AABC 2014 LLIBTA Symposium - Large Lithium Ion Battery Technology and Application - Track B: Battery Engineering and Application - Session 3

We have developed a battery state estimator based on a finite impulse response (FIR) filter. Simulation results indicate that the estimator gives accurate prediction and numerically-stable performance in the regression of filter coefficients and the open-circuit potential, which yields the battery state of charge. The estimator is also able to predict battery power capabilities. Comparison of the measured and predicted state of charge (SOC) and the charge and discharge power capabilities (state of power, SOP) of a Li-ion battery are provided. This new approach appears to be more flexible than previous battery state estimation approaches; we show that the FIR filter can capture the behavior of batteries governed by various physicochemical phenomena.

Talk outline

• Automotive use of lithium ion cells: motivation and products
• Battery State Estimation

• Trends on monitoring and control of traction batteries
• A generalized approach for linear systems
• What’s next?
• Summary and open questions

Robert Bosch Battery Systems develops and produces plug-and-play Li-ion battery packs and stand-alone battery management systems for multiple automakers.  Key functions of the resulting products are to extend cell life, predict cell performance, provide diagnostic services, and to avoid potential hazards.

This presentation illustrates steps to identify battery management system requirements, examines available options to fulfill requirements, and considers approaches to validation.

Identifying battery management system requirements:

• Understanding the characteristics of the selected cell type
• Developing a strategy to avoid potential hazards, considering the entire system
• Prediction of battery state (SOC, SOH, and power available)
• Achieving combined battery cycle and calendar life targets
• Identifying strategies for diagnostics

Examining options to fulfill requirements:

• Hardware methods
• Software methods

Validation:

• Application of MiL and HiL
• Systems-level and vehicle-level
• EMC and environmental

The presentation concludes with a summary and a discussion of future concepts for battery management.

The battery is one of the essential parts for electro mobility (HEVs, PHEVs and BEVs) and probably the most critical one with respect to reliable operation and life time expectation. A successful operation of an electric powered vehicle requires an intelligent adaptable battery management system, which observe the battery, deliver a precise status of the battery and operates the battery at the best compromise between high efficiency/power performance and best live time. Accurate and reliable knowledge about the state of the battery is a must for realization of defined energy management strategy. Well known key parameters of the battery state are e.g. state of charge (SOC), state of health (SOH) and the state of function (SOF).

The state of charge defines the “fill level” of the battery and that value is corresponding to e.g. distance the vehicle can drive electrical powered or used for calculation of acceptable recuperation power. However, the exact definition of the SOC depends on different parameters like e.g. precision of measurement electronics, time of operation, type of cell chemistry.

State of health (SOH) is standing for a measure of the battery aging (actual state of the battery related to battery condition at the beginning of battery life, BOL) and is reflecting the general condition of a battery and its ability to deliver the specified performance compared with a battery at BOL. It takes into account such factors as internal resistance, self-discharge and charge efficiency. During the lifetime of a battery, its performance or "health" tends to deteriorate gradually due to irreversible physical and chemical changes which take place with operation and with calendar age until eventually the battery is no longer usable for the target application. The SOH is an indication of the point which has been reached in the life cycle of the battery and a measure of its condition relative to a battery at BOL. In general the definition of aging is depending on the application field of the battery, like HEV, starter (SLI) or EV.

The exact knowledge of the state of charge (SOC) of the battery, state of health of the battery (SOH) and a reliable prognosis of its state of function (SOF) or power capability, respectively, are essential for the successful operation of electric powered vehicles independent on at the end on the vehicle type (HEV, PHEV or BEV).

Therefore a concept of combining different modules such as coulomb integration, OCV analysis, power analysis, which may be supported by battery specific submodules has been chosen and implemented. These modules are interacting with each other in such a way that the model which may give the best prediction for an actual state of the battery is used. The combination of different methods/modules may give the requested accuracy and allows reducing hardware costs for high precision sensors at the same time.

A Battery Test Facility (BTF) has been completed at the EPA National Vehicle Fuels Emission Laboratory (NVFEL) to test various automotive battery packs.  Battery pack tests are performed in the BTF using a power processing system, testing controller, battery pack conditioner, and a temperature controlled chamber.  For e-machine testing and power pack testing, a variety of different battery packs are needed to power these devices to simulate in-vehicle conditions.  For e-machine testing and development, it is cost prohibitive to purchase a complete range of battery pack chemistries, along with the necessary battery management system communication protocols and signals for various manufacturers.

Therefore, there is a need to accurately emulate battery pack voltage, power, SOC, etc. for testing e-machines as well as performing real-time HIL vehicle simulations by having the ability to instantly select a cell chemistry along with battery pack configuration such as cell capacity, number of cells, etc.

This paper presents lithium-ion battery pack HIL (Hardware-In-Loop) development and validation integrated into the EPA Battery Test Facility as follows.

• HIL Model (Equivalent circuit cell model, Lumped thermal model, and BMS controls)
• HIL System Setup (Power Processing, Battery Pack Conditioner, Battery Pack, Test Automation)
• HIL and BIL (Battery-In-Loop) Tests and Preliminary Validation

The validation was performed by simultaneously stimulating the HIL Battery Model and the lithium-ion battery pack with FTP UDDS, Highway (HWFET), and US06 drive cycles.

Battery pack voltage, current, SOC, pack temperature, etc. were validated with

• Nissan Leaf EV
• GM Volt Range Extended Vehicle
• Summary
• Future Work

Lithium Ion batteries are complex systems, and techniques to cost effectively monitor and manage important performance measures while predicting battery cell degradation and failure remain a key technological challenge. Traditionally, battery monitoring has been performed using voltage, current, and limited temperature measurements on individual cells across the pack.  As part of an ARPA-E funded research program, General Electric (GE), University of Michigan (U-M), and Ford Motor Company are developing a novel multi-measurand sensing system with supporting multi-physics models to improve cell utilization for electric vehicle applications.  This presentation will give an overview of that program including:

• Program objectives and scope
• Novel temperature and expansion sensor technology
• Multiphysics modeling of the coupled electrochemical, thermal and mechanical dynamics of the cell
• Parameterization of these models and coupling of the models to measurements provided by the novel sensing system
• Validation approach to evaluate the benefits of this system on 5 Ah cells (NMC/carbon)

Hybrid and electric vehicle (xEV) battery systems today use conservative design approaches and multiple redundant layers of safety to compensate for the lack of real-time, directly sensed cell internal conditions during operation. Under the ARPA-E AMPED program for advanced battery management systems (BMS), PARC and LG Chem Power (LGCPI) are developing SENSOR (Smart Embedded Network of Sensors with an Optical Readout), an optically based smart monitoring system prototype targeting batteries for xEVs. Key developments proposed under this project include:

• Cell-embedded fiber-optic (FO) sensing with a compact, accurate readout.
• Smart algorithms to maximize the potential of internal cell parameters for BMS.
• Initial technology validation for xEV applications.

This presentation will give an overview of the project, the underlying enabling technologies, and then cover some promising initial experimental results from the project. At a high-level, the following questions will be answered, focusing on the listed aspects:

1. What novel cell parameters can one monitor with SENSOR?

• FO sensors to monitor internal cell parameters
• Example of applicable FO sensor and working principle
• Utility of these internal parameters for cell state estimation
• Can SENSOR be made practical, low-cost, deployable?

• PARC’s low-cost, compact, wavelength shift detection technology
• Implications for cell-embedded FO monitoring in xEV packs
• Initial results from small-format Li-ion cells with embedded FO sensors
• Can SENSOR improve pack control, utilization?

• Intelligent algorithms for SOX (state-of-charge/health/power) estimation
• Faster charging rates feasible with internal cell state monitoring
• Enhanced pack power and energy utilization possibilities with SENSOR

The presentation will conclude with a discussion on next steps and technical targets set for the remainder of this project.

In recent years there has been increasing interest in vehicle electrification as it has the potential to both decrease greenhouse gas production and improve energy independence.  However, to realize these benefits, Plug-in Hybrid Electric Vehicle (PHEVs) must have battery packs that do not prematurely degrade.  Accurate modeling of battery health can help both manage and mitigate battery degradation in a variety of ways.  However, to have accurate health models one must conduct battery aging experiments to calibrate them. This Poster +8 will discuss an experimental framework that uses information theory and modeling to design experiments for identifying battery aging models.  A demonstration of this framework is provided by a battery aging case study.

Experimental Framework:

1. Decide the scope of the cycling and/or aging experiment
2. Create representative cycles (CCCV, drive cycle or other) – through either simulation or experimentation, to describe all of the possible experimental trials
3. Decide on a model form for the aging experiments
4. Use optimal experimental design algorithms, such as DETMAX, to select a set of cycles that maximizes the information gained and thus minimizes the uncertainty when fitting model parameters
5. Conduct the experiment and use well established methods to fit the parameterized aging model

Battery Aging Case Study:

1. Used the above experimental framework
2. Consisted of cycling 14 LiFePO₄ cells for over a year
3. Resulted in an aging model that was fit to the battery cycling data