AABC 2014 LLIBTA Symposium - Large Lithium Ion Battery Technology and Application - Track B: Battery Engineering and Application - Session 1

Figure 1 Circuit representing RCR model. Voltage (V_o), series resistance (R_o) and parallel impedance (R_p|C) depend on temperature and state of charge.

Thermal design of lithium-ion battery packs for HEV/PHEV is essential for ensuring a useful life of ≥8 years. To achieve such a long life, the typical requirements are that the pack temperature should be kept below ~50 °C and the maximum variation between and within cells be ~2 °C or less. These requirements are best achieved by a virtual design process wherein a computer model of the battery back is subjected to a drive cycle and the thermal response reported in terms of the requirements. A major challenge in realizing a successful virtual design process is obtaining an accurate and computationally efficient computer model to simulate the electrothermal behavior of the battery.

Equivalent circuit models are widely used to represent battery behavior, for example the RCR model (see Figure 1). The RCR model is computationally efficient, can be readily fit to experimental data, and applicable over a wide range of temperatures.

Figure 2. Comparison of RCR model prediction of US06 drive cycle (PHEV charge depleting) to synthetic data.

In order to characterize the applicability of the RCR model, a set of virtual cells based on a physics-based model were constructed such that each cell had a different dominant mode of voltage loss, specifically ohmic, activation, and diffusion.  Each virtual cell was exercised by simulating a USABC hybrid pulse power characterization (HPPC) test to generate synthetic data that were fit to the RCR model, and the quality of the fit assessed.  The parameterized RCR model was used to simulate a drive cycle and the quality of the prediction assessed.  This work provides useful guidance on how to test cells for creating data useful for parameterizing the RCR model, such as what currents and temperatures should be used in the tests. The case of a cell whose polarization is mainly ohmic is shown in Figure 2.
For the case where ohmic losses dominate, the RCR model gives excellent fits over a wide range of currents and temperatures.

Physicochemical processes in lithium batteries occur in intricate geometries over a wide range of time and length scales. This imposes difficulties in modeling battery responses. For example, battery “states” are not determined as point-functions, but path-dependent quantities. As the size of the battery increases for high-energy and high-power demand in electric vehicle applications, macroscopic design factors in combination with highly dynamic environmental conditions significantly influence the electrical, thermal, electrochemical, and mechanical responses of a battery system. Without better knowledge of the interplays among interdisciplinary multiphysics occurring across varied scales in the battery systems, it is costly to design long-lasting, high-performing, safe, large battery systems. Overcoming challenges in modeling the highly nonlinear multiscale response of battery systems, we suggest a nonlinear multi-scale multi-domain model. DOE’s Computer Aided Engineering for Electric Drive Vehicle Battery (CAEBAT) program has focused on developing innovative modeling capabilities to help industries accelerate mass-market adoption of electric-drive vehicles (EDVs).

In this talk, we will discuss;

• What makes modeling batteries difficult
• How the nonlinear multiscale model has been applied to investigate various scientific and engineering problems
• What we propose to advance the battery model to bring transformational change in use of mathematical models accelerating the breakthroughs necessary for battery and EDV industries

Silicon as a negative electrode material has received a lot of attention in the past few years and with good reason. Li-Si materials offer an energy density of about ten times that of conventional carbon materials. On the other hand, the volume changes associated with the incorporation of large amounts lithium in silicon raise several concerns for cell design. One major issue is the pulverization of the active material. This decrepitation and continual solid electrolyte interphase growth results in rapid capacity fade.

In this presentation, these challenges are explored both with a physics-base model of a cell sandwich and experimentally. Cell design and operation of Li-Si full cells are compared and contrasted with more traditional lithium ion batteries.

The performance of a lithium-ion battery (LIB) cell is influenced by the aspect ratio of electrodes and the size and placement of the current collecting tabs. If an electrode is not designed optimally, the utilization of the active material on the electrode will be non-uniform, inducing an inhomogeneous temperature distribution in the battery during high power extraction, which might cause battery degradation and thermal runaway. This effect becomes more pronounced as the size of the electrode increases.

This presentation outlines a modeling approach to be used when scaling up a small-scale LIB cell to a large-scale one and a thermal modeling for the thermal design of an LIB cell as in the following:

• Mathematical model for the potential and current density distributions in an LIB cell
• Cell design parameters
• Uniformity index of depth of discharge (DOD)
• Scale-up modeling and its validation
• Thermal modeling and its validation
• Effect of electrode design on the thermal behavior of an LIB cell

In this presentation we report on the improvement of cell kinetics in Lithium-ion cells by the use of hard templated porous carbons with dominant macroporosity between 100-1000nms.

NMC cathode slurry formulations have been mixed and homogenized with carbon black, PVDF binder and further addition of different amounts of the hard templated carbon Porocarb LD2N.

The cathode slurry has been coated on Al foil with standard equipment and subsequently calendared to achieve the high electrode density required by high-energy cells.

The porous particles within the electrode volume retained their initial porosity and pore size distribution whereas the remaining volume consisting of active materials and conductive additives was densified to a lower level of porosity. Thus, a smooth electrode could be produced with local areas of high porosity and macro pores.

The cathodes have been assembled to half cells and cycled at different C rates to observe the effects of the new electrode structure on Lithium-ion transport kinetics.

We are able to show:

• Electrode cycling shows improved effective Lithium-ion conductivity of the electrodes with different areas of porosity.
• A substantial capacity improvement during cycling at medium to high C rates in comparison to a standard cathode without additive can be seen.
• With small addition of the macroporous carbon particles to the slurry the electrode thickness can be at least doubled without losing discharge rate performance of the cell.

Creating local areas of high porosity within the electrode with porous carbons therefore finally leads to improved volumetric and gravimetric energy densities.

Furthermore, manufacturing costs from coating to cell assembly can be significantly lowered due to increased electrode thickness.