Large-format batteries comprised of hundreds if not thousands of cells are becoming more common as battery energy storage moves to greater applications in automotive, grid and industrial scenarios. Such battery packs require thermal management (TM) to mitigate the accumulation of heat generated during routine and above-routine use. Thus, to preserve cell-to-battery life and performance, good TM should quickly account for variable thermal loads over diverse timescales and minimize cell-to-cell temperature variance referred to as “hot spots”. The need to streamline battery packs to minimal configurations (that can cause geometric bottlenecks for TM) and to reduce their cost will no doubt lead to persistence of hot spots for many applications. Hot spots will cause local cells to age faster than cells outside the hot spot region, with consequences toward decreased energy capacity, higher electrical and thermal impedance, and reduced power. Such disparate cell aging is problematic for large-format batteries in terms of management of cell replacements and ultimate profitability of the energy storage installation. This work is based on battery aging simulations done with INL CellSage, a physics-based software tool that considers cell and series string aging characteristics dependent on prevalent stress factors over time (Gering (2017)).We seek to understand how batteries will age in their intended applications, and whether the battery will perform through its warranty period. Central to this is Aging Path Dependence, which asserts the sequence of aging conditions over time has a direct influence on irreversible aging trends of battery systems. Physics-based tools like CellSage can help guide these issues. Typical input parameters within CellSage include cell chemistry, SOC, T and DT during duty cycle, location of use (US cities), duty cycle and frequency thereof, TM metrics, string attributes, and others. Duty cycles can be customized for particular applications and up to 25 cells are considered per series string, but that can be increased.First, Fig. 1 shows how a 10 ⁰C hot spot in the center of a 7-cell series string can affect capacity loss within local Li-ion cells. The cell chemistry is Gr/NCA, and the duty cycle is defined by scaled HEV power cycling at 80% SOC. The location of operation is Phoenix, Arizona and the maximum temperature permitted under TM is 30 ⁰C, so the hot spot is at 40 ⁰C at its center. The effects of daily thermal cycling were included as a stress factor. At the end of 200 weeks the center cell has undergone nearly 15% more capacity loss than cells located at the edges of the hot spot. Increased impedance at the hot spot (not shown) together with capacity loss will render a significant impact in power delivery for this string.Fig. 2 shows results for simulating power loss in an 11-cell series string comprised of LFP 26650-type cells operating in Phoenix. No battery TM is imposed, and a 10 ⁰C hot spot (symmetrical) is assumed. Two cases of operation are considered: (1) shallow power pulses for SOC near constant around 90% SOC (solid curves), and (2) deeper power pulses that cover 90 to 30% SOC (dashed curves). The stair-step profiles in the string power curves are due to the annual temperature cycle in Phoenix. While the hot spot accelerates the loss of string power, the case of deeper power pulses over a larger SOC range has a greater influence on the aging rate. Optimization of string life can be done through simultaneous consideration of SOC and TM parameters and total cells in the string, as a start.Lastly, Fig. 3 shows the outcome for string power simulations for the same LFP cell and string basis as for Fig. 2, but including TM and considering different geographic locations in the US. Here the upper temperature limit for TM is 30 ⁰C while the lower limit is set at 10 ⁰C. Under this TM condition the aging of the strings is more consistent, with some cities producing very similar trends such as Phoenix, Miami and Honolulu.To conclude, large-format batteries carry with them a need for TM that will be consistent within spatial, temporal and power regimes. The longevity and cell replacement costs of such batteries are directly related to TM effectiveness. If string aging is understood ahead of time, large-format battery design would benefit by having a proper balance between cells onboard and the cell replacement schedule. While simple cases are given in this abstract, more complicated scenarios will be presented for larger systems and more complex duty cycles.
Read full abstract