Research

Data and code

• Battery aging dataset: We generated a dataset from 215 commercial lithium-ion batteries of the same type controlled individually at the same operating conditions until >25% capacity fade.  All the data and associated code are shared [here], and a detailed description of the experiment is given in the [article].

• A Python tutorial about Physics-Informed Neural Networks (PINN) and Model-Integrated Neural Networks (MINN) is given at [Link]: MINN is our proposed physics-based learning framework that can learn the physics-based dynamics of general non-autonomous systems consisting of partial differential-algebraic equations (PDAEs). The obtained architecture systematically solves an unsettled research problem in control-oriented modeling, i.e., how to obtain optimally simplified models that are physically insightful, numerically accurate, and computationally tractable simultaneously.