✭✭✭✭
    Re: MPPT and MPP problem Questions

    I heard back from National on these when I tried to buy one online.
    NS said that AEE and Real Goods have been signed up to sell them. Didn't have a price, but if it is going to be $200, then that will be too high I would think.

    I have a feeling that NS doesn't know this industry well enough to do a proper job.

    I saw an online video that showed what they do. They help with individual module partial shading and the reduced string voltage compared with the other strings that are un-shaded. Sounds pretty cool actually, but I have no idea of their "practical" usefulness.
    They may be a solution looking for a problem, but we do know that "shade happens", which happens to be their slogan.

    boB :D
  • Pgovetom
    Pgovetom Solar Expert Posts: 30
    Re: MPPT and MPP problem Questions

    There have been a lot of research papers with designs that use a MPP tracking algorithm and some form of DC-DC converter topology to adjust the VI apparent characteristics of the module so they all sit close to their own maximum MPP point. In the simplest terms, you can see that any series string of panels MUST pass the exact current through all panels since they are in series. That means each panel will adapt its voltage to match that current. That point may not however match the ideal MPP VI point. By adding a boost/buck DC-DC converter that can adjust the Voltage up or down so all the panels are at or close to the best MPP point, each panel can be made to seem to operate at the ideal MPP. That way all the panels are giving the best power they can for their conditions and not forcing other panels away form their ideal. The only problem is cost, reliability under a COLD or very HOT panel and the efficiency of the converter.

    If the price was $50, it was military reliable and its efficiency was 98-99% and it had performance monitoring back on the DC link it would be a miracle device. Then the normal series and parallel combiner architecture/topology plus a centralized inverter at say 96-97% would be fabulous. The combined efficiency og say 98% x 96% would drop the aggregate to about 94% but that 2-3% would easily be made up by the MPP ideal loses versus the MPP-lousy match losses.

    Unfortunately at $200 and I bet 95% efficiency, compelxity, reliability, the cost plus efficiency gain would offset the MPP optimization gain.

    For installations with lots of shading issues or panel cleanliness issues, it might pay off, but installations with no shading in clean environments, it would come down to how bad 10 years of life effected MPP mismatch losses.

    Its a tricky problem unless someone like National can get the price down, efficiency way up and reliability excellent and provide good Performance monitoring of voltage/current history, MPP tracking performance quality ( how close to ideal over time and varied problems).

    The Enphase might be nice if they can do the same but building an MPP tracking DC-AC phase locked to 60 HZ, highly reliable from -40F-250F can be done in the $100-$125 range. I doubt that and the grounding and other installation gains are one time costs. That is tricky too.

    Anybody heard of anyone taking any other approaches?

    Has anyone considered putting MPP optimizers in panels and a communication standard so panels can cheaply report their V/I history. At least the Performance monitoring makes it easy to identify where the problem panel is located easily and a common standard would be NICE.

    I see finding a really good MPP tracking and performance monitoring absolutely essential to Solar PV technology. This alone can gain up to 25% and possibly more in old systems with shading and cleaning problems.

    thanks

    Tom
  • BB.
    BB. Super Moderators, Administrators Posts: 33,422 admin
    Re: MPPT and MPP problem Questions

    Solar Guppy is the guy (and a few others here) that can give you exact efficiency gains and losses between various MPPT algorithms (and you would probably have to sign a non-disclosure).

    In the end--the difference between a well designed PWM system and a well designed MPPT system/charge controller is, at best ~25% added collection on very cold days (i.e., snow on the ground).

    Otherwise, you are probably looking at 10% differential for the "average" system...

    Remember that MPPT controllers use a bit more power to operate (Tare Losses). And that there are other design issues that also take a percent here and percent there (small diameter/long runs of wiring, less than optimum angles, morning fog and east facing panels/evening thunderstorms and westward facing panels, panels mounted very close to the roof running hot, dirty panels, fixed panels vs panels with 3 or more manual tilt positions, not using 1 and 2 axis trackers, obstructions to sun, plantings, trees, new constructions)--all add up to "losses".

    And, in the end--while it is nice to design/tune a system to 100% of maximum possible output--if the time and money was extensive--it may have been better just to add a panel or two to the string and get the additional output for the same price (or even less).

    You can play with a "derating" calculator and see how all of those 1%, 3%, 5% losses all multiplied together end up with a 77% efficiency for the "average" grid tied system (from panel rating to useful AC power).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • Pgovetom
    Pgovetom Solar Expert Posts: 30
    Re: MPPT and MPP problem Questions

    The problem with taking the "add a few panels approach" is that as a system ages, a failure or panel problem or other difficult to troubleshoot and detect problem will still happen. A series parallel array of panels will almost collapse in power output if the MPP balance point of just one or a few panels is bad. Its just like batteries. Ever wondered why one battery takes down a flashlight when there are 3 new ones. Its hard to find the problem panel without getting under the array with a voltmeter and looking for the offender ( if its bad enough to spot).

    I realize with brand new wiring and panels, the MPP balance will be pretty good ( without shade issues) but after 5 or 10 years, all the panels will age differently and the array could easily drop to 60% power with a scattering of MPP point drifting and problem panels sprinkled around.

    How do you keep the array operating near full power even with a few added panels after 5 years when you have 50 panels all aging differently and some going really weird. Only a per panel MPP optimizing plus some kind of per panel performance monitor will keep things going close to full power and allow the problem to be found and fixed/replaced quickly after 5-10 years.

    When a brand new panel that is similar to the old ones is used to replace a bad panel after 5 years and the original is out of production, that brand new panel with a DIFFERENT MPP point will then also mess things up. Same problem same fix.

    Tom
  • BB.
    BB. Super Moderators, Administrators Posts: 33,422 admin
    Re: MPPT and MPP problem Questions

    The MPPT problem needs to be addressed like every other engineering problem.

    If you have two strings of ten panels each (series parallel) -- like my GT system. There is only for things I can watch.

    1. Total system watts (at any one time)
    2. Total system watt*hours
    3. Panel Voltage
    4. Panel Current

    I don't log anything (other than paper/pencil of 1&2--and throw into a spread sheet to compare year over year usage--and how much power my home, itself, consumed).

    If there is any one panel that fails (shorts/open)--then I will loose ~50% of may system's output (even if the panel fails shorted, I will suffer a pretty large drop in power).

    I will never "see" a 10% or less drop in power due to aging/etc... If I had "Enphase" modules (or voltage/current measurements for all 20 panels)--I would probably find 5% or less variations in output (mostly, because I have 20 units to compare in the same sun/mount--so it is easy to find differences).

    Solar Guppy has said before, that he has not observed any aging in old (good quality silicon) solar panels over 20+ years.

    Yes, there are failures--but generally those are obvious.

    The "cheap" thin film type panels do not last as long (cheap mfg or panel physics?) and, they degrade ~20% in the first 6 months. And, from what I could find--they do not seem to degrade much faster than standard crystalline solar panels after that "burn in" period (other then when thin film panels fail around 10+ years -- instead of the 25-40 years of a good quality crystalline panels).

    Finding "new" panels that match the Imp/Vmp (series or parallel) of an existing string is a problem--but given that new panels are 1/2 or 1/3 the price of the panels they are replacing--that is not a huge issue if the "new panel" does not have optimum MPPT tracking with the rest of the string (or if you have to add a "new panel" to several strings to match series parallel stings into a MPPT controller).

    In the end--I like the simple solutions for many problems.

    Adding complexity usually adds too much cost, hurts overall reliability, and ends up trying to solve problems that 80% of the systems never see in the first place.

    For certain applications--perhaps using Enphase would help--but it certainly is not a magic bullet at this time.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • Pgovetom
    Pgovetom Solar Expert Posts: 30
    Re: MPPT and MPP problem Questions

    Hi...Thanks for thoughts!
    Otherwise, you are probably looking at 10% differential for the "average" system...

    Is this a guess at the average with impairments ( some shade, some clouds, some bird droppings, some dirt, etc..) or simply the mismatch due to brand new panels being off their MPP. Or another way of putting it -- Would some ideal cheap MPP optimizer that was installed for each panel ( 100% match/efficiency each panel with boost/buck with an MPP tracking input/reference) would add 10% more output to the average system ( medium to large -- very bad through very good)

    Solar Guppy has said before, that he has not observed any aging in old (good quality silicon) solar panels over 20+ years.

    That is just one technology and one manufacturer tested. Betting on that would not be good unless it was the same as monitored.

    Solar Guppy is the guy (and a few others here) that can give you exact efficiency gains and losses between various MPPT algorithms (and you would probably have to sign a non-disclosure).

    There seem to be many IEEE papers on algorithms but I've yet to see a good data set that shows complete parallel combined and series stringed topologies with varied impairments and the total impact. For example, if one could model a large series string and multiple parallel combined with plug-in VI models for various panel technologies and manufacturer processes and add impairments ( shade with parameters like background irradiance, blocked irradiance, duration etc..), it would be interesting to try various real world topologies. I have 14 series and 2 combined and I could enter my panel model see what the instantaneous, per day and yearly impacts the impairments I expect.

    Has anyone seen a software package or other simulation tool like MatLab where this has been done?
    Yes, there are failures--but generally those are obvious.

    They might not be so obvious finding them. I have a ground mount 2 x14 series arrangement ( not uncommon) and locating a bad panel if I couldn't see the problem ( dirt or bird droppings etc..) would not be simple. A communication device that monitored the VI per panel would be handy.

    In the end--I like the simple solutions for many problems.

    AGREED. Plus any add-on should always fail such that it does not impair the system ( as Enphase would). What does that Solar Magic do?

    T:confused:
  • BB.
    BB. Super Moderators, Administrators Posts: 33,422 admin
    Re: MPPT and MPP problem Questions

    Regarding the Solar Magic--I don't know--I have not seen a detailed explanation of what it does (just marketing dancing around the functions).

    It sounds like each module is a digital power supply with either (or both) boost / buck (raise voltage, lower voltage) based on the overall maximum power output of the panel and the overall power output of the string.

    Sounds like, it adjusts for a panel's Vmp*Imp curve--and it monitors the current in the series string and can adjust that for more current too (for example, two parallel strings of 10 series panels each; one in shade, then the other 9 panel Solar Magic Modules must "boost" the total string voltage to equal the unimpaired 10 string in parallel--otherwise, the string with the "dark module" would drop most of its current output due to its lower Vmp with the "missing" panel's output).

    If the Solar Magic Module could do something like I described--it would put it in a similar function/category with the Enphase. However, I would wonder if the Solar Magic ends up being less efficient because it would add another ~5% loss because of the the additional "converters" (Solar Magic times Central GT Inverter losses).

    The above is just all guess-work... I have no idea if I am correct or not.

    There was the link to an Enphase type unit (no idea of brand) that compared a distributed inverter vs central inverter topology showed a ~0.5% improvement with the distributed inverter when using "out of the box" new panels and their own manufacturing variations.:
    BB. wrote: »
    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).

    As Solar Guppy said in the above "Enphase" thread--with very large solar panels of today (250 watt)--you can design single series string central GT inverter type systems. In those cases, dirt/bird-droppings would be managed exactly the same as with distributed or central inverter--each would adjust the I*V operating point for optimum power output... And neither technology could "fix" the lost power from the shading event itself (i.e., a panel simply produces less power with shade--no technology can recover that lost power until the shadow is removed).

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • lorelec
    lorelec Solar Expert Posts: 200 ✭✭
    Re: MPPT and MPP problem Questions

    National just issued a press release today (May 27); apparently they are now shipping their Solarmagic devices. MSRP is $199...too expensive in my opinion. Will be interesting to see how they work out "in the real world."

    Marc
  • boB
    boB Solar Expert Posts: 1,030 ✭✭✭✭
    Re: MPPT and MPP problem Questions
    lorelec wrote: »
    National just issued a press release today (May 27); apparently they are now shipping their Solarmagic devices. MSRP is $199...too expensive in my opinion. Will be interesting to see how they work out "in the real world."

    Marc


    Yep, too high $$ in my opinion also. Maybe if you could add one S.M. to one series string that had some partial shading on one or two modules, then that might be OK. One per module on the shaded modules ???..... Probably not worth it, but I haven't done the math.

    Also, for a shaded series string of modules in parallel with another string that is un-shaded, Mr. Guppy pointed out another manufacturer that makes a unit for that. Somewhere on this forum and not too long ago of a thread.

    S.M. is evidently being sold by AEE and Real Goods, both in California.

    boB
  • niel
    niel Solar Expert Posts: 10,300 ✭✭✭✭
    Re: MPPT and MPP problem Questions

    "If there is any one panel that fails (shorts/open)--then I will loose ~50% of may system's output (even if the panel fails shorted, I will suffer a pretty large drop in power)."
    bill,
    if you have bypass diodes or if the pvs have them builtin there shouldn't be a total loss on that string to warrant a 50% loss of power. prove it to yourself and shade 1 pv.
  • BB.
    BB. Super Moderators, Administrators Posts: 33,422 admin
    Re: MPPT and MPP problem Questions

    Mine are on top of a 2nd story roof--Not real easy to get up there and do the test... Next time I have the ladder out--I will give it a try.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • Solar Guppy
    Solar Guppy Solar Expert Posts: 1,989 ✭✭✭
    Re: MPPT and MPP problem Questions

    on a large string, losing one panel isn't that bad ( bypass diodes kick in ) ... from the work I have done, its was less than 10% hit ... more panels in the string the better as the remaining panels can run a volt or two higher with reduced current also the parallel string can run at a lower voltage and higher current ... the blend is nothing near 50% hit seen on the many configurations I've run over the year
  • BB.
    BB. Super Moderators, Administrators Posts: 33,422 admin
    Re: MPPT and MPP problem Questions

    OK--I will admit to being a bit "pessimistic" on how much shadows can affect arrays...

    A major portion of that is I want new users to realize that solar panels only generate power when they are in full sun and virtually nothing in the shade. And definitely nothing in starlight (yes, there is a web article where they still claim solar panels will work under starlight--debunked here).

    For large arrays (like mine with 2x strings by 10 panels each)--yes, I will admit that one shaded panel will probably only drop the ~10% that Solar Guppy says vs the 50% I warn everyone about.

    But, also, as SG says--there are a lot of variables in how solar systems are designed and installed. If the system is build with "high voltage panels" and only 5 panels in a string--the drop in parallel string output will be much more severe with one panel shaded. Off-Grid systems with similar series/parallel strings will be similarly affected by one panel shading (larger drop-offs in power output).

    So, while I promise Niel I will do a shade test on my array next time I am on the roof with a ladder--I will still want to ensure that people do their best to install an optimal system--and ask questions rather than just accept that solar panels are magic and can produce large amounts of power under any condition.

    -Bill
    Near San Francisco California: 3.5kWatt Grid Tied Solar power system+small backup genset
  • niel
    niel Solar Expert Posts: 10,300 ✭✭✭✭
    Re: MPPT and MPP problem Questions

    ok, but if you fall, don't blame me.
    hey, what are you talking about that the pvs don't work on starlight? they do to when they're aimed at one that is 93 million miles away.;)

    i know, shut up niel.:p