Bypassing faulted panels in long strings

A friend of mine and I were discussing automated strategies for dealing with faulted strings in large grid tied installations and he referenced some sort of technology whereby you can bypass a single problem panel in a string and continue generating power with the other panels and attached complete strings. I didn't think this sounded right, unless you're using small inverters and have each string independently feeding its own inverter.
Am I missing something? Wouldn't you effectively disable the whole string when you bypass the bad panel if the other strings are at full power? Indeed, if it weren't for the blocking diodes wouldn't you have yourself a serious problem with strings of different lengths connected to each other?
Personally, I'm off grid and manually manage my three little strings so this isn't of particular utility to me, but the domain of the problem is very interesting and I'd like to learn more about it.
Am I missing something? Wouldn't you effectively disable the whole string when you bypass the bad panel if the other strings are at full power? Indeed, if it weren't for the blocking diodes wouldn't you have yourself a serious problem with strings of different lengths connected to each other?
Personally, I'm off grid and manually manage my three little strings so this isn't of particular utility to me, but the domain of the problem is very interesting and I'd like to learn more about it.
Comments
You reasoning is correct for parallel strings. The single panel bypass diode will continue to be reverse biased by the other parallel strings. Little current will flow through that string.
This is the reason for the move to "smart modules" that use a converter module to manage and optimize each module individually.
I'd replace the faulty module.
If it's because of a shaded module, a chainsaw may be necessary.
boB
"Smart modules"?!? What are these and how do they work?
I think that is the idea behind Enphase inverters. If one panel goes down, the string continues to opperate without any decreas in production (save the one pane).
Tony
I think he was talking about these:
http://news.cnet.com/8301-11128_3-20006418-54.html
$0.12/watt? Hmm.
You can read some discussions here:
Solar Edge [from: help me design the most efficient system]
National Semiconductor - Solar Magic Devices ?
National Semi "Solar Magic"
-Bill
They are yet another option on to where to partition up the system.
These are MPPT control and boost to 350 to 500 vdc. The central inverter is just a transfomerless PWM chopper to convert the high voltage DC to AC.
Panel drops out of contribution when its mini-module cannot create at least 350vdc output or matching the voltage produced by other panels based on loading from central inverter. Central inverter will adjusts its load pull so that voltage never is allowed to drop below 350 vdc which is minimum voltage required to make 240 vac output.
It does have an advantage over the Enphase system in that it does not require the large electrolytic capacitors in the mini-inverter. The single phase AC power profile smoothing can be done in the central inverter with large caps that are more reliable and not subjected to the high temperatures under the PV panels.
High voltage DC is also more efficient at reducing wire loss then the Enphase 240 vac.
I do not work for, nor am I invested in Enphase, but I felt I have to give my opinion....
The 350 volts needed to make a central inverter work is exactly one place that Enphase has an advantage. An Enphase setup can, and does, get one watt out when one watt is all there is to have. With a central inverter you get nothing up until the panels "make the minimum".
You say "large electrolytic caps" in the Enphase? They are quite small when compared to the ones in a central inverter, then right after you imply large caps are not as reliable, you go on to say the large ones in a central inverter are more reliable? I'm confused.
You'll be hard pressed to sell me that I will save any money at all by running high voltage DC the fifty feet from roof to inverter that most systems run. In fact, don't most use 10 ga wire for each string? Enphase uses common 12, AND once it gets in the roof it does not required to be in conduit, another savings.
When you toss in the added cost of these "smart modules" why wasn't Enphase used instead? No need for smart modules on an off grid system there doesn't seem much need as once the batteries are charged the rest is lost, and with a grid tie system I think you'd be better off using micro inverters. The smart module people always worry about the heat effecting the Enphase, but forget to mention the smart module is in the same place enduring the same heat. Need I mention they are both built to take it? If either one fails you are headed to the roof to fix it.
Enphase is not the best smart-module solution to my mind.
"The 350 volts needed to make a central inverter work is exactly one place that Enphase has an advantage."
- All distributed inverters (SolarEdge, Tigo, Solar Magic...) have this same advantage.
"You say "large electrolytic caps" in the Enphase? They are quite small when compared to the ones in a central inverter, then right after you imply large caps are not as reliable, you go on to say the large ones in a central inverter are more reliable? I'm confused."
- Electrolytic caps are the least reliable type of capacitor even when properly designed. Notwithstanding Enphase's assertions, putting them in the harsh roof-top environment is a design error which they have recently corrected but doesn't fix the 200,000 they already made. Central inverters use much larger ones yes, but in a cooler location where they can be serviced much easier. The bigger the caps are, the more reliable they are too.
"You'll be hard pressed to sell me that I will save any money at all by running high voltage DC the fifty feet from roof to inverter that most systems run. In fact, don't most use 10 ga wire for each string? Enphase uses common 12, AND once it gets in the roof it does not required to be in conduit, another savings."
- In the other distributed inverter systems, you don't have to wire the modules at all. They just plug together. Only wiring needed is the home run to the inverter. With SolarEdge for example, you can do strings of 25 modules with just one pair of wire coming off the roof. Plus, the Enphase wiring system is expensive.
"When you toss in the added cost of these "smart modules" why wasn't Enphase used instead?"
- The other module converters are really not added cost. In most cases, the additional power gained, offsets any additional cost. And in the case of SolarEdge, the DC-DC converter stage which all central inverters have - is just relocated to the PV modules where it is more effective leaving just the low cost DC- AC stage to be done centrally. The system is good value.
I know the Enphase system seems like such a simple elegant solution, and it is for small arrays, but people really ought to look into the other options and find the best one for their situation.
I am not promoting either system. Solar Magic is just a different partitioning of the system.
The Enphase inverter must also creat 350 vdc minimum before it can produce any output. Unless the panels are very poor quality with excessive leakage current a boost inverter will be able to produce the required minimum voltage for AC conversion as soon as there is any minimal illumination across the panel. Partial shading of a panel will have the same dropout on either system.
The major differences is where the energy storage capacitor is placed in the DC path and it potential reliability. Large energy storage capacity (not talking physical size) is necessary to filter out the pulsing sine^2 power profile of single phase AC. Putting it on the HV boosted DC has an advantage compared to putting the storage directly at the panel terminals that requires very little ripple voltage to keep the panel operating at maximum efficiency.
What has Enphase done differently on newer units? There is two choices to place the storage capacitor, on the input or output side of the DC-DC converter. To keep the size down, must use electrolytics which takes away the advantage on putting the storage on the HV side since they cannot tolerate high ripple voltage.
Back up & run me over again? Each panel and each Enphase inverter make a complete separate system. My panels are open circuit voltage of 36.7 volts DC. That is the highest DC voltage that any of my inverters can possibly get. That being said, I do not know what happens inside the inverter. For all I know it takes the 36 volts and makes it a million before it inverts it to AC, but, the point I am making is that my panels will in fact produce one watt of 240 V AC in early morning or late evening.
Your second paragraph is over my head.
Of course! Personally I put most home solar arrays in the small system category.
There are no strings in an Enphase system. All the outputs are in parallel.
Enphase must produce the 350 volts internally before it can push any power to AC grid. Just like the central system, just distributed (exposed) differently.
Enphase takes the output of a single module at <50VDC and converts it to 240VAC. Where does 350V (DC or AC?) come in to it?
The peaks in the sine wave. it's 240VAC RMS, not P-P
|| Midnight Classic 200 | 10, Evergreen 200w in a 160VOC array ||
|| VEC1093 12V Charger | Maha C401 aa/aaa Charger | SureSine | Sunsaver MPPT 15A
solar: http://tinyurl.com/LMR-Solar
gen: http://tinyurl.com/LMR-Lister ,
Actually it needs to produce a bit higher then 350 vdc to make the peak of the sinewave given the high end tolerance of the grid and losses for the PWM switcher and output filter.
Wow, great information. Thank you all. This makes perfect sense now, as my friend was talking about large systems that have to be remotely monitored (in other words: not a system on someone's house). Moving the DC-DC conversion to the individual modules and allowing for monitoring of the same seems like a no brainer at scale.
A guy I know who has done lots of small to medium scale installations with both Enphase and SMA inverters told me a year or so ago (so it may have changed since then) that the break point where central/string inverters are less expensive than micros is around 3-4kW.
Enphase to their credit, or at least in response to their critics, have redesigned their microinverters, replacing the electrolytic capacitors with some other type. No doubt takes up more pcb space and costs more but reliable. Ceramic or film caps go forever.
do you know what cut-off in serial numbers exists in transitioning from the old electrolytic type to what the newer design is?
There is really no alternative to electrolytics for Enphase unless they increase the box size by about 5 times. They require too much capacitance. Choices are aluminum or tantalum types. Hermetic tantalums would be an improvement but they are expensive.
They could drop the capacitance down and raise the rated working voltage on the caps but that would cause more ripple current on panel which would reduce efficiency.
Panel ripple current degradation would likely not show up in their specs for efficiency as they would rate it with inverter driven from a lab power supply not a PV panel.
There are very high K dielectric ceramics but they have their own set of problem. Their capacitance is a function of bias voltage and they are not reliable against humidity intrusion that causes leakage current.
Every other type of capacitor would be a large increase in physical size. Film caps at the required capacitance would be almost the size of a 12 oz beer can.
Ummmm, beeeerrrr. Is it 5 o'clock yet?
So far, I haven't discovered what capacitor is used in Enphase's new M215 microinverter. Can anybody take one apart and see?
The old ones used four 2200uF 63Vdc 105degC Nichicon UPW1J222MHD $.94 ea
In ceramic, this would take 880 of the largest 10uF/50V/$.109 Panasonic ECJ-4YF1H106Z SIZE 1210 (3.2 x 2.5mm)
In Tantalum, 400 of 22uF/50V/$.029 Vishay 595D226X9050R2T (7.2 x 6.0mm)