18650电池的变形和失效外文文献英文

Deformation and failure mechanisms of 18650 battery cells under axial compression Juner Zhu a Xiaowei Zhanga Elham Sahraeia b Tomasz Wierzbickia aImpact and Crashworthiness Lab Massachusetts Institute of Technology 77 Massachusetts Ave Room 5 218B Cambridge MA 02139 USA bElectric Vehicle Safety Lab The George Mason University Nguyen Engineering Building 4400 University Dr Fairfax VA 22030 USA h i g h l i g h t s Short circuit happens when an 18650 battery cell is axially compressed to 4 mm Deformation is localized in the positive terminal endcap region Axial compression of jellyroll creates a circumferential crack in the separator in top region Deformation of shell casing and jellyroll can be simulated by FE model and explained analytically a r t i c l ei n f o Article history Received 12 July 2016 Received in revised form 28 September 2016 Accepted 19 October 2016 Available online 29 October 2016 Keywords Lithium ion battery Axial compression FE simulation Analytical model Micro CT scan a b s t r a c t An important deformation mode during ground impacts of battery packs made of cylindrical battery cells is axial compression This type of loading subjects the cell to a complex deformation pattern and failure mechanism The design of endcaps plays an important role in such deformations To explore the sequence of deformation and the underlying failure mechanisms a combined experimental numerical study was carried out Tests were conducted on 18650 cells and the deformation of each component was carefully investigated and documented There are four different stages in the force displacement curve corresponding with deformation of various components in the endcap assembly A short circuit happens at a displacement of 4 mm To clarify these observations a detailed Finite Element model was set up covering the geometry and the mechanical property of almost all the components of the cell Using the simulation results the sequence of the axial compression was revealed which was subsequently vali dated by Micro CT scans as well as analytical solutions Based on the precise analysis of the mechanical behavior the cause of the short circuit during axial loading was clarifi ed Two failure mechanisms in the separator at the top section of the cell explain the possible causes of short circuit 2016 Elsevier B V All rights reserved 1 Introduction Cylindrical lithium ion battery cells have been a major power source for Electric Vehicles like Tesla Model S The vertical confi g uration of these cells in the fl oor mountedbattery packs make them prone to axial deformation in case of a ground impact Most of the work on mechanical loading of batteries have been focused on transverse direction for both pouch and cylindrical cells 1e7 Researchers at MIT Impact and Crashworthiness Lab ICL have designed experimental methods to characterize through thickness mechanical propertied of pouch cells under local indentation They havedevelopedcomputationalmodelsthatcloselypredict response of the cell in such loadings 2 3 Similar experimental and computational methods were applied to cylindrical cells under lateral loading by ICL team and this line of research was followed byVolkswagenresearchers 1 4 Results havealso been reportedby ICL as well as University of Michigan on axial loading of pouch cells in which local short amplitude buckling was observed 8e10 An initial investigation was performed on the Tesla S ground impact accident and it was found that the road debris can produce severe indentation in the protective aluminum armor plate 11 The computational model predicted a considerable shortening of the cylindrical cells more than 10 mm before a sheet separating the battery pack from passenger compartment is punctured The cells Corresponding author Impact and Crashworthiness Lab Massachusetts Insti tute of Technology 77 Massachusetts Ave Room 5 218B Cambridge MA 02139 USA E mail addresses zhujuner mit edu J Zhu zhangxv mit edu X Zhang elhams mit edu E Sahraei wierz mit edu T Wierzbicki Contents lists available at ScienceDirect Journal of Power Sources journal homepage http dx doi org 10 1016 j jpowsour 2016 10 064 0378 7753 2016 Elsevier B V All rights reserved Journal of Power Sources 336 2016 332e340 in that study were previously represented by using homogenized material for all interior of the jellyroll Although the assumption of homogenized material properties works well for transverse loading of cells it loses much of its ac curacy for axial loading due to special assembly of anode and cathode layers having a mismatch on top and bottom parts of cy lindrical cells Additional complexity in axial loading is the exis tence of cell endcap structure at positive terminal A deformation in axial direction may cause short circuit in the cap assembly well before any failure happens in the interior of jellyroll In axial loading a large deformation in jellyroll will lead to local buckling patterns Such patterns are also detected on compression of lami nated and sandwich structures as studied by Karam and Gibson 12 The objective of this paper is to investigate the modes of failure for various components of 18650 cylindrical cells under increasing axial loading Deformation and failure in lateral loading cases are more straight forward as it only involves tension and compression of various layers in a continuous uniform way Axial loading how ever involves several phenomena in a step by step sequence of deformation for various components arranged in series Axial deformation bending and buckling defi ne different stages of shell casing deformation Also the jellyroll deformation is non uniform and localized Mechanisms of short circuit could involve failure in the plastic gaskets causing contact between shell casing and pos itive terminal or failure in separator leading to contact between electrodes or positive electrode and shell casing To identify how the deformation progresses a detailed fi nite element model of an 18650 cell was developed Special emphasis was put on high fi delity representation of endcaps involving gas kets safety vent axisymmetric groove in shell casing and several other small components The electrodes and separators were modeled as individual components and a great deal of attention was put on representing the mismatch and interaction between the layers in the top section of jellyroll right under the positive ter minal To represent all the above details around one million ele ments were used in the model The numerical simulation provided a clear picture of the progression of the complex deformation pattern on the upper part of the cell thus providing a much needed insight about possible causes of short circuit The predictions were verifi ed by CT Scan of cells tested in identical loading confi gura tions It is believed that present results provide great insight for designing next generation of safer cylindrical cells 2 18650 battery cell 2 1 Composition of the cell An standard 18650 battery cell and its three major components are shownin Fig 1a Components include a safety valve at the top of the cell Fig 1b jellyroll Fig 1c and shell casing made of mild steel Fig 1d Safety valves made by different manufactories differ in structure but commonly several important sub components are included such as positive temperature coeffi cient PTC device aluminum safety vents steel positive terminal and gasket seal see Fig 1e The jellyroll which includes anode cathode and two layers of separator is the most essential part of the battery cell Fig 1feh From the structural point of view the anode is comprised of a copper current collector with active positive electrode material graphite coated on its surfaces and the cathode is an aluminum fi lm current collector with Lithium metal oxide coating here LiCoO2 The separator is a polymeric porous membrane placed between anode and cathode permeable to ionic fl ow but pre venting electric contact of the electrodes The separator of the studied 18650 is made of polyethylene PE 13 The basic dimensions of the geometry of the battery cell and its components are carefully measured and listed in Table 1 The readers are referred to the authors previous publications for more details 1 3 11 14 15 2 2 Material properties of the components The present research team have devoted great efforts to obtain the mechanical behavior of the individual components of the bat tery cell covering elasticity plasticity and fracture The properties of the essential components are shown as Fig 2 including the hardening curves fl ow stress plastic strain of the aluminum cur rent collector 16 and the copper current collector 3 the engi neering stress strain curves of the separator the tensile true stress strain curves of the coating material for anode and cathode 5 and the hardening curve and ductile fracture locus failure strain stress triaxiality of the steel of shell casing 15 More details about the materialpropertiesarereportedintheabovereferenced publications 3 Axial compression test Axial compression tests were carried out on 18650 cells using a universal testing machine 200 kN MTS under a loading speed of 5 mm min quasi static range The setup of the test is shown in Fig 3a and b Digital Image Correlation DIC method was used for recording the deformation of the cell and the displacement of the loading end 17 A voltmeter connected to the data recording system of the computer was employed to monitor the voltage of the cell The synchronization among the measurements of force displacement and voltage was realized by computer All the battery cells were fully discharged before test SOC 0 Fig 3c presents the results of four compression tests In three tests both the resistance force black curves and voltage blue curves were measured In the fourth one yellow curve for resis tance force the exterior accessories of the cell top cover and conductors attached on the positive negative terminals had been removed Results show satisfactory repeatability of the complete cells However it should be pointed out that the battery cell is a Nomenclature s stress tensor smmean normal stress svon Mises equivalent stress syyield stress suultimate stress equivalent strain hstress triaxility Fresistance force udisplacement puniformly distributed pressure Rout Rinouter and inner radius of shell casing rradius of battery cell Mfully plastic bending moment per unit length hthickness of shell casing EYoung s modulus Eeffeffective Young s modulus of jelly roll Acontactarea of contact Llength of battery cell components DLdeformation of battery cell components J Zhu et al Journal of Power Sources 336 2016 332e340333 complex assembly so that there are initial gaps between different components For this reason there is an obvious displacement shift of 0 9 mm between the last test and the other three Note that by removing the exterior accessories only a small part of the initial gaps was eliminated and the interior gaps still exists Therefore the total amount of initial gaps should be larger than 0 9 mm Despite the displacement shift all the resistance force curves share the same interesting feature e they follow a trend of slow increase e rapid increase e slight decrease e rapid increase In other words four different stages clearly exist in the compression process which implies that there must be different underlying mechanisms behind each stage of the process Another important observation of the test is the voltage drop Although there is a small deviation in the voltage drop displace ment among the three complete cell test it is still safe to draw the conclusion that 18650 cells come to failure after a shortening of 4 mm approximately 3 mm subtracting the initial gaps Moreover this voltage drop turned out to be caused by a short circuit rather than a broken circuit because a large amount of heat was produced and the tested cells were warmed during the tests The main task of the present paper is to investigate the mech anism of the deformation of cells under axial compression and investigate the cause of the short circuit To accomplish this task various approaches are utilized The Finite Element method is described in Section 4 followed by the analytical modeling in Section 5 as well as the experimental method in Section 6 Besides Micro CTscans are carried out to further validate the conclusions of the present paper 4 Finite element simulation of the axial compression process 4 1 FE model A FE model representing almost all the details of the battery cell was set up in Abaqus explicit using the geometric parameters in Table 1 and the material properties in Fig 2 For the sake of computational robustness the simple elasto plastic material model is used for most of the metal materials The steel of shell casing and the aluminum foil of the cathode current collector both show a certain degree of anisotropy less than 10 15 16 which is characterized by the Hill 48 model For the separator material a constitutive model combining anisotropic elasticity and Hill 48 plasticity is employed For all of the involved materials a fracture locus which describes the fracture strain as a function of stress triaxiality is defi ned The stress triaxialityhis a frequently used parameter to characterize the stress state and is defi ned as h sm s 1 3 tr s ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffi ffiffi 3 2 s s r 1 wheresmis the mean nornal stress sis the von Mises equivalent stress and s denotes the stress tensor Theapproximatetriaxialityforuniaxialtension uniaxial compression and pure shear are 1 3 1 3 and 0 respectively It is Fig 1 Composition of the 18650 battery cell a 18650 battery cells b safety valve c jellyroll d shell casing e sub components of the safety valve structure f anode g cathode h separator Table 1 Geometric parameters of the 18650 battery cell ParameterValueParameterValue Height of the cell mm64 7Diameter of the cell mm18 2 Height of top end mm3 5Thickness of shell casing mm0 17e0 25a Thickness of coating on cathode mm70 Width of copper fi lm mm58 0 Thickness of coating on anode mm70 Thickness of copper fi lm mm10 Width of separator mm58 5 Width of aluminum fi lm mm56 0 Thickness of aluminum fi lm mm20 Thickness of aluminum fi lm mm20 a Not uniform The top part is thicker than the bottom The two parts are welded together J Zhu et al Journal of Power Sources 336 2016 332e340334 well known that the fracture strain of materials especially for metals is highly dependent on the triaxialilty 18 For example most metals can undergo much larger strain under uniaxial compression than under uniaxial tension A fracture locus model shown in Fig 2b can conveniently describe the fracture initiation of typical metals The detailed FE model is shown in Fig 4a The overall structure is axis symmetric which implies that a 2D model would be suffi cient However the applicability and robustness of the 2D simula tion in Abaqus is not as good as the 3D case 19 especially for such a model with great amount of contact surface pairs and various sizes of components For this reason a quarter model using 3D shell and solid elements were preferred 3D shell element were used for thin aluminum and copper fi lms Two different sets of mesh size was employed and compared One is fi ne mesh with 917 415 ele ments as illustrated in Fig 4a and the other is a coarse mesh with 405 486 elements Simulation results show that the force responses with these two mesh sizes are identical but the fi ne mesh can provide much smoother results for local deformation To speed up the simulation a loading rate of 1 m s is adopted after a careful study on the dynamic effect Under such a loading rate the kinetic energy of the entire structure is negligible F 2pp R2 out R 2 in Force equilibrium 2M u Z Rout Rin 2ppr r Rin drMoment Equilibrium 2 where p is the uniformly distributed pressure u is the displace ment M is the fully plastic bending moment per unit length and is function of u considering a linear plastic hardening model M u My M u My u u max 3 and My syh2 4 0 0016 kN Mu suh2 4 0 0043 kN and umax Fig 3 Setup and results for the axial compression tests a overall setup b compression of battery cell c results of the tests black curves force blue curves voltage yellow dot curve force of Test 4 initial gaps were partially eliminated and the four stages of the process For interpretation of the references to colour in this fi gure legend the reader is referred to the web version of this article J Zhu et al Journal of Power Sources 336 2016 332e340336 Fig 4 a Finite Element model of a quarter of the 18650 battery cell mesh grids are not plotted with illustrations of fi ne mesh sizes and unit of the jelly roll Simulation results of force displacement response b un shifted curve and c shift the displacement of experimental curves to compare with simulation Fig 5 a sequence of the deformation process with A E corresponding to Fig 4c b validation of simulation results by comparing cross section with Micro CT scan c validation of simulation results by comparing deformed jelly roll with test pictures in b and c are captured at the moment of short circuit J Zhu et al Journal of Power Sources 336 2016 332e340337 is the maximum shortening at the end of Phase I Solving the above equations for F a linear function is obtained given by the segment AB in Fig 7 5 2 Stage II III IV indentation of jelly roll roll 4 where Eeffis the effective Young s modulus of the layered structure of the jelly roll 10 64 GPa Eeff P i hiEi P ihi i layers in the jelly roll 5 and Acontactis the area of contact 7 79 mm2 measured from simulation The resistance force of the latter part is more complicated since the axial compression process consists of elastic range shell buckling and post buckling In the elastic range the resistance force can be calculated as F2 EA E 2phRout DL L0 casing 6 The fi rst peak in the experimental and simulated load displacement curve is due to the process of progressiv