ARPA-E awards WUSTL $2M to develop predictive battery management system for plug-in vehicles; targeting more efficient use
6 August 2012
A team of engineers at Washington University in St. Louis (WUSTL) will receive $2 million from the US Department of Energy’s ARPA-E to design a predictive battery management system for lithium-ion batteries to guarantee their longevity, safety and performance. The School of Engineering & Applied Science is providing $1.2 million in matching funds, for a total project investment of $3.2 million.
The APRA-E award goes to the Modeling, Analysis and Process-control Laboratory for Electrochemical systems (MAPLE) in the Department of Energy, Environmental & Chemical Engineering, led by Venkat Subramanian, PhD, associate professor. The targeted predictive battery management system will use innovative modeling software to optimize battery use. The system will predict optimal charge and discharge of the battery in real-time, enhancing battery performance and improving battery safety, charge-rate, and usable capacity.
The project is one of 12 that won funding from the DOE’s Advanced Research Projects Agency-Energy (ARPA-E) under the new AMPED program that focuses on innovations in battery management and storage to advance electric vehicle technologies and to help improve the efficiency and reliability of the electrical grid. (Earlier post, earlier post.)
I want to give credit to my doctoral students. Without their efforts, we wouldn’t have been able to submit a proposal. The solicitation came and we had two weeks to respond after the team was formed, and then we got a review and we had to respond over the weekend.—Venkat Subramanian
In addition to Subramanian, the team includes doctoral students Venkatasailanathan Ramadesigan, Paul Northrop, Sumitava De, Bharatkumar Suthar and Matthew Lawder.
Earlier this year, Hyundai Motor Company (HMC) awarded a $100,000-contract to the MAPLE Lab to incorporate electrochemical model-based code in an advanced battery management system for automotive batteries. (Earlier post.) While the HMC project was only for modeling efforts, Dr. Subramanian notes that the main difference for the ARPA-E project would be the conversion of the fast codes to a microcontroller environment and the development of a patentable product, and significant validation.
If Li-ion batteries are charged too quickly, they can heat up and may explode. To avoid catastrophic failure, manufacturers overdesign the batteries and use only part of their energy capacity per cycle, Subramanian says.
The goal of the AMPED program is to push the current technology to 100-percent efficiency, while making sure battery lifetime is not compromised. If you can predict what will happen inside the battery, you can push the battery to do more per cycle. Currently empirical (experience based) models that have no predictive capability are used to manage the batteries. This is why manufacturers over-stack the material; they have no idea what’s happening inside.—Venkat Subramanian
The Battery Management System (BMS) the MAPLE lab will develop will keep the battery operating optimally, enabling maximum utilization of energy at all times. Pushing the battery technology closer to 100% efficiency would ultimately reduce the weight of the car by enabling the use of a smaller pack and improve its overall energy efficiency.
There are physics-based models of lithium-ion batteries but they are computationally intensive and can’t be solved in real time by the usual methods.
The MAPLE engineers plan to use a class of simulation techniques called spectral methods aided by mathematical analysis to solve a physics-based model’s differential equations. Spectral methods should allow them to cut down on the model’s computational demands so that it runs faster.
In general, people write mathematical models and then plug them into commercial software to solve them. We relish solving the models ourselves to see if we can find more elegant ways to do it. That’s the overarching theme of our work.
We are also interested in re-examining predictive models of importance for medicine, such as those used in medical imaging, to see if we can solve them faster but with the same accuracy so that they can be used in real-time sensing and control.—Venkat Subramanian
V. Ramadesigan, P. W. C. Northrop, S. De, S. Santhanagopalan, R. D. Braatz, and V. R. Subramanian (2012) Modeling and Simulation of Lithium-Ion Batteries from a Systems Engineering Perspective, J. Electrochem. Soc., 159(3), R31-R45 doi: 10.1149/2.018203jes
TrackBack URL for this entry:
Listed below are links to weblogs that reference ARPA-E awards WUSTL $2M to develop predictive battery management system for plug-in vehicles; targeting more efficient use: