The IEEE Control Systems Society (CSS) selected Anna Stefanopoulou and Jason Siegel to receive the 2016 Control Systems Technology Award "for the development of an advanced battery management system accounting for electro-thermo-mechanical phenomena."
The Control Systems Technology Award is bestowed "to a team or individual for an outstanding control systems technology contribution in either design and implementation, or project management."
This award recognizes the collaborative effort among researchers at the University of Michigan, Ford Motor Company and General Electric. Leveraging novel sensors, prognostics, and controls, the team has developed new, accurate methods for battery state estimation and algorithms which improve the utilization of existing Li-ion batteries.
Utilizing high-volume low-cost flex manufacturing, GE developed temperature and expansion sensors that monitor the real-time temperature and swelling of Lithium-Ion batteries due to Li-intercalation (Li-infusion) and thermal expansion. Due to their large size (1-2mm thick), packaging of existing temperature sensors such as thermistors limits their placement to areas away from the areas of highest temperature. “The new sensors are 10 to 20 times smaller, which means they can fit in between battery cells and give more precise and relevant temperature readings,” said Brian Engle from Amphenol Advanced Sensors, a partner in the work and battery sensor supplier. In addition, GE developed a new sensor for measurement of cell expansion based on the same sensor platform allowing for integration of that measurement into the battery management system.
Siegel and Stefanopoulou developed models and control algorithms guided the sensor placement and interpreted the information of small thin film sensors for temperature and strain. They collaborated with experts from the fields of mechanics (Krishna Garikipati, UM), vibrations (Bogdan Epureanu, UM), and mass transport phenomena (Charles Monroe, Oxford) to quantify the cell swelling at various charging and temperature conditions based on innovative experimental methods including neutron imaging (Dan Hussey, NIST) and specialized laboratory fixtures (Yi Ding, TARDEC, US Army).
“This is a great example of what is possible with multidisciplinary research and collaboration between industry and academia. On their own, new battery sensors or elaborate battery models have limitations, but together, they provide actionable and robust insight,” said Aaron Knobloch, Senior Scientist and Principal Investigator from General Electric Global Research.
The model-based estimation of critical states merging data from thin-film temperature and strain sensors led to four innovations in battery control.
- Real-time estimation of power capability enabling aggressive cell utilization at comparable aging. This innovation allows the control system to run closer to the performance limits accounting for the multiple time-scales of the electric, thermal, and mechanical stress.
- Optimized control for fast warm-up. A model-predictive algorithm of the temperature-dependent limits of battery charge acceptance were developed to achieve fast heating while protecting the battery from dangerous Li-plating and internal shorts.
- Improved state of charge estimation (SOC). Voltage measurements combined with Bulk stress measurements allow for higher accuracy at low SOC where higher confidence level is necessary.
- Advanced state of health estimation. By monitoring shifts in bulk stress peaks, the control system can quantify capacity fading and actively shift (control) the battery operation to maximize life.
Several of these innovations were tested on modules and a full battery pack from a Ford Fusion HEV by Dyche Anderson’s team at Ford demonstrating that the electro-thermal model-based algorithm provided a systematic way for
- Controlling the battery pack power under transient warm-up conditions
- Increasing the utilization of the cells by widening the operating state of charge (SOC) range
- Enforcing the time-average cell operation towards lower SOC for reduced degradation and higher acceptance of regenerative energy.
As part of an ARPA-e funded program, the team estimated the system impact of the enhanced control based on a 20% increase in energy utilization per cell. In the Ford application, this would result in cost reduction from battery downsizing and improvements in driveability from a faster cell warm-up. Initial long term test results in battery modules show a projected decrease in capacity of only 0.5% after 100,000 miles. Additional demonstrations are underway to show that the control techniques and measurements are broadly applicable to a variety of cell chemistries and packaging. The sensors, estimation, and control methods developed by the team have the potential for broader impact both across a variety of automotive platforms but also in grid storage systems. Predictive controls such as these will accelerate low-cost vehicle electrification, reduce fossil fuel use, and help the environment by safely storing large amounts of renewable energy.