RESEARCH
Energy Group
ENERGY GROUP - OVERVIEW
Mission Statement
"At the forefront of battery informatics, our group is pioneering novel approaches for analysis, diagnosis, and prognosis of rechargeable batteries to ensure their efficient and safe usage."
Research Interests
- Model development
- System identification
- Aging diagnosis
- Optimal scheduling
Our detailed Works
1. Model development
Mathematical models are milestone for battery informatics, so we are trying to develop an advanced model for lithium-ion batteries.
We develop meso- and micro-scopic electrochemical models as well as continuum scale models. Using them, we mainly conduct morphological analysis such as investigating of battery performance or degradation affected by morphological features. Also, an analytic steady state model is used to prove other numerical models.
2. System identification
Battery system(or state) identification is a necessary step to practically use the models, which is an optimization to match our models to target real batteries. In general, this process includes solving a highly nonlinear and complex problem, which requires a very expensive computational cost and is hard to obtain an accurate solution. Thus, we are trying to improve optimization algorithms for that in novel ways.
3. Aging diagnosis
Battery diagnosis is crucial to ensure battery safety. We develop advanced battery aging estimation algorithm using empirical models, electrochemical models, and neural networks. Further, we are trying to improve aging estimation by reflecting several degradation phenomena, e.g., particle cracking, lithium plating, and so on.
4. Optimal scheduling
Based on battery model and state estimation, we design optimal control methods. Specifically, we control input current to decrease battery degradation or improve state estimation. For this, we usually exploit reinforcement learning.