-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Solar Panels UK - is it worth it?
- Thread starter jdwuk
- Start date
I have had the same issue and lived with it for 8 years! It is inevitable that multiple ’generators' on the same side of the current transformer can't identify the difference between solar power and battery power once it's converted to ac.
I'm just about to start trying to come up with a work around. For me, a solution may exist by configuring a volt free contact on one of the devices. But in truth, I'm not confident in finding a complete solution as the devices are not capable of 'talking' to eachother.
WannabeOwner
Well-Known Member
Not sure I'm talking about the same thing ... so put me straight if not.
I've become of the opinion that the Charger having some control / "intelligence" is not what I want.
I want the Car to decide that it will charge (well .. something telling the car to charge)
For example, if I use Zappi to dump to car when it detects Export then a couple of things happen. #1 my battery charges at a Max Rate which is probably not good for it and #2 when car starts to charge (after battery full) I have more than 7kW PV and so I will export anyway ... or a second car charger needs to charge the other car (and the two chargers are going to have a tough time working things out).
Plus, from my perspective, I am likely to want Car-A to be charged in preference to Car-B because tomorrow I'm off somewhere. And I won't want to have to be arsed to make sure that I have plugged the right car into the right charger (if you don't have two cars / chargers yet then maybe consider that that might be a future problem
)So ... I'm working towards something that predicts tomorrows Solar Radiation, gets some input from me as to which car to charge and by how much, and sets a plan a accordingly (and will charge the car and/or battery during Off Peak if the Solar Radiation is predicted to be insufficient/lousy). If it looks like there will be enough for some excess than I will start charging car as soon as the battery charge gets to the point of "more than is ideal for battery longevity (**)", and that will mean battery takes longer to charge and then once it is full I'm like to be passed the mid-day PV peak and unlikely to then reach "Battery full but PV generating more than 7kW"
My experimentation so far (manual) has not done a good job of guarantying that Battery is at 100% just as the sun gets to the point where PV is less than house usage ... and (my setup) I need very close to 100% battery to get through the night (contemplating another battery to make that easier to achieve, and allow more power draw overnight ... and will definitely make use of that in Winter when I need battery from end of Off Peak to start the following night!)
But once I use some software, instead of my "When I remember to check"! approach, I reckon it will feather the various players (and it could wake up some dump-sources - immersion, aircon, etc. on days once battery and cars are full already)
I've been using solcast.com.au for prediction of solar radiation and it looks promising.
(**) I don't have the exact numbers to hand, but its something like 10kW MAX and 3kW overnight (which is deemed to be best for longevity)
That crimp tool out the whole damn network, just plugging it into the powered down network
.The Andersen has an Advanced Solar option in beta that I'm going to investigate. Otherwise I'll take back manual control and just choose to charge at a limited rate for a few hours. But that sucks.
The puredrive has a local web interface which I can hopefully scrape some data from then interact with the car or something as v1.5.
My setup is different – Zappi, Victron inverter, Pylontech, 3kW AC-connected PV – but maybe the concepts will carry across. No big changes were needed to get the Zappi to play nicely with the battery.
In the morning the inverter software does a good job of taking all the surplus PV for the battery, keeping the export to the grid in the range 0 to 50W approx. If the car is plugged in the Zappi will be in "Eco+" mode and won't do anything while it waits for some minimum level of export, around 1.5kW.
When the battery reaches 100% the Zappi notices the export and (after some seconds) begins charging. It starts with an optimistically high current, 1 or 2 kW more than the PV is producing, so the inverter prevents any import by supplying the difference from the battery.
Now the cunning part ... the Zappi also has a target for the grid connection: its "export margin" is configured to a relatively high value, say 150W, meaning its goal is to have 150W export at all times. So while the battery is draining, the Zappi spends about 60 seconds steadily reducing the charge current to achieve that goal. Then at first the battery takes part of the PV power to top itself up, and after a minute or two we have the battery sitting at 100% and the Zappi current following the PV up and down, all the time aiming for 150W export.
In summary: no special hardware, just the normal clamps on the grid connection so the inverter and Zappi can measure the import/export. No special configuration beyond the export margin in the Zappi (Advanced - Supply Grid - Network).
In the morning the inverter software does a good job of taking all the surplus PV for the battery, keeping the export to the grid in the range 0 to 50W approx. If the car is plugged in the Zappi will be in "Eco+" mode and won't do anything while it waits for some minimum level of export, around 1.5kW.
When the battery reaches 100% the Zappi notices the export and (after some seconds) begins charging. It starts with an optimistically high current, 1 or 2 kW more than the PV is producing, so the inverter prevents any import by supplying the difference from the battery.
Now the cunning part ... the Zappi also has a target for the grid connection: its "export margin" is configured to a relatively high value, say 150W, meaning its goal is to have 150W export at all times. So while the battery is draining, the Zappi spends about 60 seconds steadily reducing the charge current to achieve that goal. Then at first the battery takes part of the PV power to top itself up, and after a minute or two we have the battery sitting at 100% and the Zappi current following the PV up and down, all the time aiming for 150W export.
In summary: no special hardware, just the normal clamps on the grid connection so the inverter and Zappi can measure the import/export. No special configuration beyond the export margin in the Zappi (Advanced - Supply Grid - Network).
Sounds spot on! Will explore the Andersen software some more to see if I can find something similar!
WannabeOwner
Well-Known Member
My (subjective
) view:I upgraded my PV a few months ago (48 panels now), and added (to Tesla Charger) a Zappi at that time (assuming I'd be using its "Divert to EV when export detected")
But ... I'm not sure I want to charge my EV at 1.5kW ... as I think the losses are significant.
Thus I would prefer to divert part of my PV to EV, at a high rate, using the PowerWall as the Ebb-and-Flow, whilst attempting to get PowerWall to 100% at end-of-day.
The Zappi approach (on a good-sun day) is charge the PowerWall flat out, then charge the car flat out. The car is going to be happy with flat-out 'coz it will be perfectly happy when it max's out at 7Kw (although, in my case, I have enough PV that that will at times go well over 7kW so I'd still export at that point), but based on the rate the PowerWall charges on Off Peak (which I read as "ideal for longevity") I'm not sure that "flat out" is best for PowerWall - on a good-sun day mine will get from 5% to full by 11AM (10 kW Max I think, my peak PV is around 15kW)
My approach is, of course, a lot more complicated ... particularly as it involves forecasting whether this afternoon's sun will be good enough to actually fill the PowerWall the last few percent to 100% by sun-down ...
Given the complexity and custom method your looking to achieve, and probably various modding going forward, the only way I can see you getting this to work is to make your own custom program setup.My (subjective
I upgraded my PV a few months ago (48 panels now), and added (to Tesla Charger) a Zappi at that time (assuming I'd be using its "Divert to EV when export detected")
But ... I'm not sure I want to charge my EV at 1.5kW ... as I think the losses are significant.
Thus I would prefer to divert part of my PV to EV, at a high rate, using the PowerWall as the Ebb-and-Flow, whilst attempting to get PowerWall to 100% at end-of-day.
The Zappi approach (on a good-sun day) is charge the PowerWall flat out, then charge the car flat out. The car is going to be happy with flat-out 'coz it will be perfectly happy when it max's out at 7Kw (although, in my case, I have enough PV that that will at times go well over 7kW so I'd still export at that point), but based on the rate the PowerWall charges on Off Peak (which I read as "ideal for longevity") I'm not sure that "flat out" is best for PowerWall - on a good-sun day mine will get from 5% to full by 11AM (10 kW Max I think, my peak PV is around 15kW)
My approach is, of course, a lot more complicated ... particularly as it involves forecasting whether this afternoon's sun will be good enough to actually fill the PowerWall the last few percent to 100% by sun-down ...
I have a fairly complex setup of my own, with 2x 22kw chargers, 30kw PV Array and 56kw Combined Heat and Power unit (CHP) at my farm. Plus a 60kw paper shredder and other stuff onsite. My chargers are not near electric meter, so AC meter clamps are of no use.
To minimise export/limit import i need my own setup.
First, all my 'stuff' 'talks' either on MQTT protocol or Modbus TCP or Modbus RTU.
I use node red Node-RED to create flows which determine what should happen given a set of parameters.
My electric meter I poll every 2sec using Modbus RTU to get export/import readings
The chargers talk MQTT to nodered, so I can change the current up and down as needed and read kw, kwh current etc (every 5sec).
The CHP and PV inverter talk modbus RTU, so nodered can read data from them.
Once you have all the data coming into nodered, it's just a matter of creating the flows to get the outcome you want.
Nodered can run on a Raspberry Pi and is all web based, some simple Javascript coding maybe needed in places, but it's mostly graphical boxes with wires between.
As with all things like this, you start small and build up, such as get the data coming in and the control going out first.
Chris
WannabeOwner
Well-Known Member
Yes, I'm expecting that (although I was expecting to find some APP that I could pay real money for that would do the job ... seems that all the GitHub stuff has not been monetised into an APP as yet ...)
Very useful read of what you are doing, thanks.
I'm a database developer by trade, so able to tackle this, but the whole "Raspberry Pi and Node-RED" is a level of fiddling that I would prefer not to have to find the time for ... but as you say, likely there will be no other way (although I am hopeful I can get enough data from PowerWall API to drive my process)
I've had similar thoughts: I've started recording data, and I am manually fiddling with "Charge car at X AMPs" and then, hopefully!, remembering to check it soon enough that I can intervene if weather has clouded over and I'm draining the PowerWall
. Those manual actiosn are reflected in the actual (logged) charging usage by the cars, so I have data for "what I tried manually" too.I have data feed from solcast.com.au which is the best source I've found for insolation prediction - I'm figuring that I need to take a view at midnight about whether to charge the powerwall during Off Peak (rare this time of the year, but keen to do that if tomorrow there won't be enough sun), and then at 5AM-ish I will drop the Reserve from 10% to 5% as most nights I am just about running out of PowerWall around 5AM, (and put it back up as soon as PV has replenished it), then at 6AM-ish, just as the PV starts to generate enough to cover my background usage, start to take a view on how much excess I might have during the day - the prediction from Solcast seems to be fine-adjusted for the next 3 hours and, other than that, is in the main following what was predicted 24 hours ago.
Then I am recording the daily log from PowerWall which gives me 5-minute logging of Home/Solar/Powerwall/Grid (kW) and PowerWall SoC. So far I am reckoning that will be good enough for how much the home used (including EV charging) and what Solar generated, how much went into/out of the PowerWall and Export/Import from Grid.
I also have TeslaFi for the cars, and can export data from that, which will give me Charge AMP and time - which I can combine with the above to see what charging I did during the day (and subtract that from "Home" for nett home usage), and also to give a figure for any overnight charging I have needed to do during peak summer.
So, I think?, I am recording data for Inputs and Outputs to the historically / retrospectively review
I haven't found time to chuck that data into database, yet, but once I have I ill expect to then be able to retrospectively predict (from data) what I should have done, and compare that against actual, and then - my hope is - be able to refine my prediction-algorithm to then recommend what I need to do to the cars.
I will then need to build something to talk to the cars and powerwall APIs which will I expect will take a huge amount of time as I figure out what I am doing wrong with the various APIs!
WannabeOwner
Well-Known Member
Sorry, forgot the pretty picture
Top one is 5AM this morning, bottom is 10AM-ish
Left shows actuals (well ... its Solcast's data for my region for what it thinks actually happened, I can use my own data for that too and see what variance I get from Solcast)
Right is its forward prediction - as the bottom one shows fine-adjusted for "next 3 hours", but rest of forward prediction seems to stay the same
Getting the Pi and Node red running, probably a 15 min job, (not hard).
The getting of the data arriving in node red is where you have to think...
Also MQTT, modbus, database connectors etc are all available for node red, just a couple of clicks and their installed, all from the browser interface.
But node red can easily poll databases for data and then 'do stuff' based on that.
It can write to databases very easily also,
I output data into google spreadsheets via node red, like every time a charge is completed on the chargers, with the amount used etc.
Or use it to message me, the charge is complete with the kwh used.
For me anyway, I'd go for actual export/import as 'seen' at the meter rather than predicted solar data.
As a farmer we find the weather forecast is regularly wrong for the next 24 hours never mind longer periods.
There is even a node for node-red someone created for the tesla car api, although I've not used it, I use the MQTT protocol via Teslamate with node red.
Since MQTT is very easy and flexible to use.
I have dual immersun diverters which can dump up to 6kw excess into thermal store and using the export margin on different devices is how I try and control their behaviour.
Yes, I agree. If you really have nothing else to do with a small amount of export power, then charging the car even at the lowest possible level is maybe just worth doing.
Loads of posts since I started typing this!! They make the point that prediction is not straight forward, and it's easy to waste ages and achieve little.... especially if you try to factor in battery type, energy throughput, losses at various charging rates, tomorrow's weather, next Friday's weather, if the wood burner is on, whether to run the heat pump preemptively, how cold swmbo decides to feel, charging car vs heating house vs charging battery etc etc... Talk about rabbit hole!
I think aiming for big picture, low hanging fruit and trying to understand what your aiming for as well as what you are actually managing to do is all I can hope for.
BTW, have any of the assembled company come across any generic G100 devices that might convince the DNO they don't have to worry about what I might export! I don't really get why a DNO accepts that g83 and now g98 can dependably keep export to sub 3.6kw or whatever, but not accept that the same control systems can be relied on (suitably configured) to ensure a g99 inverter can't export more than say 2kw.
with regard to this, I have little to offer...
Except when I've dealt with the DNO to get export connections for my kit (90kw export in my case)....
They seem to have a rule book they follow, if you try to deviate from that, you get told 'no' , there's no negotiation.
If you don't fit the general profile... forget it.
'computer says no'
Could you potentially really export that amount of power? I don't want to export anything if can avoid it.
I'm leaning towards an off-grid with grid support configuration as a way of taking DNO out of the equation at least as far as any export. Down side is that my inverters have to be sized to handle max continuous load, but the upside is I get to pretty much do what makes most sense as far as the rest of the system design, and grid-fail backup comes as part of the system design.
The other day, when v sunny, and we were not shredding paper (weekend)...
We hit 80kw (might have been a little over, 82?) export.
If are onsite usage is very low and it's sunny the PV will hit 30kw , and the CHP does 54.5kw (after parasitic load)
So we get quite close and I always like a bit of head room, so as to not upset the DNO...
We balance against the grid, our kit cannot run without a powered up grid connection....
It's the nature of the kit we have...
WannabeOwner
Well-Known Member
. Scaffolding didn't come down, and it was really cloudy (can't workout where the sun is levels of cloudy). Hopefully not too many more days like this until autumn!
RedMod3
Member
Yesterday my second PW was installed, and today it was commissioned, and so I can now soak up all my excess solar and not give any back to the grid, although today I was extremely generous and gave them 2.5kWh (12.5p worth). A second PW means that between Spring and Autumn, I should not be drawing any power outside the off-peak GO period as solar will keep me topped up, and I'll use energy at 5p kWh if necessary.
I'm most impressed with the Tesla product, and I wait patiently for PW3 in October, which will keep me on off peak power during most of the winter and keep our total energy bill to under £1,000 per annum (or zero if I knock off my FIT payments).
So yes, solar is definitely worth it and adding a PW significantly enhances your system.
I'm most impressed with the Tesla product, and I wait patiently for PW3 in October, which will keep me on off peak power during most of the winter and keep our total energy bill to under £1,000 per annum (or zero if I knock off my FIT payments).
So yes, solar is definitely worth it and adding a PW significantly enhances your system.
Similar threads
