Knowing exactly how power and capacity is fading within an electric vehicle can improve driving-range estimates, cut costs for manufacturers, and prolong battery life.

A new model from Stanford University offers that kind of precise understanding of a lithium-ion battery's internal workings, in real-time.

To predict the battery’s remaining storage capacity and charge level, the new algorithm combines sensor data with computer modeling of the physical processes that degrade lithium-ion battery cells.

The research appears in the journal IEEE Transactions on Control Systems Technology .

The Physical Processes Behind a Degrading Battery

When a battery charges and discharges, lithium ions continuously shuttle between electrodes. The model developed at Stanford University provides a better understanding of the lithium concentration – a measurement that has been frequently miscalculated, according to the researchers..

Models that support existing battery managements systems assume that the amount of lithium never changes in each electrode. In fact, the lithium fades.

"Lithium is lost to side reactions as the battery degrades,” said lead study author Anirudh Allam, a PhD student in energy resources engineering . “So these assumptions result in inaccurate models.”

Along with Stanford energy resources engineering professor Simona Onori, Allam designed a system that measures the lost lithium and offers continuously updated estimates of lithium concentrations.

A dedicated algorithm for each electrode adjusts based on sensor measurements as the system operates. Onori and Allam validated their algorithm in realistic scenarios using standard industry hardware.

With a dSPACE MicroAutoBox-II 1401/1513 prototyping unit, the team tested its algorithm on the Urban Dynamometer Driving Schedule (UDDS), which represents city driving patterns.

Smaller Cells, Bigger Range

A battery pack, consisting of hundreds or thousands of small battery cells, typically accounts for about 30% of total vehicle cost. Better estimates of a battery’s actual capacity will enable a smaller buffer and a smaller battery pack, according to Onori.

“If you have more certainty around how much energy your battery can hold throughout its entire lifecycle, then you can use more of that capacity,” said the Stanford professor. “But our system reveals where the edges are, so batteries can be operated with more precision.”

According to the researchers' study, validation results against experimental data demonstrate a bounded capacity estimation error within 2% of its true value. The accuracy of the predictions may pave the way for greater driving range in electric vehicles.

In a short interview with Tech Briefs below, Onori explains how soon this type of battery model could be take the road.

Tech Briefs: In simple terms, what is your model able to determine about a battery?

Prof. Simona Onori: Degradation phenomena are very complex to understand and streamline and their interaction is even more intricate – and poorly understood. We focused on extracting – from voltage and current measurements – the change in internal electrochemical parameters (not one but a few) that are linked to this loss of performance over time. The fact that we use the already on-board available sensors make this a very promising and cheap solution to adopt.

Our solution makes use of electrochemistry, control system theory, optimization, and advanced mathematical modeling. In my opinion, it’s not easy nor common to have all those skills and knowledge at once.

Tech Briefs: In your press release you said: ""We have exploited electrochemical parameters that have never been used before for estimation purposes .” Can you explain more about what these electrochemical parameters are, and how they lead to a better assessment of the battery?

Prof. Simona Onori: What we did in our work was to investigate the main cause – in an electrochemical and mathematical framework – of degradation. When a battery ages, we observe power and capacity fade; these indicate an irreversible loss of the ability of a battery to store charge, and irreversible reduction of the rate at which electrical energy can be accepted or released by the battery, respectively.

It’s a commonly accepted fact that the main cause of degradation is the Solid-Electrolyte Interface (SEI) growth at the negative electrode  (for graphite-anode lithium-ion batteries). This is usually related to capacity fade by a quite-well-known mathematical relationship which is used for control/estimation tasks by the battery management system (BMS).

We weren’t happy with using only this established method, as the power fade was not directly accounted for by this model. We wanted to understand in a more comprehensive way the effect of degradation on other battery dynamics and parameters.

Tech Briefs: So, where did you go next?

Simona Onori, Stanford
Simona Onori, assistant professor of energy resources engineering in Stanford's School of Earth, Energy & Environmental Sciences

Prof. Simona Onori: We went further and found out a relation that links the ionic conductivity in the SEI layer (which decreases as the battery ages) to a resistance increase, and thus power fade. Porosity also decreases with age, and we linked this to the power fade.

Finally, we were able to write mathematical formulas to connect all of these parameters to the system level aging effects. This is the first new contribution in our research.

Tech Briefs: What other kinds of monitoring is occurring?

Prof. Simona Onori: The second contribution is the estimation algorithm that adaptively monitors such parameters – which change over the battery lifespan – to then provide a real-time capacity estimate. The new model, along with the novel estimation algorithm, have the ability to capture degradation effects due to calendar aging (as they will be manifest in terms of an impedance increase).

We show estimation range between two percent, in simulation. We went further, and tested the algorithm in a real hardware – which per se is a pretty exciting result as this has never been attempted before – and we obtained the same performance we obtained in simulation. Hopefully, one day – not too far in the future – our algorithm will be used in real cars.

Tech Briefs: What is possible when you have a such a precise understanding of a battery’s capacity?

Prof. Simona Onori:

  1. The driver can have a better estimate of the driving range.
  2. You have a more accurate estimate/knowledge of when the battery reaches end-of-life, and this is valuable information for the drivers/consumers, but also for the automakers as they can provide longer warranty periods.
  3. This is a very valuable information for recycling and reuse purposes, as knowing how much life is still left in the device would definitely facilitate the adoption to second-life application.
  4. The algorithm can even be used off-board of the vehicle in a separate device to run diagnostics on the battery.
  5. The calibration effort of this model/algorithm is quite limited as opposed to the expensive (in time and hardware resources) calibration process of equivalent circuit models which are the adopted in todays’ BMSs
  6. Our technology is not limited to a predefined chemistry but it is applicable to all the graphite-anode lithium-ion batteries.

Tech Briefs: How do you envision this technology being used on future vehicles?

Prof. Simona Onori: This technology could be used as soon as some automakers would be willing to take the leap. Maybe Elon Musk?

What do you think? Share your questions and comments below.