Not everybody has $6500 to toss into a Tesla Powerwall (and that’s a low estimate), but if you want the benefits of battery storage for your house, [Matt]’s modular “microbattery” storage system might be right up your alley. With a build-as-you-go model, virtually any battery can be placed on the grid in order to start storing power from a small solar installation or other power source.
The system works how any other battery installation would work. When demand is high, a series of microinverters turn on and deliver power to the grid. When demand is low, the batteries get charged. The major difference between this setup and a consumer-grade system is that this system is highly modular and each module is networked together to improve the efficiency of the overall system. Its all tied together with a Raspberry Pi that manages the entire setup.
While all of the software is available to set this up, it should go without saying that working with mains power is dangerous, besides the fact that you’ll need inverters capable of matching phase angle with the grid, a meter that handles reverse power flow, a power company that is willing to take the power, and a number of building code statutes to appease. If you don’t have all that together, you might want to go off-grid instead.
35 thoughts on “Microbatteries On The Grid”
You know, rather than dealing with the inefficiencies of charging and discharging, and the $0.50/kWh cost of the batteries, and the fire risk and capital and maintenance cost of a pile of power stuff, I’d be really happy to just be able to tell my fridge and freezer when to do a defrost cycle. It irks me when they kick in in the middle of the day at high time-of-use rates and burn a kWh. It would make a lot more sense to hold off and do it at 2 am. Or do it when I have a surplus of solar power to burn, for that matter.
That’s called feedback, and you’d have to consider what happens when enough people are doing it to matter.
Though if you just set it on a timer, then the cycle time will be long enough that it won’t create destructive oscillations. By the time you notice the prices change, the information is already stale.
Oh this is the kind of thing I would like to have source code from to adjust settings :( Like in my car… ah :(
It would be a lot better if appliances were online and the power company charged you less for the right to switch them on and off to meet demand load. This is much cheaper than massive on-demena power stations such as pumped storage and could, if all devices were so controlled, to around 20% less power stations.
Time of use charges and/or smart hours already exist. The utilities used to insert a signal into the 50/60Hz carrier to signal power meters to switch heating loads on and off, and in most places they still do.
The issue is that most demand is not flexible like that. Heating and cooling might be, but if it comes to playing around with the electricity rates, most people simply switch over to gas.
Load adaptation would be much better received if it’s centered around turning things *on* rather than turning things off. That said, thermal storage for HVAC is orders of magnitude cheaper than batteries and would be something great to invest in.
Gas is orders of magnitude cheaper than thermal storage.
How do you air condition with gas? Absorption systems have been tried but the efficiency is so poor that you might as well run a conventional air conditioner from a generator.
In cold climates with cheap gas, residential CHP would be a good investment.
Just another reason I honestly don’t really like it when products aren’t connected. I can think of very few devices I own that wouldn’t benefit from an RF module(Assuming it was an open-ish API).
Everything should know about the grid, and everything that gets hot should text you if it’s on too long.
Yikes. The day my appliances try to connect to the web is the day that I replace their CPUs with relay logic or cheap PLCs.
Why not just put them on an inexpensive relay tied to turn off at a set hour? You don’t need any expensive ADR functionality to turn off an appliance when peak hours are so predictable.
Unplug your fridge for 12 hours, it has a simple clock timer
Hot water tank timers can save significant money if you are on time of use rates. I use a Nodemcu controller, a clock module and a high power relay. Web page hosted on the controller is used to monitor and override auto settings if needed. Other controllers are used to save money on utilities as well.
I have the same thoughts all the time, now that I monitor my fridge’s power cycles. The consumer has very little control over it.
Or that they not do it all and stop trashing the food in my freezer…
– At least in my neck of the woods, there’s a monthly fee to go to net metering. It can pretty quickly put a small production setup into the red rather than having even a minor monthly savings. I’ve been curious / contemplated that I believe if I kept track of demand (monitor a fewer high-demand appliance status – furnace, AC, air exchanger, etc – along with which side of the split-phase they are on), and kept production below that point, I *in theory* should be able to reduce my demand without going to net metering. Depending on the metering, you’re likely to be charged for any over-production rather than credited, if the meter doesn’t phone home you’re producing while it’s at it(?) – but if you leave a of margin of error to your production, use inverters made for grid-tie, seems feasible? – Thoughts?
I don’t think any net metering scheme punishes for over-production. It’s rather that they’re designed to force the utilities to accept whatever random power in order to benefit the companies that import Chinese solar panels.
– Agreed – I’m talking staying off of net metering though, to avoid the monthly fee of going to net metering. From what I understand, older analog meters will will spin the same whether you are producing or consuming – being just an absolute current meter (which makes sense, since it’s AC, and tracking direction of flow would be more involved). Supposedly newer non-net ‘smart’ meters work the same, but I’ve also heard comments of even though they are not intended for net metering (power vendor replaces your smart meter if you go to net-metering plan), they have enough smarts to sense production rather than consumption, and report it back to the vendor (probably not ideal if you’re looking to produce to lower your bill while avoiding going to a net-metering plan).
If you want to be a responsible gird user, you don’t mix your renewable power with your grid power anyways. The intermittent power you don’t use shows up as demand variations just the same, which creates troubles for the utility.
– Isn’t that pretty much saying any non-utility production is bad (more clarifying than being argumentative)? – any consumer-level net-metered solar or wind install is going be forcing power back into the grid when it happens to be available, since the utility doesn’t (generally) have any control over your production output. While in my scenario, I’m more looking to lower my demand, rather than forcing the utility to deal with my excess production when they may not need/want it, let alone essentially charging them retail rate rather than wholesale while I’m at it. From the utility side it more looks like I wasn’t home for the week and my demand was low, rather than a net-metering producer of ‘here, take all my wind power @ 2am, and I expect (net) retail rate credit for the production while I’m at it.
– Seems ‘more responsible’ to me to lower your personal demand, especially if you can do it on a somewhat consistent basis, at least as opposed to economics of net-metering. I think (?) I get where you’re coming from, but short of leaving all production purely to the utility (or where they wholesale excess demand from) to be in full control of all production, or going off-grid completely with loads, it doesn’t seem to leave any happy-ish medium.
> Isn’t that pretty much saying any non-utility production is bad
Yes and no.
Suppose you have a water turbine in a stream. That’s going to be pretty much 24/7 so the utility can predict what you will put out at any time. That’s no problem – they can make you a deal. Then suppose a million people have solar panels which reduce their power consumption randomly by the amount of clouds passing by – that’s going to be an issue because on aggregate it’s a huge amount, and it goes up and down at a tremendous rate, and at best you’ll get a couple hours heads up.
>it doesn’t seem to leave any happy-ish medium.
There is no happy medium here. It’s because renewable power is so intermittent, and when the effect is spread over millions of customers the grid goes unstable. It makes no difference whether there’s net metering or just demand reduction – it’s the same swing in output that the utility has to provide in order to meet the grid load.
Though, assuming that nobody makes any surplus, demand reduction would technically be not as bad as net metering because the amplitude of the power variation would be less, but you get the point.
Net metering assumes that you’re making more than you’re consuming – but that’s the usual case especially with solar panels, because people aren’t at home at the peak production hours. Then they come home at night when the sun isn’t shining, start cooking and having hot showers, turning on their AC and watching TV, etc…
Is it a real problem, or just an engineering challenge they should solve, as part of the service they’re providing?
If it makes the whole thing unreliable or inefficient, that’s a problem, but if it’s just a matter of adding some tech, seems like that shouldn’t get in the way of having solar on every roof.
Failing that, since solar is often cheap compared to batteries, can inverters just reduce output to match demand, and we all just accept having lots of unused capacity, until loads get smart enough to take advantage of the sunny hours?
That way, grid tie inverters would *be* the load following power plants when possible. There’s already increased demand in the daytime, for air conditioners, so I’d imagine there’s plenty of demand to keep panels busy.
>If it makes the whole thing unreliable or inefficient
That is exactly the problem. The power variations have to be diluted into the wider grid, which means the average transmission distance increases and efficiency goes down, and the infrastructure has to be over-built relative to the actual demand.
>can inverters just reduce output to match demand
Yes, but in order to make your investment back in 20 years assumes you’re selling all the power. If you don’t have net metering, the solar panels will never pay for themselves and people obviously won’t be buying them.
The main problem is that it’s the same sun for all people. Everyone from San Fransisco to Seattle produces at the same time, so the peak goes up and the price goes to zero because there’s no buyers, only sellers. Without investment and rate subsidies and forcing utilities to take the power, solar panel just couldn’t exist.
When I first had my solar system in stalled I had an old analog meter with the ol’ spinning disk and it would spin rather rapidly in reverse in the middle of the day and my consumption meters would go back wards as well.
You can take a look at “zero export” inverters. Sadly, there’s not that many options especially if you only want a little battery backup but want to be able to offset any load in the house. Therefore, I’m building my own out of a Prius inverter.
– Interesting, I hadn’t come across zero-export inverters being a thing. That’s essentially what the idea was, more of a DIY zero export setup from load monitoring and a handful of grid-tie microinverters. Microinverters you can pick up on the (relatively) cheap; not so much zero-export inverters, metering swaps, transfer switches, etc. Good luck with the Prius inverter, seems like that could give you a pretty good hardware starting point if you can pull off the control of it! – from what I remember of the battery/MG1/MG2 setup, seems that could give you some pretty nifty consumption/production/charging capabilities!
Damien Maguire has figured out how to take control of the PWM signals. The biggest appeal is the low cost – $100 or so for an inverter able to power the whole house.
Where I’m about to move there’s a minimum fee monthly for being connected to the grid ($24) and then an addition fee if you net meter (+$10).
Personally, I’m leaning towards not connecting to the grid. In the end I think it’ll be the same price or lower for me to just be off grid. I mainly wanted grid power for when I get into welding… But will probably just fire up a generator for that instead.
Instead of a requirement to receive a special signal from the power company, it might be almost as good for small amounts of generation/load to simply key off of frequency. Generate under 60 Hz, charge over 60 Hz as generally utilities aren’t penalized and possibly rewarded for providing frequency support.
You have to tune the system or you’ll create a feedback oscillation in the grid – so you have to turn the control over to the utility anyways.
And when the electric companies get wise that that’s what everyone is doing, they’ll put the night rates up.
A Tesla powerwall is only $6500?
I’ll take two yesterday. Seriously.
Here in Japan, the equivalent devices start at about $20,000…… and they only really act like crappy UPS systems with 5 second switchover times, or you only get a fee hundred watts output, or, or… the selection is abysmal here.
Please be kind and respectful to help make the comments section excellent. (Comment Policy)