the national renewable energy lab will be presenting a track on battery management at coolingzone 14 - click here to learn more

lifetime predictive models are needed to better understand design and operational consequences on performance degradation of li-ion batteries (lib). calendar-life models describing lib degradation have shown reasonable promise in predicting rate transport and reactions leading to lithium(li) loss and can be closely matched to data (ploehn, 2004; safari, 2011). detailed elementary chemical reaction models (christensen, 2004; colclasure, 2010) provide a framework for studying degradation reactions for different chemical systems. cycle-life models of libs, however, have yet to offer a method to predict capacity fade for a wide range of cycling and environmental conditions.  coulombic-throughput or energy-throughput are sometimes used as proxies to describe mechanical stress-induced fade and are regressed to experimental capacity data (peterson, 2010). these models are difficult to extend to a wide range of cycling conditions (wang, 2010).

mechanical stress has gained increasing attention in lib modeling literature with one motivator being to create physical models of capacity fade that can help guide cell design. mechanical stress effects have been modeled at length scales ranging from particle-level (christensen, 2006) to electrode sandwich level (renganathan, 2010; xiao, 2010), cell level (sahraei, 2012) and pack level (sahari, 2010). particle stress investigations have shown, for example, possible failure during fast-rate charging where intercalation-rate of li into a negative electrode active particle drives a faster rate of expansion at the outer radii of the particle, generating high-tensile stress in the inner core that may lead to fracture. during de-intercalation, high-tensile stress occurs at the outside of theparticle. models predict theoretical stress levels below which no failure should occur, albeit for simplified electrode geometries without pre-existing flaws. while these stress models already provide useful directional input to the design phase, further extensions are needed to predict capacity fade. particle fracture models must be extended beyond crack initiation to describe crack propagation on a time scale relevant for life prediction. computational models have not yet captured realistic geometries including dispersed flaws consequent of the manufacturing process. fracture leading to apparent active-site isolation has been studied; however, additional factors that may lead to active-site isolation—such as binder failure coupled with differential expansion of various cell components—have not been studied as they relate to cycling fade rate. cell and pack level models have mostly considered the impact of crush on various cell and pack geometries.

in this paper, we hypothesize a fatigue model to describe mechanical-stress-induced capacity fade. the model is regressed to capacity fade data from the literature for a commercial iron-phosphate cell. based on the model, the relative importance of various mechanical-induced degradation mechanisms on cycle life is discussed.   click here to read the full pdf article on the nrel web site, there is no cost for this article.

More >>