> To get any sort of stable battery-free power, you have to run the panels well, well below peak power (30-50% of peak is as high as you can really run), and even then, you have a horrifically unstable system. If the array power briefly drops below demand (perhaps an airplane has flown over), you shut down the entire output for a while
It doesn't seem like a battery or capacitor to handle blips like an airplane flying overhead would need to be very large?
No, but using a small battery means using a battery that get heavily charged/discharged by such events, and therefore has a short lifespan (less than a year, probably, if you used the kind of battery that's in a UPS.) The reason the battery banks for solar are so large isn't just to capture the entire daytime output for night-time use, but also so that they can have a lifespan closer to that of the solar array itself.
> heavily charged/discharged by such events, and therefore has a short lifespan (less than a year, probably, if you used the kind of battery that's in a UPS.)
But surely such events (i.e. grid power outages) are pretty rare. In Australia it's rare for the power to be out more than once a year.
That's interesting! Is it like an SSD where you plan for a certain amount of cells to fail or does it reduce load on the cells so they last long individually?
Notice how running the current the other way will break up the formed crystals, but it won't actually re-dissolve them. That's what it means for a battery to wear out. (Battery electrolytes are chosen specifically to be resistant to this, so the effect is mainly visible as a thin "rust" of deposited crystals on the battery's anode, rather than the full-scale crystals seen here.)
A similar but distinct process occurs when overcharging a battery, where, instead of splitting the electrolyte molecules apart, you're reacting and bonding them together to form new molecules or molecular complexes, with the reaction usually being one-way rather than an equilibrium reaction. (For lithium batteries this process is exothermic and catalyzed by the presence of the product—thus lithium battery explosions.)
It's much better for the life of the battery if it basically just hovers around 40-60% charge for its whole useful life, since then you're just generating tiny seed crystals (on discharge) and then re-dissolving them (on charge), where those crystals are small enough that they can be fully re-absorbed.
And this is true even in battery-cell technologies that require a "deep charge cycle" to erase their "battery memory." Battery memory is basically the electrolyte causing enough crystalline rust specifically on the anode to increase its resistance. A deep discharge can capture and erase this rust—but it still shortens the battery's lifespan, because you're still producing non-reabsorbable large crystals within the electrolyte.
Yup. Though I've heard (as it concerns deep-cycle batteries) not to let them drop below 50% charge rather than keep them in a certain range. So, for example, when I purchased a new house battery for our RV, I got a 200ah battery knowing I can only make use of 100ah before recharging. Or put another way, figure out how much you'll use between charges (or in the case of solar, how much to tide you over until the sun comes up), and double it.
Not quite. Planning for battery failures is hard because they have different failure modes. For example, the battery could fail to dead short, which is the functional equivalent of an cell in the SSD dying and then setting the SSD on fire. Or the cell might just actually catch fire during normal operation if you overload it.
To my knowledge, the best and safest battery system is to have each battery individually monitored by a charge controller with the capability to fully disconnect the battery at will.
The downside is that this will be expensive so people usually settle for just eating the rare chance of a dead short battery.
Additionally, batteries that you just leave around doing nothing will probably die at some point too. You don't want it sitting around you want to use it for efficiency or else you swap your battery and find out it was dead too.
To reduce burst loads on batteries you could use supercaps if you find ones that can handle the voltage and current (that will be very expensive).
> Additionally, batteries that you just leave around doing nothing will probably die at some point too.
This wouldn't happen if batteries were designed to act like nuclear reactors, where the "fuel" (the fuel rods; the electrolyte) can be completely removed from the substrate that makes it react (the neutron medium; the anode+cathode.)
But a battery that can withdraw its anode+cathode from the solution would be damned expensive. It'd make more sense as an architecture if you had just a few, super-large battery cells, e.g. giant vats of lead-acid.
It might be possible to design regular battery cells such that they wouldn't start degrading until they were first exposed to a voltage load, though. (I think the "50 year" Duracell NiCd batteries have this property—they probably have an antifuse oxide layer between the anode/cathode and the electrolyte, that gets broken down when you put load on the circuit.)
That reminds me of a recent NASA project where they kinda built small nuclear fission batteries with less than a couple tons of weight (IIRC down to 100kg)
These would be kinda neat as off-grid generators and you can take the nuclear element out (and it would last longer but still be radioactive).
That you could do with a super capacitor, but not with a battery.
I know of some railway train setup, which is deployed in Saudi Arabia by Siemens, that uses a combination of battery and super caps to power the train. While the train starts it takes the sparking current from the super caps. While riding it will take the energy slowly from the batteries. While deceleration it will recuperate and will charge up the super caps, because they can be charged easily with a high current. So the super caps are used as a high available energy buffer with very fast charge and discharge capabilities. While the battery is only used during the times of less energy consumption. Also at the stations the super caps are super charged very quickly. With that concept, they can run the train the whole day with only 40% discharge of the battery. The battery is actually way smaller then in a Tesla.
This concept can be applied also locally. The problem: these super capacitors are way expansive so you need to design them to your specific requirements.
You can do it with a battery. An 18650 that can supply a hundred watts is $5 or less. Even limited to half its discharge depth for longevity, it can take you through "blips" two minutes long.
I wouldn't want to ride a battery pack like that all day and every day, but it should be fine for emergencies.
Also, while not necessary trivial, it is not hard to throttle consumption of devices, and if it's with two circuits, one for devices that care about voltage, and one for devices that actually brown out. The latter is easy to scale by just reducing the voltage far enough that it doesn't sink too much power, and devices like e.g. desktop or server computers are theoretically easy to rapidly throttle, e.g. by reducing cpu clock to a few hundred megahertz, halting HDD operations or even cutting their power (most are made to handle hot-swap well).
A large, modern CPU with 140 W TDP can cut usage within 10 ms to 10W, but dram can't scale that way due to how it works, or rather, this scaling is not supported in mainstream hardware.
If you use a small supercapacitor, which commonly has a minimum discharge time of about 2~20 seconds (similar to Lithium-ion pouch cells "LiPo", which are available with different "C" ratings, and negatively correlated gravimetric/volumetric energy density and maximum average discharge power (i.e., optimizing discharge current vs time to maximize the ratio of total power extracted divided by time until empty). You need more metal in the electrodes and the plate/foil that aggregates the current of the individual electrodes, leaving less space/weight for the parts that actually store energy.),
you can use much more advanced power reduction techniques for electronic devices, as you have time for something approaching a proper shutdown. Compare e.g. the time your Laptop takes until it is off from the moment you close the lid.
Oh, and yes, these capacitors are cheap. One about 1kg / 1 liter size for about 100$ can provide 20kW average power, for 2 seconds. The main difference is that they don't care (much) if you cycle them at 100% depth of charge, apart from internal losses heating it up and thereby causing damage.
It doesn't seem like a battery or capacitor to handle blips like an airplane flying overhead would need to be very large?