Dr. Kai-Philipp Kairies working in a meetingDr. Kai-Philipp Kairies working in a meeting

This paper will explain:

Battery lifetime is a concern that many people share. Whether it’s a smartphone, home storage battery, or electric vehicle, there is a nagging worry that the battery will not last as long as expected.  The key to getting the most life out of a battery is knowing what is happening within the battery. As soon as we understand the influencing factors and quantify their impact on the degradation of the battery, we can find the most suitable countermeasure to aging. This is where the different aging models to predict battery lifetime come into play.

Four Common Aging Models to Predict Battery Lifetime

Aging models for predicting battery lifetime
Aging models for predicting battery lifetime
  1. Empirical/Data-driven
    These models are purely based on empirical observations using large amounts of data. They aim to find and leverage the strongest correlations between a set of parameters that characterizes the usage of a battery and its health status. Accordingly, these models are objective in the way they operate.
  2. Semi-empirical
    These models resemble purely empirical aging models but also include some battery expertise. For example, when characterizing the usage of a battery, a battery expert can tell which characteristics knowingly influence battery aging. These parameters are then extracted from the data and may serve as inputs to the aging model. Battery expertise can also be used to categorize influencing factors and be applied to other relationships that have been investigated thoroughly.
  3. Mechanistic
    The mechanistic modeling approach goes one level deeper into what is happening within the battery. In addition to estimating high-level characteristics like residual capacity, it tracks and predicts a minimum of three battery internal parameters: loss of lithium inventory (LLI), loss of active material on the negative electrode (LAMNE), and loss of active material on the positive electrode (LAMPE). This adds complexity to the model but increases the accuracy when forecasting future degradation.
  4. Physical-chemical
    Physical-chemical aging models go even one step further than mechanistic aging models. This approach aims to model aging mechanisms separately using the necessary physical-chemical equations. Evidently, this model provides the deepest look into what is causing the battery to age. But these insights come at the cost of extensive parametrization efforts. It requires a large set of parameters that often depend on the materials used as well as the cell design. Shortcuts are not always practical since an inaccurate parametrization can easily become a bottleneck for a physical-chemical aging model's overall performance and accuracy.

Benefits of a Mechanistic Battery Aging Model

First things first: All these modeling approaches have their pros and cons. None of these approaches is generally superior to all the others. So, in order to find the best fit it is important to evaluate the target use case. We, at ACCURE, make sure we use the right modeling approach under the right conditions.

The benefit of the mechanistic model is the compromise that it represents. On the one hand, it provides a deeper understanding of what is happening inside the battery while it is aging. While on the other hand, it does not require complex parametrization. The mechanistic model is a compromise. It provides accuracy and insights without the complexity of a physical-chemical aging model.

Accurate LLI and LAMPE Estimation Using the Mechanistic Modeling Approach with Layered Oxides
Accurate LLI and LAMPE Estimation Using the Mechanistic Modeling Approach with Layered Oxides

A mechanistic model breaks down aging from cell level to electrode level. It specifies in which ranges the negative and positive electrodes are utilized when the cell is operated. The mechanistic model is also able to imitate any arbitrary combination of loss of lithium inventory (LLI), loss of active material on the negative electrode (LAMNE), and loss of active material on the positive electrode (LAMPE). Moreover, it can detect when an electrode is too close to an unsafe operating range. A popular example is the detection of thermodynamic lithium plating on the negative electrode. 

Mechanistic modeling is anything but a trivial task. Recently, we collaborated with Matthieu Dubarry (Ph.D.) and the Hawai'i Natural Energy Institute (HNEI) to push the boundaries of mechanistic modeling. The results of this collaboration can be found in our recent open-access publication.

Interestingly, we found that most state-of-the-art mechanistic models are inaccurately modeling the loss of active material on the positive electrode. 

More importantly, though, we propose the necessary changes to fix this issue. 

Through our work with leading battery researchers, ACCURE is helping the world better understand lithium-ion batteries.  By building better mechanistic models, we will all get a little more life out of our batteries.  

About ACCURE Battery Intelligence

ACCURE helps companies reduce risk, improve performance, and maximize the business value of battery energy storage. Our predictive analytics solution simplifies the complexity of battery data to make batteries safer, more reliable, and more sustainable. By combining cutting-edge artificial intelligence with deep expert knowledge of batteries, we bring a new level of clarity to energy storage.  Today, we support customers worldwide, helping optimize the performance and safety of their battery systems. Visit us at accure.net.

Matthias
Kuipers
Senior Battery Expert at ACCURE
About the author

Matthias

Kuipers

Matthias helps customers gain actionable insights into their battery systems. As Senior Battery Expert and leader of the battery expert team at ACCURE he is responsible for the development of leading-edge battery diagnostics and analytics. As an accomplished systems engineer, Matthias has extensive experience in automotive battery system development. He holds a Master of Science in Electrical Engineering, Information Technology and Technical Computer Science. In his free time, he enjoys sports including soccer, skiing, snowboarding, and jogging.

Connect with me

You may also like...