Time to brush up on current affairs. Because we're predicting Li-ion batt lifetimes using impedance and AI

Machine-learning software can predict the remaining useful lifetime of a lithium-ion battery by seeing how it reacts when a rapidly oscillating voltage is briefly applied across it, according to a study published in Nature Communications. The constant cycle of discharging and charging Li-ion batteries gradually knackers their …

[Anonymous Coward]

They hope the algorithm....

It's not an algorithm, its a black box, if they can tell me why specific voltages returning specific impedence values means a specific number of charge cycles, fine, but I bet they can't.

This type of thing will reduce our advancements in the future as it will reduce our understanding of how and why things happen. Taking the easy route will make things harder in the future.

[diodesign]

"It's not an algorithm, its a black box"

It's linked to from the article - you can download it and run it in Matlab.

"if they can tell me why specific voltages returning specific impedence values means a specific number of charge cycles, fine, but I bet they can't."

It's in the paper. You train a model to take these variables - frequency, temperature, impedance, etc - and match them to battery lifetime. So that when you show it arbitrary EIS values, it predicts the lifetime.

C.

[bombastic bob]

my guess is that it would be hard to use that kind of algorithm with a microcontroller, so NOW "someone out there" will need to produce a charge controller that can be queried [let's say using I2C or SPI] for the charge cycle calculation. That should also work for phones, laptops, and even cars [if it's designed for multi-battery systems].

At least, that's what I'd LIKE to see happen - intelligent LiPo/LiIon charge controllers that can also measure battery capacity.

I worked on something a while back that would estimate battery life using the known charge characteristics, the known discharge current, and the number of 'ticks' it would take to cross specific voltage thresholds during the charge and discharge. It duplicated (more or less) my understanding of the 'smart battery' tech that goes into phones and laptops. Still it was easiest to rely mostly on the charge controller IC to make sure the battery was properly charged and then just intelligently measure things to estimate capacity. [and I don't really want to see the universe encumbered with SOFTWARE PATENTS over this...]

So in my view, THAT is the best place to put this kind of AI batterty tech - in the charge controller, with a way of reading it (like I2C or SPI).

[Anonymous Coward]

It's not an algorithm, its a black box

You don't know what's happening inside a black box, by definition.

However, when you generate an algorithm using machine learning, as in this case, you can see exactly what it's doing i.e. you can see how it works, you just don't have a clue why it works.

[Electronics'R'Us]

Battery analysis

Having once spent a year (at least it felt that way) analysing the charge state of NiCd batteries, I can say that impulse response (a form of ac excitation which does indeed measure the impedance) is a pretty accurate measurement method for rechargeable batteries.

The problem is that battery construction (there are a lot of ways to construct a battery in a given technology depending on the end use) requires each type of construction to have its own training data which I had to do by hand (this was 30 years ago and I tested literally thousands of batteries from multiple manufacturers and batches) although the 'typical' construction for a given battery yielded type consistent results across manufacturers (the manufacturers said it was impossible to know the charge state but they would - they were in the business of selling new batteries).

There are a lot of gotchas in this field but with sufficient training data the analysis can be quite accurate.

I ended up building an analyser with a microcontroller and the results were tested by actually discharging the battery and confirming that the test had reported the correct charge state - it was always within 5% of the reported state of charge and typically within 2% so it is possible.

[Mike Shepherd]

Re: Battery analysis

You speak from a time when it was fashionable to think. Now we say "It's all too complicated, so give it to a neural network and we'll move on to something else". Oh, and invent some pretentious term to impress the natives. We measure the battery impedance at two or three different frequencies, so call it "electrochemical impedance spectroscopy".

[Electronics'R'Us]

Re: Battery analysis

The up front analysis took many forms before I found one that appeared to correlate (pulse testing); what is interesting is that the entire analysis was done with a precision discharge circuit (I already had that available), a charging setup, a multimeter, a cheap oscilloscope, a few resistors and a timer (and a lot paper and simple mathematical analysis).

Oh - you are right; and a lot of thought.

[Citizen99]

Re: Battery analysis

"pulse testing" - rich in harmonics, more than just 2 or 3 frequencies , perhaps relevant :-)

[Electronics'R'Us]

Re: Battery analysis

Indeed; the impulse response test had significant harmonics out to at least the 7th and I found that a couple of pulses of different durations (and therefore differing frequency content) increased the accuracy of the test.

NiCd batteries are far more difficult to analyse than Li+ which do have a reasonable relationship of terminal voltage to state of charge; The open circuit and loaded voltage of NiCd tells virtually nothing about the state of charge of the cell.

What I did find was that by loading the cell with a pulse and measuring the complex recovery curve (RLC) relative to the unloaded voltage I could determine the state of charge which was accurate and even accounted for memory effect (where an apparently fully charged battery may be able to provide much less than the nominal full charge).

So hardly a new idea; the company I was at did file disclosure to the USPTO (I was in the USA at the time) so it should stand as reasonable prior art.

[Pascal Monett]

Sounds nice

Call me when you sell it in a handheld format, where I put the battery I want to test in, press a button and get a cycle number on a LED screen.

And, obviously, it has to be able to test all three common battery formats (AAs and AAAs especially).

[jonathan keith]

Re: Sounds nice

... and it has a USB-C connector allowing testing of phones and laptops.

[Bogbody]

When will a certain Fruity brand use it to predict battery life and then arrange for your new Fruit to be delivered, debiting your account of course.

Cynic? Moi? :-)

[Chemist]

Whilst on this topic..

.. for a while now I've been using a bash script to control the charging of a laptop.

Script monitors the limits set for chsrge/low charge - signals one of the house pis equipped with remote control board that can turn (4) on/off mains sockets - laptop thus kept within limits. Anyone care to suggest what the limits should be ?? . I'm using 55%-85% at the moment so that there is a reasonable reserve for off-mains running. Of course you need a scripted way of turning off charging before hibernation etc.

Using this on an older laptop has kept the battery capacity at 96% after 4 years.

If you have a spare pi the remote sockets are useful for all sorts of things and very cheap

[Version 1.0]

Re: Whilst on this topic..

Laptop chargers are designed to keep the battery charged, nothing else. A dead battery often means a new laptop sale so there's virtually no reason to design a system that extends the battery life. It makes much more economic sense to predict the end of life.

[bombastic bob]

Re: Whilst on this topic..

actually... battery life _IS_ a selling point for laptops, slabs, and even phones.

There are 2 factors to consider:

a) weight/size

b) capacity

You can have a lightweight laptop with lousy capacity. You can also have a laptop with a bozillian hours of operation, but it weighs a lot or is 3 times as thick because of all of the batteries. Then there's replaceable vs soldered-in batteries [a big problem nowadays with slabs and phones at least].

So if you can predict the number of charges on a battery BEFORE you install it, you can (literally) grade them and select "the best" to go in the higher end devices, etc.. and ALSO let people know when it's time to "buy a new one" for the cheap stuff and/or automatically adjust for a repaired device [new battery].

Anyway, this is how I see it. But I also expect that only very new designs will start using something like that, maybe as soon as a year or two from now. [there is a predictable time delay from innovation to implementation and finally getting out to market].

[a_yank_lurker]

Is this new?

This sounds like the patents that basically took a tedious manual process and did it on a computer with a bunch of buzzword bingo. Reasonably accurate determination of battery's useful life has been something people have been interested to some extent at least for 100+ years because batteries have been in use for numerous semi-critical applications even then.

[Anonymous Coward]

Re: Is this new?

But historically, it was about the new market.

Now the used market is going to be a lot bigger.

[rcxb]

There's nothing necessarily AI about this at all. AI seems to be a buzzword for just any computer algorithm at this point.

[Smndnm]

Stop making up words

Oscillatory signal.... really? A signal that has oscillations in it? So perhaps it is an oscillating signal, yeah that’s it, an oscillating signal... being a signal that oscillates... ffs... don’t bastardise the English language more than you have to.

[General Purpose]

Re: Stop making up words

"Oscillatory signal" is correct and not a neologism - try searching for it. Waving a white flag is an oscillatory signal. An oscillating signal is one that signals different things from one moment to another, like waving a flag that's white on one side and a skull-and-crossbones on the other.