This paper presents a numerical model used for analyzing heat propagation as a safety feature in a custom-made battery pack. The pack uses a novel technology consisting of an internal short circuit device implanted in a cell to trigger thermal runaway. The goal of the study is to investigate the importance of wrapping cylindrical battery cells (18650 type) in a thermally and electrically insulating mica sleeve, to fix the cells in a thermally conductive aluminum heat sink. By modeling the full-scale pack using a 2D model and coupling the thermal model with an electrochemical model, good agreement with a 3D model and experimental data was found (less than 6%). The 2D modeling approach also reduces the computation time considerably (from 11 h to 25 min) compared to using a 3D model. The results showed that the air trapped between the cell and the boreholes of the heat sink provides a good insulation which reduces the temperature of the adjacent cells during thermal runaway. At the same time, a highly conductive matrix dissipates the heat throughout its thermal mass, reducing the temperature even further. It was found that for designing a safe battery pack which mitigates thermal runaway propagation, a combination of small insulating layers wrapped around the cells, and a conductive heat sink is beneficial.
This paper presents a novel model for analyzing thermal runaway in Li-ion battery cells with an internal short circuit device implanted in the cell. The model is constructed using Arrhenius formulations for representing the self-heating chemical reactions and the State of Charge. The model accounts for a local short-circuit, which is triggered by the device embedded in the cell windings (jelly roll). The short circuit is modeled by calculating the total available electrical energy and adding an efficiency factor for the conversion of electric energy into thermal energy. The efficiency factor also accounts for the energy vented from the cell. The results show good agreement with the experimental data for two cases – a 0D model and a 3D model of a single cell. Introducing the efficiency factor and simplifying the short-circuit modeling by using an Arrhenius formulation reduces the calculation time and the computational complexity, while providing relevant results about the temperature dynamics. It was found that for an 18650 NCA/graphite cell with a 2.4 Ah capacity, 28% of the electrical energy leaves with the effluent. Lithium-ion batteries are gaining more and more popularity in the field of electric energy storage. 1 This trend is followed by an increase in safety, energy density, and cycle life requirements. The in-crease in energy density brought a significant contribution to this trend, but it came with a trade-off concerning safety. 2,3 When operated un-der abusive conditions such as overcharging, over-discharging, object penetrations or even operation under high ambient temperatures, etc., Li-ion batteries can undergo internal short circuits between the current collectors or electrodes, leading to thermal runaway. 3 The reactions with electrolyte inside the cell decompose the battery components, generating a significant amount of heat, which, if not properly man-aged can lead to fires and explosions. 4 To assist the design of thermal management systems in mitigating the effects of thermal runaway, it is important to be able to model thermal runaway and account for the energy contributions in the process. Modeling thermal runaway has been the focus of many researchers, but the authors in Refs. 5,6 brought a substantial contribution to the field. The authors found the activation energies and the enthalpies of the different decomposition reactions for the components in an 18650 LCO (2.6 Ah) Li-ion battery and proposed a model for predicting thermal runaway based on Arrhenius formulations. Papers such as Refs. 7,8 added new decomposition reactions (cathode, electrolyte) and extended the model, from a simplified lumped model to complex 2D and 3D geometries for a single cell. Based on these models, some authors extended the models to simulate the thermal behavior of single battery cells. 9 A comprehensive list of references and studies of modeling safety in Li-ion is given in Ref. 3. The activation energies and the enthalpies found by the authors in Refs. 5, 6 are crucial for predicting the energy released during thermal runaway and are used in this paper. During a thermal runaway, an internal short circuit (ISC) can occur in the cell due to a conducting metal particle, component defects or melting of the separator, causing the adjacent electrodes (the anode and the cathode) to come into contact. Analyzing the ISC is a chal-lenging task being tackled by more and more authors. Some focused on experimental studies on different types of cells, 10–13 but only a few studies have been performed in modeling the ISC.
Rotating ring disc electrode (RRDE) experiments are a classic tool for investigating kinetics of electrochemical reactions. Several standardized methods exist for extracting transport parameters and reaction rate constants using RRDE measurements. In this work, we compare some approximate solutions to the convective diffusion used popularly in the literature to a rigorous numerical solution of the Nernst–Planck equations coupled to the three dimensional flow problem. In light of these computational advancements, we explore design aspects of the RRDE that will help improve sensitivity of our parameter estimation procedure to experimental data. We use the oxygen reduction in acidic media involving three charge transfer reactions and a chemical reaction as an example, and identify ways to isolate reaction currents for the individual processes in order to accurately estimate the exchange current densities.
A nonlinear state-variable method is presented and used to solve the pseudo-2D (P2D) Li-ion cell model under high-frequency input current and temperature signals. The physics-based governing equations are formulated into a nonlinear state variable method (NSVM), in which the mass transfer variables are evaluated using a 1st order exponential integrator approach at each discrete time point and the electrochemical kinetics (Butler-Volmer) equations are solved by either an iterative or an explicit method. This procedure provides an accurate, computationally efficient method to develop physics-based simulations of the performance of a dual-foil Li-ion cell during practical drive cycles. (C) The Author(s) 2017. Published by ECS. All rights reserved.
This paper presents a mathematical model built for analyzing the intricate thermal behavior of a 18650 LCO (Lithium Cobalt Oxide) battery cell during thermal runaway when venting of the electrolyte and contents of the jelly roll (ejecta) is considered. The model consists of different ODEs (Ordinary Differential Equations) describing reaction rates and electrochemical reactions, as well as the isentropic flow equations for describing electrolyte venting. The results are validated against experimental findings from Golubkov et al. [1] [Andrey W. Golubkov, David Fuchs, Julian Wagner, Helmar Wiltsche, Christoph Stangl, Gisela Fauler, Gernot Voitice Alexander Thaler and Viktor Hacker, RSC Advances, 4:3633-3642, 2014] for two cases - with flow and without flow. The results show that if the isentropic flow equations are not included in the model, the thermal runaway is triggered prematurely at the point where venting should occur. This shows that the heat dissipation due to ejection of electrolyte and jelly roll contents has a significant contribution. When the flow equations are included, the model shows good agreement with the experiment and therefore proving the importance of including venting.
A particle filter (PF) is shown to be more accurate than non-linear least squares (NLLS) and an unscented Kalman filter (UKF) for predicting the remaining useful life (RUL) and time until end of discharge voltage (EODV) of a Lithium-ion battery. The three algorithms, i.e. PF, UKF, and NLLS track four states with correct initial estimates of the states and 5% variation on the initial state estimates. The four states are data-driven, equivalent circuit properties or Lithium concentrations and electroactive surface areas depending on the model. The more accurate prediction performance of PF over NLLS and UKF is reported for three Lithium-ion battery models: a data-driven empirical model, an equivalent circuit model, and a physics-based single particle model.
