Tight battery management in the applications is key to securing long life and safe operation. In this session we will review reliability requirements and the use of simulation tools to aid the design and validate reliability and safety.
Session Chairman: Bob Taenaka, Technical Leader, Advanced Battery Systems Electrified Powertrain Engineering, North American Product Development, Ford Motor Company
Bob Taenaka is a technical leader in Advanced Battery Systems at Ford Motor Company, responsible for battery cell selection/validation in support of Ford's present and near-term future production hybrid and electric vehicles. Bob's team also carries out battery system sizing, performance, and life modeling/validation activities. In this role, Bob is also responsible for technical oversight of battery cell suppliers, helping to bring their cell design and manufacturing quality processes to automotive standards. Prior to joining Ford in 2001, Mr. Taenaka spent 18 years with Hughes Space & Communications, serving as battery engineer for the Galileo Probe mission to Jupiter; principal investigator or program manager for several nickel-hydrogen and sodium-sulfur battery development efforts, and had responsibility for in-orbit support on battery usage for satellite customers and ground stations.
SESSION AGENDA
Battery Cell Engineering and Product Reliability Bob Taenaka, Technical Leader, Advanced Battery Systems Electrified Powertrain Engineering, North American Product Development, Ford Motor Company
Electrified vehicle batteries are commonly designed to provide at least 10-year, 150K-mile life. To meet this target, automakers must carry out a series of steps to both understand and demonstrate battery cell reliability prior to vehicle launch.
This presentation will outline those steps, which include:
Selection of an appropriate battery cell design and supplier
Design and process Quality documentation and review
Design verification testing and analyses
Ford launched its first production electrified vehicle -- the Escape Hybrid -- in 2004; this presentation will also address demonstrated battery cell reliability across more than 10 years of electrified vehicle production at Ford.
Close Abstract
Simulation of Battery Mechanical Abuse and Thermal Runaway John Turner, Chief Computational Scientist, Oak Ridge National Laboratory
We report on progress and remaining challenges related to simulation of mechanical abuse of batteries leading to internal short-circuits and mechanisms leading to thermal runaway. An integrated program including fundamental experiments, abuse tests, and simulation is under way and will be described as follows.
Review of motivation and goals
what are the challenges and how can simulation provide insight?
discussion of layer-resolved and homogenized models
Experiments
abuse tests being used to trigger internal shorts and thermal runaway (e.g. pinch tests)
fundamental tests under way to determine properties of components and also behavior of components in a layered configuration
Simulations
the Virtual Integrated Battery Environment (VIBE), an open source tool for simulating mechanical, electrochemical, and thermal responses of prismatic and cylindrical batteries
mechanics simulations used to initiate virtual short-circuits
VIBE offers fast-running approaches for electrochemistry such as the Newman-Tiedeman-Gu (NTG) and Doyle-Fuller-Newman (DFN, a.k.a. DualFoil) models as well as a new highly-resolved 3D volume-averaged formulation known as AMPERES (Advanced MultiPhysics for Electrochemical and Renewable Energy Storage). We provide an update on deployment of VIBE and how it is being used in virtual crash scenarios.
Battery State and Internal Variables Estimation Using a Reduced-Order Physics-Based Model of a Lithium-Ion Cell and a Nonlinear Kalman Filter Gregory Plett, Professor, University of Colorado at Colorado Springs
Battery management systems (BMS) must continuously provide accurate estimates of available energy and power. These computations are based on estimates of internal battery state (e.g., state-of-charge). Present BMS base these latter estimates on equivalent-circuit models (ECM) of cells; however, ECMs are unable to provide information on the internal electrochemical processes that drive cell degradation. In order to provide estimates of available power and energy that take cell aging into account, physics-based models (PBM) must be used instead. In this paper, we address the problem of estimating the internal state and values of the internal electrochemical variables of a cell using a nonlinear Kalman filter and a reduced-order PBM. The method uses readily available measurements of voltage, current, and temperature only; simulation results agree closely to truth values and are robust to incorrect initialization; and automatic confidence intervals on estimates allow their use in follow-on applications.
This presentation addresses the following:
Background
The need for state estimates
Equivalent-circuit models versus physics-based models
Practical reduced-order physics-based models
Method
What is a Kalman filter? How does it work?
Extended Kalman filters and sigma-point Kalman filters
Estimating internal electrochemical variables in addition to model state
Results
State-of-charge estimates: robust even to bad initialization
Internal electrochemical variables estimates
The presentation concludes with a summary and a discussion of future application of the internal-variables and state estimates produced by these methods.
How to Design Battery Packs for HEVs: Cell Modeling and Thermal System Gaetan Damblanc, Technical Lead, CD-Adapco
This two-part talk will show attendees how to improve the specifications of a commercially available cell and simulate the performance of an HEV battery module using commercially available CAE software. Utilizing cutting-edge technology, this talk will help demonstrate how you can derive real benefits.
Within the presentation, the characterization process is explained to highlight and explore the key parameters necessary for the cell performance model. Presenters examine how software technology like can quickly and accurately adjust cell design with regard to enhancing performance.
For the second part of the presentation, utilizing software resolving simultaneously for the electrochemistry and heat transfer problems, the enhanced cell will be used to virtually generate a module of an HEV battery pack to simulate its thermal signature under a standard drive cycle.
What We Can Learn From Transient Methods in the Characterization of Li-Ion Batteries Ed Fontes, Chief Technology Officer, Comsol
This presentation is about modeling of the Li-Ion battery using transient models based on accurate descriptions of transport and reactions in the cell, solving the numerical model equations, and performing parameter estimation using experimental data. Despite the insight in the words of John Von Neumann, “With four parameters I can fit an elephant, and with five I can make him wiggle his trunk”, we can show that although we may have precisely five parameters to fit to experimental data, these parameters have varying relevance for different time scales. Using accurate models in combination with transient experiments may therefore give us relatively accurate values of key parameters in the performance of a battery system.
The major steps in modeling and simulation evolve as models are understood and validated. The list below describes the most common steps.
Understanding the studied system: In this step, the mathematical model is formulated, discretized to generate a numerical model. The numerical model is then solved. Studying the results from the solution of the numerical equations, for many sets of input data and conditions, yields the understanding that makes it possible to take the next step, which is to make predictions about the system.
Making predictions from the model results: Once the system is understood, it is possible to make predictions about the system’s behavior and, if relevant, improve it. Modeling and simulation software is then used as a “what if”-tool for checking the outcome of these predictions. The predictions from these studies are then validated to experimental data.
Performing parameter estimation, optimization, and automatic control: Once the first validations are done, modeling and simulation software can be used for further polishing the models to make even better predictions. Parameter estimation implies extracting data by combining experiments and solutions from numerical models. With an accurate set of parameters, such models can be used for optimization of a design or process. If relevant, these accurate models can be used for automatic control of a process or a device.
Modeling and simulation software is a very cost effective tool for developing ideas, making predictions, and optimizing a device or a process.
The presentation further shows how the methodology in 1, 2, and 3 can be made useful and efficient for scientist and engineers that have only a limited experience in modeling and simulations, but that have expertise in battery systems or the use of battery systems.
