Looking for Knowledgable People on PV Inverters and Battery Augmented PV Systems

[TSLA Pilot]

I called Solar City and I thought for a moment that I was talking to Tesla

I spoke with a nice enthusiastic young gentleman who was perfectly happy to talk leasing a full house system for California. When we got to me being in Florida, wanting access the the battery back up, wanting information on the size and capability of the systems,,,, we ran out of run way. SC does not sell anything. They are still in pilot with the battery system and, from what I gathered, do not have enough batteries to sell systems even if they wanted. The person I spoke with had no idea if their battery was integrated at a higher than common 24V-48VDC level. I did learn that they are packaging someone else's inverter with their systems but he did not know who's inverter. Anyone out there with a SC installation that knows which inverter they are using?

Excellent thought process LolaCarChamp.

FYI: SolarCity isn't even operational in Florida--they're only in 14 states and I'd guess that FP&L's grip on the FL State legislature is quite firm. So nothing for Florida . . . sounds somewhat like NJ perhaps?

Regardless, in Texas we have a SCTY PV system (7.6 kW DC) on our roof; it's been there for almost 12 months, and we LOVE it. However, we HATE the fact that we're only getting 75% of our "push back" into the grid, once we reach 500 kWh of production and that happens every month, even with our MS.

In my quest for a solution, we found Ideal Power's site:

Ideal Power | Home

I'm so impressed I purchased about 1,400 shares a few days ago.

However, my concerns are:

1. IPWR is a very young company.
2. Are their massive efficiency gains supported by test results?
3. How much does their unique inverter cost?

My long-term goal is to capture the DC power from the PV system and shunt it to a Tesla "home" battery as well. Then any excess (beyond the MS's and home's usage/needs) was to go back to the grid. Then, when the MS returns home every night it would charge off of the DC battery first, then the grid, if additional power was still needed.

Overall, there should be a massive increase in efficiency as we're losing (ostensibly) about 15% going from the grid's 240V AC to the MS's DC battery. Even worse: I'm losing a similar (or larger?) amount of power going from the DC PV system, through our two Power One inverters, to convert the sun's power to AC/grid power, just to turn around and change it BACK to DC for the Tesla's battery!

This is borderline crazy, but it's an AC/DC world that has a lot of low-hanging fruit for efficiency gains.

I can easily see SCTY and/or Tesla buying an inverter company so that they can nip this problem in the bud; let's hope they beat us to it as I'd much prefer a turn-key solution.

Please keep us posted on how your project progresses.

- - - Updated - - -

Thank you for that. I was curious how PV systems handled too much capacity (home needs X, battery is full and panels producing X + Y). I guess making heat is one way to do it.

On our SCTY PV system, any grid power loss results in an immediate and automatic PV system shut down, ostensibly for Elec Co. lineman protection. As far as I know, there is no power nor heat being generated when the system if off line.

Frankly, I don't think any power is being generated if the system is being isolated, although I do wonder why that's the case. The initial install had the panels exposed to the sun and no "shocking" occurred . . . I'd guess that there's voltage generated with solar exposure, but with no connection there's no amps flowing.

In the long-term, I expect that the ability to isolate from the grid during a local power failure AND still have some home circuits powered up (while the sun is shining, at least), will be a major selling point for SolarCity and other PV installers. I can't wait for that day to arrive as the current situation is remarkably unsophisticated.

- - - Updated - - -

Oh, forgot to give you more details on the inverters SCTY installed. At least as of 12-months ago, they used:

www.power-one.com

We had to use two inverters with a 4 circuits total due to 4 different mounting angles on my roof.

Otherwise, I believe they try make a PV system work with a single inverter (for cost savings?) when the PV array will allow--one large set of panels, all on the same roof angle and without shading.

 

[JPP]

I called Solar City and I thought for a moment that I was talking to Tesla

I spoke with a nice enthusiastic young gentleman who was perfectly happy to talk leasing a full house system for California. When we got to me being in Florida, wanting access the the battery back up, wanting information on the size and capability of the systems,,,, we ran out of run way. SC does not sell anything. They are still in pilot with the battery system and, from what I gathered, do not have enough batteries to sell systems even if they wanted. The person I spoke with had no idea if their battery was integrated at a higher than common 24V-48VDC level. I did learn that they are packaging someone else's inverter with their systems but he did not know who's inverter. Anyone out there with a SC installation that knows which inverter they are using?

I'm in the SF Bay Area (SC's home base) with a 16kW PV system that uses 63 Canadian Solar panels and SolarEdge inverters. As we have an east-west panel install (due to the geometry of our house), our system has variable output based on time of day/sun angle. We have 4 strings (wired in serial) of panels with anywhere between 11 and 21 panels per string. Each of our panels has some kind of special SolarEdge module mounted on the back, that is able to track power throughput and 'bypass' an under performing panel/string to maintain optimum power output. Due to the special 'bypass modules', we have 4 single inverters (rather than double inverters). Each SolarEdge inverter talks wirelessly to a monitoring base station in our house, provided by SC, which uses an Ethernet connection to talk to the SC servers. I can monitor hourly/daily/monthly output (by string, not by individual panel), and SC does the same. I have a Power Purchase Agreement, so I do not own or 'lease' my system, but simply buy 16kW of power at an agreed upon $0.16/kW (with 3% annual increase). Yes, sounds expensive, but in NoCA PG&E charges $0.35/kW at summer peak, and we have a big house with a black shingle roof and use lots of AC in the summer. We have of course switched to Net Metering and TOU, and our best rates are at night (about $0.09/kw at night), so of course if I need to charge the S I do it on the timer from 12mn to 6 am. With the PPA, SCs goal is to generate about 70% of your power needs (averaged/annualized) but about 90% of your bill. They do not want to overgenerate, as PG&E buys back excess at a poor rate, and also because PG&E charges basic meter fees and hookup fees even if you net over-generate.

Under my PPA, SC must guarantee output and performance/uptime (since they would then need to 'buy' power from PG&E to cover any shortfall at 'retail' prices). My assumption is that SC purchases and installs cost effective technology, maybe not the best or most efficient panels or inverters, but over a time/$$ basis the quality and durability that meets their business model. It's not my problem/nickel if they need to replace panels or inverters, as they are in effect 'renting' my roof and they are just one of my utility providers.

FWIW, I have signed up on the SC waiting list for battery storage, and, as others have noted, this is not ready for prime time. I will let SC do the engineering and work on this, as this will entail not only mods to my extensive PV installation, but a new subpanel (I have 400A service and 4 subpanels currently) and break out of loads to allow the battery backup to carry an appropriate 'emergency' load of critical items in our home.

More info that you wanted...and don't complain about your electric rates in FL...

 

[ecarfan]

TSLA Pilot thanks for posting the link to Ideal Power. Reading through their site it sounds like they are targeting businesses rather than homes but I sent them an email anyway asking if they had a PV/battery storage system in the 7kW size range.

UPDATE 03-14-14: I contacted Ideal Power and they told me, quote "Our systems are sized for commerical and industrial applications: 30kW. May do something in the 10kW range in a year or so."

So they are targeting business customers and not small residential installations. Disappointing. I have no need for more than a 7kW PV system.

 

[lolachampcar]

There are some interesting things out there....

From what I have gathered, DC serial connection of panels has drawbacks as any shaded or otherwise non-performing panel brings down the set. If you are using four series sets (which I would be doing at roughly 400 VDC and 8KW DC power) one bad or shaded apple brings down 25% of your capacity.

Apparently there are two ways of dealing with this. The first is the mini inverter per panel (got a panel and an inverter on my roof now playing with it) where each panel/inverter syncs with your 220 VAC and pushes whatever AC it can generate. Each panel/inverter stands on its own as a generating source.

The second way appears to be some sort of DC-DC inverter to boost/buck (sorry, not looked into it deep enough) the DC to a level that is then fed to a standard PV system inverter which syncs to the home 220 VAC. Any one panel not performing does not impact other panels again providing isolation.

It is looking like the most intelligent approach would be a PV system sized for the house usage during the day and then a separate battery system sized to power the house for one portion of a normal 24 hour cycle when the sun is not producing. The battery system would have its own dedicated panel set(s) to charge the battery when the sun is shining. Volker, I think this is where you were trying to point me.

I looked into the Sunny Island inverter but it appears to be 110V only. I really need a system that does 220 VAC split rail so it will integrate seamlessly with my house.

One thing is for certain. I have no desire to net meter back to FPL as I think they only pay 2 or so cents.......

One of the most interesting thoughts for me right now is the idea of current monitoring at the FPL tap then dynamically regulating the PV system to divert PV DC to a battery such that I never push power to FPL nor take any when the sun is shining. Not sure that is possible at a reasonable cost nor really necessary but the nerd in me still wants to chew on the problem.

 

[InTheShadows]

There are some interesting things out there....

From what I have gathered, DC serial connection of panels has drawbacks as any shaded or otherwise non-performing panel brings down the set. If you are using four series sets (which I would be doing at roughly 400 VDC and 8KW DC power) one bad or shaded apple brings down 25% of your capacity.

Apparently there are two ways of dealing with this. The first is the mini inverter per panel (got a panel and an inverter on my roof now playing with it) where each panel/inverter syncs with your 220 VAC and pushes whatever AC it can generate. Each panel/inverter stands on its own as a generating source.

The second way appears to be some sort of DC-DC inverter to boost/buck (sorry, not looked into it deep enough) the DC to a level that is then fed to a standard PV system inverter which syncs to the home 220 VAC. Any one panel not performing does not impact other panels again providing isolation.

It is looking like the most intelligent approach would be a PV system sized for the house usage during the day and then a separate battery system sized to power the house for one portion of a normal 24 hour cycle when the sun is not producing. The battery system would have its own dedicated panel set(s) to charge the battery when the sun is shining. Volker, I think this is where you were trying to point me.

I looked into the Sunny Island inverter but it appears to be 110V only. I really need a system that does 220 VAC split rail so it will integrate seamlessly with my house.

One thing is for certain. I have no desire to net meter back to FPL as I think they only pay 2 or so cents.......

One of the most interesting thoughts for me right now is the idea of current monitoring at the FPL tap then dynamically regulating the PV system to divert PV DC to a battery such that I never push power to FPL nor take any when the sun is shining. Not sure that is possible at a reasonable cost nor really necessary but the nerd in me still wants to chew on the problem.

Modern panels have bypass diodes so the series degradation is not as bad as it used to be.

2 sunny islands can be hooked together to make 240v.

 

[lolachampcar]

Can the two Sunny Islands use the same PV array or do they have to have dedicated panels?

Thanks for the bypass diode info.

This is a nice place to hang and discuss these types of issues. Thanks to all for your input.

 

[InTheShadows]

Can the two Sunny Islands use the same PV array or do they have to have dedicated panels?

Thanks for the bypass diode info.

This is a nice place to hang and discuss these types of issues. Thanks to all for your input.

The sunny islands only work with the batteries.

The wiring goes from the grid - mains panel - sunny islands - backed up panel - sunnyboy solar inverters - PV array

The sunny islands have the batteries hooked up to them, the backed up panel has the emergency load circuits hooked up to it. If the grid goes down the sunny islands create an island and disconnect from the mains panel and keep everything on the backed up panel humming. Including the sunnyboy inverters which backfeed solar to the backed up panel and if there is extra the sunny island recharges the batteries creating a completely off grid system when the power is out.

This setup is much more efficient because the old school off grid systems were much less efficient because the panels went through a charge controller and maintained the batteries first then the DC got inverted. The new systems are grid tie first, battery backup second.

 

[lolachampcar]

Keeping up with posts here and doing some research on hardware is helping me better understand my goals and shedding some light on where my goals may/are not consistent with available systems. I think my targets are the following three in order of importance-

First
A PV system with tracking inverter that will supply most if not all of my needs when the sun is shining. I'll need to look at my TED data to size that system. Trying to make it any larger than my needs results in net metering back to FPL at something like two cents per KWHr so I most certainly want to err on the just a little less than needed side for this portion of the task.

Second
A battery system that charges when there is sun and provides a tracking inverter function to power most if not all my house's requirements when the sun is not shining. The Sunny Island stuff seems to be a back up power system whereby the grid mains are disconnected. My needs are more of reducing my grid usage while still having it in place. The battery contribution needs to be as transparent as the PV contribution above.

Lastly
It would be nice to have the battery capable of several days of backup. The only reason I considered this was, if I were going to solve the battery during off peak problem, it would seem a MS battery would be a cost effective choice if procured from a properly wrecked MS. The normal non-sunlight 24 hour cycle would only partially cycle the battery allowing it to be managed in an ideal charge range thus extending the life of an already long life battery. Any capacity reduction over time would only eat into the loss of grid and sun backup time and not the daily usage function.

Considerations for the above are becoming clear.
The ideal would be the ability to portion DC to the tracking inverter such that my grid meter runs forward slowly if at all. That is a dynamic control function.
The PV should be apportioned between the inverter and the battery as needed to achieve the above result.
The battery should be supplied to the tracking inverter in place of PV when the sun is not shining. There is no reason to purchase a second tracking inverter if the PV one will do the job.

The problem with the above is that I do not believe tracking inverters have a control input such that they regulate the amount of AC current they generate. They simply take what the have in DC input and push it out on the AC side up to the rating of the inverter. Some level of control would seem to be possible on quad DC input inverters by simply switching off one or more of the DC inputs. This would provide 2 KWHr resolution on an 8 KWHr inverter assuming this input switching works.

Some form of control will absolutely be needed when powering the inverter DC input with a battery as the battery is a low impedance source more than willing to provide whatever the inverter will draw (within limits of course but we are only talking an 8 KWHr inverter which is peanuts in MS battery land ). Again, perhaps input switching on the four DC inputs to the inverter will achieve this goal.

I'm collecting inverter information now. It would seem tracking is important so as to eliminate the need to isolate the house from the grid. However, the tracking function requires grid 220 VAC to sync with thus eliminating the possibility of using the system for a backup when the grid is down. If not available as back up (I do have a natural gas generator for this function and the grid is almost NEVER down) there is no need for a big battery as only off sun time in a 24hr cycle is required.

This is interesting stuff. I am surprised that there has not been more work on a simple 24 hr solution for people like me where net metering is simply not financially viable.

 

[InTheShadows]

The reason most people do not backup their whole house is costs.

The sunny islands can handle 40 amps 120 volts each. To have the same rating as a typical 200 amp house panel you would need 10 inverters, 5 for each leg. Then putting enough battery into the system for that size system would cost $100k+ for batteries alone.

This is overkill for most people, however if you want your mains panel to be backed up (USA electric) code requires you provide enough power for the entire panel.

Most homes in our area can be 100% solar powered with less than a 30kw system. A250w backup system is overkill.

 

[TSLA Pilot]

Keeping up with posts here and doing some research on hardware is helping me better understand my goals and shedding some light on where my goals may/are not consistent with available systems. I think my targets are the following three in order of importance-

First
A PV system with tracking inverter that will supply most if not all of my needs when the sun is shining. I'll need to look at my TED data to size that system. Trying to make it any larger than my needs results in net metering back to FPL at something like two cents per KWHr so I most certainly want to err on the just a little less than needed side for this portion of the task.

Second
A battery system that charges when there is sun and provides a tracking inverter function to power most if not all my house's requirements when the sun is not shining. The Sunny Island stuff seems to be a back up power system whereby the grid mains are disconnected. My needs are more of reducing my grid usage while still having it in place. The battery contribution needs to be as transparent as the PV contribution above.

Lastly
It would be nice to have the battery capable of several days of backup. The only reason I considered this was, if I were going to solve the battery during off peak problem, it would seem a MS battery would be a cost effective choice if procured from a properly wrecked MS. The normal non-sunlight 24 hour cycle would only partially cycle the battery allowing it to be managed in an ideal charge range thus extending the life of an already long life battery. Any capacity reduction over time would only eat into the loss of grid and sun backup time and not the daily usage function.

Considerations for the above are becoming clear.
The ideal would be the ability to portion DC to the tracking inverter such that my grid meter runs forward slowly if at all. That is a dynamic control function.
The PV should be apportioned between the inverter and the battery as needed to achieve the above result.
The battery should be supplied to the tracking inverter in place of PV when the sun is not shining. There is no reason to purchase a second tracking inverter if the PV one will do the job.

The problem with the above is that I do not believe tracking inverters have a control input such that they regulate the amount of AC current they generate. They simply take what the have in DC input and push it out on the AC side up to the rating of the inverter. Some level of control would seem to be possible on quad DC input inverters by simply switching off one or more of the DC inputs. This would provide 2 KWHr resolution on an 8 KWHr inverter assuming this input switching works.

Some form of control will absolutely be needed when powering the inverter DC input with a battery as the battery is a low impedance source more than willing to provide whatever the inverter will draw (within limits of course but we are only talking an 8 KWHr inverter which is peanuts in MS battery land ). Again, perhaps input switching on the four DC inputs to the inverter will achieve this goal.

I'm collecting inverter information now. It would seem tracking is important so as to eliminate the need to isolate the house from the grid. However, the tracking function requires grid 220 VAC to sync with thus eliminating the possibility of using the system for a backup when the grid is down. If not available as back up (I do have a natural gas generator for this function and the grid is almost NEVER down) there is no need for a big battery as only off sun time in a 24hr cycle is required.

This is interesting stuff. I am surprised that there has not been more work on a simple 24 hr solution for people like me where net metering is simply not financially viable.

Don't forget there may be a problem with using an MS pack--it's part of a BMS that requires a cooling circuit, and that circuit has at least two electric pumps, three heat exchangers, two cooling fans, and LOTS of software to keep the pack "happy" during its long life. Don't know if the loading/charge cycles for home use would generate similar waste heat, but this could be a major consideration as the last thing you'll want is an uncooled/overheated 85kWh Li-ion battery in your garage . . . .

I think Elon is on this as a long-term solution for both Tesla and SolarCity. If one runs the numbers (500k packs/6,500 employees), it's but 77 packs per employee per year. That's seemingly not a lot of output.

My guess is that some of those cells/packs aren't going into cars, but also other applications such as home energy storage, thus the 500k pack output may be strategic deception/partial truth on Elon's part, but only to keep competitors at bay (and that's fine by me!).

In a nutshell, I don't see how you can pull this off just yet, but we'd love to hear about it if you do.

 

[lolachampcar]

The key piece of information I was missing was a full understanding of FPL's net metering. I did not fully understand that FPL lets you "bank" power produced during the day and retrieve it during off PV hours. FPL is my battery WRT a 24 hour cycle which was my primary concern. Battery back up for outages is a very small concern (maybe out three times in ten years for hurricanes for which I have a backup generator). I was trying to solve a no problem problem........ Now I know.

Any yes, there would have been some work on the battery system but the nerd in me REALLY wants any reasonable excuse to start reversing the battery/inverter CAN interface.

 

[Pollux]

First
A PV system with tracking inverter that will supply most if not all of my needs when the sun is shining. I'll need to look at my TED data to size that system. Trying to make it any larger than my needs results in net metering back to FPL at something like two cents per KWHr so I most certainly want to err on the just a little less than needed side for this portion of the task.

I understand your point but wonder whether you are planning for your current-day needs versus allowing the possibility of a future increase in your needs.

Alan

 

[InTheShadows]

I understand your point but wonder whether you are planning for your current-day needs versus allowing the possibility of a future increase in your needs.

Alan

Or decrease in needs. (Excluding adding an EV of course.)

 

[lolachampcar]

I've been schooled a bit here.
FPL allows you to effectively bank what you do not use during the day and retrieve it at night with their implantation of net metering. They only account for any plus on the customer side once a year.
The system I am contemplating is 10 KW which should yield 8 KW on the AC side or roughly 40 KWHr per day. Given that I use more than that on average every day, I'll never see a plus thus will never provide FPL power at 1/4 what they charge me. It is a non-issue.

 

[InTheShadows]

Our state pays you on the annual basis for your net excess generation at the average generation cost. On top
Of that you get renewable energy credits.

The financial payback is maximized at 100%, however if you have a green side then going over 100% is better because your excess goes to your neighbors house thus lessening the losses on the power lines in sending them power from the central generation station miles away. Sometimes the line losses can be upwards of 20%, which means the power plant has to produce more power than you use to get you the power you need. If you can supply the power for your neighbors house as well as yours then that has a multiplying effect for the environmental benefits. So your environmental return is greater.

 

[Sunfishsolar.ca]

Have you heard of PVWatts, for calculating solar needs?

Our state pays you on the annual basis for your net excess generation at the average generation cost. On top
Of that you get renewable energy credits.

The financial payback is maximized at 100%, however if you have a green side then going over 100% is better because your excess goes to your neighbors house thus lessening the losses on the power lines in sending them power from the central generation station miles away. Sometimes the line losses can be upwards of 20%, which means the power plant has to produce more power than you use to get you the power you need. If you can supply the power for your neighbors house as well as yours then that has a multiplying effect for the environmental benefits. So your environmental return is greater.

PVWatts will guide you through to calculate your annual needs. I install these up in Canada. If you take your house demand, then do a calculation on your average tesla kW usage per year. You will be able to hit your mark within a few percentage points. Keep in mind if you are limited in Florida on inverter size to tier 1 (10kw) it works out to 11,700 watts of DC power and 10 kws of AC power output. Most likely around 12-13,000 kW/ hrs per year. PVWatts will tell you. I find PVWatts is conservative to about 4%. You should produce more than they say. If you have a choice pick the east roof. Morning can be cooler and may produce more. Every roof is different but consider temp vs time of day, and wind over panels to max output.

ignore the signature, I've got 32,000 Kms on it.

 

[InTheShadows]

PVWatts will guide you through to calculate your annual needs. I install these up in Canada. If you take your house demand, then do a calculation on your average tesla kW usage per year. You will be able to hit your mark within a few percentage points. Keep in mind if you are limited in Florida on inverter size to tier 1 (10kw) it works out to 11,700 watts of DC power and 10 kws of AC power output. Most likely around 12-13,000 kW/ hrs per year. PVWatts will tell you. I find PVWatts is conservative to about 4%. You should produce more than they say. If you have a choice pick the east roof. Morning can be cooler and may produce more. Every roof is different but consider temp vs time of day, and wind over panels to max output.

ignore the signature, I've got 32,000 Kms on it.

If using pvwatts v1 I find that setting the derate factor to 82 gives a good conservative value in our area. 84 is what I expect from the systems we install. 82 is what I give our customers though. Your mileage may vary but this is what I have found after installing close to 100 systems.

We have only had one system underperform the 82 and I think it was because of the inverter we used. Our installation team hated that inverter's install process so we never used that brand again so I don't have a way to verify it.

 

[mwulff]

I would recommend that you read this thread.

Plan: Off grid solar with a Model S battery pack at the heart

He already solved a lot of problems and posted a ton of details.