If you have selected the Life Model option (Electrical Parameters group box), you must specify the following life model parameters:

If you have select Using Profile/Data Table (Solution Options group box), you can use one of the following options in the Profile Types group box:

For the Time-Scheduled or Event-Scheduled options, you must provide time-dependent or event-dependent input (respectively) for the C-rate, current, voltage, or power. The time-scheduled and event-scheduled profiles enable you to change the electric load type and electric value in the simulation on the fly. When either the Time-Scheduled or Event-Scheduled profile type is selected, you can browse to the ASCII text file with either the .dat or .txt file extension that you generated beforehand. The inputs for the Time-Scheduled and Event-Scheduled profiles are described below.

Time-Scheduled Profile Specification

The Time-Scheduled profile file must be organized in three columns:

For example, the below time-scheduled profile:

0        15    1
500      15    1
500.1    60    3
1000     60    3
1000.1   -1    0
1500     -1    0

specifies the following battery charging scenario:

Ansys Fluent solver uses linear interpolation of your time-scheduled profile data in the problem simulation. If you want to capture a sudden change in the profile accurately, use smaller time gaps to describe the profile.

Event-Scheduled Profile Specification

The Event-Scheduled profile file must be organized in the following five columns:

For example, the below event-scheduled profile:

0    1    4  3.5    0
0    0    8  300    0
1  -15    5  4.1    0
2  4.1    3   -1    1

specifies the following scenario:

Any time-scheduled profile can be rewritten in an event-scheduled profile format, but not vice-versa. The example shown in the time-scheduled profile can be redefined in an event-scheduled profile format as follows.

1    15    1    500    0
3    60    1   1000    0
0    -1    1   1500    1

Both profile types can be read into Ansys Fluent. You can choose the most convenient one to use in your specific application.

Table Data Specification

The fast charging 2D lookup table defines the charging current as a function of battery's average DOD and temperature. To use this specification method follow the steps below:

2D and 3D lookup tables are used in the battery model to specify various quantities as tabular functions of other solution variables when the tabular input option is available. Lookup tables are tab- or space-delimited ASCII text files that support the following formats:

Usually if you transfer table data from one simulation to another, you can use the compact scheme format. If you prepare table data manually through a file, the plain text format is more readable and may be a better choice.

Figure 31.8: Model Parameters Tab—CHT Coupling Method

You need to specify the following CHT-Related Parameters:

For both heat source and tab electric current, you can use either constant data type or any of the imported transient profiles.

If constant is selected, you need to specify a value in the adjacent field.

To use transient profiles for energy sources, you first need to read all profile files into your simulation as outlined in Reading Profile Files. Once the profiles have been imported into your analysis, their names automatically appear in the drop-down lists for each active zone and for tab electric current. The profile file format conforms to the Ansys Fluent built-in transient profiles described in Transient Cell Zone and Boundary Conditions. The below example shows the file format of a tabular transient profile (.ttab).

energy-source  2  4  0
time       energy-source
0          5.0
100        5.0
100.1      10.0
10000      10.0

In this example, the first line contains the profile name, followed by the number of fields, the number of lines, and the flag indicating whether the profile is periodic.

The second line contains all the field names. One of the field names must be time for the profile to be correctly recognized.

The remaining rows contain the profile data.

All the imported profiles will be saved within the Ansys Fluent case file.

If in a pack simulation, the same heat source is used for all active zones, you can select Use Same Settings for All Zones and define only one set of data. Ansys Fluent will automatically update all zones with the same cell information.

Figure 31.10: Model Parameters Tab—NTGK/DCIR Model

For the NTGK/DCIR Model, define the following operating conditions:

Figure 31.16: Model Parameters Tab—Equivalent Circuit Model

For the Equivalent Circuit Model (ECM), define the following operating conditions:

 

31.2.3.4.4.1. The HPPC Library

Ansys Fluent provides a small HPPC library that you can use for importing HPPC test data into your case. Currently, the library contains only a single battery, called Generic-Cell. In future releases, the library will be expanded to include commonly used batteries. The purpose of the Generic-Cell library is to show the folder structure of test files as further described.

The Ansys Fluent HPPC library is located in the following directory:

path/fluent24.2.0/addons/msmdbatt/hppc_database/

where path is the Ansys Fluent installation directory (for example, C:\Program Files\ANSYS Inc\v242\fluent\fluent24.2.0).

In a standard HPPC test, a pulse of current is applied to a battery at a particular temperature and SOC level, and the voltage response of the battery is measured. The voltage response curve depends on:

• the nature of the pulse (discharging or charging)

• the magnitude of the current

• the temperature level

• the SOC level of the battery

To reflect such testing procedure, the test files are organized in the folder structure as shown in Figure 31.19: Overview of the Folder Structure in the HPPC Library.

Figure 31.19: Overview of the Folder Structure in the HPPC Library

The naming convention is as follows:

• The battery name can be any string. The capacity of the battery must be put in the parenthesis with a number followed by the unit "Ah".

• The names of the first-level folder must be discharge and charge.

• The number of the second- and third-level folders and their names can be arbitrary.

You can create your own library following the outlined conventions and put it in the hppc_database folder. The name of your library will appear in the HPPC Data Library dialog box. This way, the HPPC library can be expanded with the proprietary data and shared between colleagues within an organization.

---

Note:  Currently, the parameter estimation tool can only take data from the discharge branch and only handle data from one C-rate level. Data from different C-rate levels are treated as different battery libraries. You should select the library with the C-rate level close to that at which the battery is operating in your application. The charge branch is not used so far. However, support for the charge branch is expected in future releases as the solver capability will be expanded in Ansys Fluent.

---

Figure 31.20: Model Parameters Tab—Newman’s P2D Model

Theoretical Capacity is computed automatically from the Newman input parameters you specify and reported in the Model Parameters tab.

Specify the following settings in the Model Parameters for the Newman’s P2D model:

Figure 31.23: Model Parameters Tab—User-Defined E-Model

The following parameters are available under User-Defined E-Chem Model Parameters:

Prior to running a coupled electrochemical-thermal simulation, you should perform a quick analysis of your electrochemical model by simulating it in a standalone mode. The standalone model calculation uses the operating conditions that you have specified in the Model Options tab (such as electric load type and value) and the echem model parameters that you have defined in the Model Parameters tab. This preliminary analysis will allow you to confirm the electrochemical behavior of the selected submodel and, if necessary, adjust the model settings.

To run the echem model in a standalone mode:

In practical Newman’s P2D model simulations of battery aging through electric load cycles, the physics-based battery life model is solved on the electrode scale in the standalone mode. Further, the standalone simulation results can be included in a 3D CFD simulation to account for the battery aging effects as described in Inputs for the Newman’s P2D Model. For more information about the theory behind the physics-based battery life model, see Physics-Based Battery Life Model in the Fluent Theory Guide.

To run the physics-based battery life model in the standalone mode:

Aging Effects Profiles

The 1D spatial results of the physics-based battery life model are saved in aging effects profiles, which are Ansys Fluent built-in format profiles (see Profiles for more information about profile files). In the standalone mode, all the generated aging effects profiles are saved in the file called aging-field-profile.prof. The profiles are named aging-cycle-number, where number is the cycle number. Each profile contains the following field variables sampled at the cycle end:

Apart from postprocessing, the aging effects profiles can be used in the following two scenarios:

You can also create your customized aging effects profiles based on experimental data and import them into case. A customized profile should conform to the format of the Ansys Fluent built-in profiles described in Profiles and may contain different field variables based on different scenarios:

The customized profile can optionally contain cs_li, which is otherwise treated as zero.

Many industrial applications use a battery pack in which multiple battery modules are connected in either series or parallel. Within a pack, modules are usually laid out in some repeated pattern. In addition, a cold plate may be attached to the batteries. You can use the brute force method to build the whole pack geometry and then perform a simulation of the battery pack in Ansys Fluent. However, before running the pack simulation, separate simulations of individual battery module and a cold plate are often required. The Battery Pack Builder tool allows you to construct a battery pack directly within Ansys Fluent from existing Ansys Fluent simulations of the battery module and cold plate.

Figure 31.27: Battery Pack with a Cold Plate shows an example of a battery pack where the cold plate (a) and a single module (b) have been already simulated separately in Ansys Fluent. The Battery Pack Builder constructs the pack geometry (c) within Ansys Fluent.

Figure 31.27: Battery Pack with a Cold Plate

To use the Batter Pack Builder tool:

Note the following limitation that exist with the Battery Pack Builder tool:

Simulation of a battery thermal management system can be performed accurately using computational fluid dynamics (CFD). However, CFD models can be too large and too slow for a system-level transient thermal analysis. Reduced order models (ROMs), such as LTI-ROM and SVD-ROM, can be used in the system-level analysis with minimum loss of accuracy if certain conditions are satisfied.

In the LTI-ROM [70], the battery thermal problem is regarded as a system that processes a set of input signals (simply called inputs) and yields a set of output signals (simply called outputs). For a thermal problem, the inputs are the power (or the heat source), and the outputs are the temperature rise at user-specified locations or the average temperature rise across a zone. The benefit of the Linear and Time Invariant (LTI) system is that it is completely characterized by its impulse response, and the output of an arbitrary input function can be readily computed as the convolution of the input function and the system impulse response function.

A thermal system becomes an LTI system under constant flow rate, constant material properties, and linear boundary conditions. The temperature and heat flux boundary conditions are linear, and the radiation boundary condition (if present) is nonlinear. Constant density and constant properties are fairly accurate assumptions for water-cooling battery systems. For air-cooling battery systems, these assumptions can still be valid for the range of temperature variation under consideration.

For such an LTI thermal system, given an arbitrary time history of the battery heat source, the temperature evolution at a point can be obtained solely from the impulse response function at that point (that is, from the history of the temperature increase for a given unit amount of the heat source applied to the battery cell at “zero” time). The impulse response is obtained by performing a CFD calculation. In practice, instead of using an impulse function, a unit step function is applied, and the step response is monitored. The impulse response function is the time derivative of the step response function.

The LTI-ROM can calculate only cell-average temperature or temperature at a user-specified point. To obtain the temperature field, in theory, you can monitor the response function at each control volume and use the LTI-ROM for all the control volumes. However, the computational cost of such an approach would be too high, making it impractical. The dimension reduction method [12], [68] can be used for such cases to reduce the simulation time. In this method, Ansys Fluent first runs a CFD analysis to obtain the step response of the temperature distribution. Then, the singular value decomposition (SVD) is used to separate the spatial and temporal contributions of the temperature field. The dynamics of the temperature response is approximated in a largely reduced dimension space, which is formed by keeping only the first several leading singular values and their corresponding orthonormal left singular vectors. The LTI-ROM is then applied in such a reduced dimension space. Temperature at any physical space location can be obtained by a linear combination of the temperature points in the reduced subspace.

Similar to the LTI-ROM, the battery SVD-ROM tool kit in Ansys Fluent creates only the step response functions. The SVD-ROM files generated by Ansys Fluent can be used to run simulations in Ansys Twin Builder. Contact your technical engineer for detailed information and tutorials on using SVD files generated by Ansys Fluent in Ansys Twin Builder.

To use the battery ROM tool kit, click the Battery ROM Tool Kit… button in the Advanced Options tab of the Battery Model dialog box.

Figure 31.29: The Battery ROM Tool Kit Dialog Box

In the Battery ROM Tool Kit dialog box that opens, you can:

 

31.2.3.6.4.2. Computing Coefficient Matrices of the State Space Equations (LTI-ROM Generation Tab)

In a system simulation, an LTI-ROM model is usually solved by state space equations. Ansys Fluent provide a tool to calculate the coefficient matrices of the state space equations.

Figure 31.32: The Battery ROM Tool Kit Dialog Box - LTI-ROM Generation Tab

You can use the LTI-ROM Generation tab to perform the calculation as follows:

1. In the Response/Input File Location text entry box, specify the directory path where the response files are stored either manually or by using the Browse... button.

2. In the Control Parameters group box, specify the following parameters:

   State Space Model Min Order

   is the minimum order of the state space model for each element. A higher order yields more accurate results, but the computation time will also increase. Note that if zero is entered, then a contribution of cross heating from other volumes may still be ignored depending on the value you specify for Tolerance for 0th Order.

   State Space Model Max Order

   is the maximum order of the state space model for each element.

   Target Relative Error

   is the relative error that determines the order of the state space model fitted. In general, a smaller error creates a higher order state space model.

   Tolerance for 0-th Order

   allows you to ignore contributions from some inputs (for example, negligible contributions from certain cross-heating) if they are less than the specified tolerance multiplied by the largest contribution. This can significantly reduce the size of the overall model.

   NonZero StepRes Slop

   is an indicator of how the end slope of the response curves is handled. You can select from the following options:

   • No

   • Calculate

   • Provided

   For instance, in a case with all adiabatic boundary conditions, the step response will never reach a steady state condition, and the curve will have a slope. You can let the solver calculate it by selecting Calculate or provide a specific value by selecting Provided.

3. Click Generate LTI-ROM.

   The LTI-ROM tool will use all the necessary files (including Input_PortNames.txt, Input_StepInputValues.txt, and all the response files) from the folder specified in the Response/Input File Location field, run the fitting process, and generate the matrices of the state space equation. The resulting matrices will be saved in the Thermal_ROM_Matlab.m file in the same location. This file can be used in any system simulation software that solves a state space model. The GUI settings you specified will be automatically saved in a file called Input_ROMconfig.txt, which can be read back in a later session using the Load Config... button.

   Contact your technical engineer for detailed information and tutorials for the LTI-ROM generator tool.

A battery is made of a rolled or stacked sandwich structure of layered electrodes. Usually, the material properties in the in-plane direction of the electrode are different from those in the cross-plane direction. For example, the thermal conductivity in the in-plain direction is typically much larger than that in the cross-plane direction.

In battery applications, a battery module or pack may contain many batteries. In Ansys Fluent, you can define the orthotropic material for each battery in two different ways:

To use the Ansys Fluent built-in tool for defining orthotropic thermal conductivity for all battery active zones:

If you use the Battery Pack Builder tool, then material property settings specified in this way will be carried over to each battery module. That is, if a battery pack is constructed from a module that uses the orthotropic thermal conductivity, then the material properties specified in the Orthotropic Thermal Conductivity dialog box will be automatically applied to all batteries in all other modules. The cylindrical origin and axis of batteries in other modules will be updated automatically according to the module location and its orientation.

For the NTGK and the ECM model, the empirical-based battery model can be used in an FSI coupled simulation (see Empirical-Based Swelling Model in the Fluent Theory Guide).

To include the swelling effect, perform the following steps:

To use the Ansys Fluent battery venting model described in Battery Venting Model in the Fluent Theory Guide in your simulation, perform the following steps:

Once the solution is completed, in addition to the battery variables listed in Postprocessing the MSMD Battery Model, the following variable becomes available for postprocessing of simulations that involve the venting model (in the Battery Variables... category):

This variable can be used to display where the venting gas is added in the domain. The volume integral of Venting Gas Mass Source gives the total venting gas release rate in kg/s.

When any of the electrochemical submodels is used in the calculation, you can include entropic heat in the heat source calculation (Equation 19–8, Equation 19–13, and Equation 19–30 in the Fluent Theory Guide). To do so, follow the steps below.

To model the thermal abuse, select the Thermal Abuse Model option. The Battery Model dialog box expands to show additional model parameters (see Figure 31.25: The Battery Model Dialog Box (Advanced Options Tab)).

You can select the following options to model thermal runaway reactions:

Further details on these models are provided below.

For theoretical information about the thermal abuse model, see Thermal Abuse Model.

 

31.2.3.6.9.1. One-Equation Kinetics Model

Figure 31.38: The Battery Model Dialog Box - One-Equation Kinetics Model

For the One-Equation Kinetics Model, you can select the Abuse Reaction Rate Type from the following options:

• Arrhenius

• Table Data

• UDF

If you have selected Arrhenius, you need to specify the battery constants , , , , and used in Equation 19–33 through Equation 19–34, as well as the initial value for (Parameters for Kinetics Model group box).

If you have selected Table Data, you need to specify:

• Initial Reaction Progress Variable: Is the initial value of in Equation 19–33 and Equation 19–34.

• Specific Heat Release, HW: Is in Equation 19–34.

• Data for the abuse reaction rate table where the temperature increase rate is defined as a function of the temperature. You need to enter the following inputs in the Abuse Reaction Rate Table dialog box that opens by clicking the Abuse Reaction Rate Table... button:

  Figure 31.39: The Abuse Rate Data Table Dialog Box

  
  
  

  • Temperature Levels

  • Temperature

  • Rate dT/dt

  The controls in the Abuse Reaction Rate Table dialog box are similar to those in the Battery Aging Table dialog box described in Setting the Life Model.

If you have selected UDF, you need to specify Initial Reaction Progress Variable and Specific Heat Release, HW, and select the name of your compiled DEFINE_BATTERY_ABUSE_RATE UDF from the Abuse Reaction Rate UDF drop-down box. The UDF will be used to compute in Equation 19–33 and Equation 19–34. See DEFINE_BATTERY_ABUSE_RATE in the Fluent Customization Manual for details.

 

31.2.3.6.9.3. Standalone Thermal Abuse Model

Before running a CFD simulation, you can run the standalone thermal abuse model to compute the abuse reaction behavior and display the results in the graphics windows. This tool can be used to quickly assess the abuse model you are using in your simulation. To use this tool, click the Preview Thermal Abuse Model... button, and in the Standalone Thermal Abuse Model dialog box that opens, specify the following parameters in the Model Inputs group box:

Figure 31.41: The Standalone Thermal Abuse Model Dialog Box

• T0: Is the initial temperature

• rho*Cp: Is the volumetric heat capacity

• Volume: Is the battery volume

• Area: Is the battery external area

• T_conv: Is the environment temperature for convective heat transfer

• T_rad: Is the environment temperature for radiation heat transfer

• h: Is the heat transfer coefficient between battery and its surrounding environment

• epsilon: Is the battery surface emissivity

• Time Step: Specifies the time step size used in the calculation

• Total Time: Specifies the total calculation time

If you want to compare the temperature prediction with experimental data, enable the Compare with Experimental Data option and specify the name of the ARC experimental data file in the File Name text box. Alternatively, the file can be selected using the Select File dialog box, which is opened by clicking the Browse... button.

Once you entered the model inputs, you can click one of the plots (Temperature Profile, Heating Rate vs Time Profile, Heating Rate vs Temperature Profile, or Reaction Progress Variable Profile) in the Plots of group box to display the simulation results in the graphics window. Note that the standalone solver automatically performs computations in the background once you click one of the available plots. After reviewing the results, you can adjust the rate parameter if needed.

The NTGK model parameters are fitted as functions of DOD and temperature. To fit the NTGK model parameters, you need to have battery's discharging curves at different constant C-rates. You can use testing data at different temperatures.

To use the parameter estimation tool for the NTGK model:

If you use testing data from only one temperature level, the fitting results in both polynomial and table data types will be transferred to the Battery Model dialog box and to the parameter data table dialog boxes. In this case, the Polynomial data type will be automatically selected in the Battery Model dialog box, respectively. You can use either data type in your simulation.

The following is an example of using the parameter estimation tool for fitting the NTGK model parameters:

Parameter Estimation for Model:
	1: NTGK Model
	2: ECM Model
	3: Thermal Abuse Model
Model option:  [1] 1

-- Make sure every input file has this format --
        Crate     1.0
        Temperature   300.0
        time_1    voltage_1
        time_2    voltage_2
        ...       ...
------------------------------------------------

 file name for curve (wildcards "*" and "?" can be used within the file names) [] ntgk-*.dat
The following 15 new curves have been added:
ntgk-3C-290K.dat
ntgk-3C-300K.dat
ntgk-2C-310K.dat
ntgk-4C-300K.dat
ntgk-5C-310K.dat
ntgk-1C-310K.dat
ntgk-4C-310K.dat
ntgk-2C-290K.dat
ntgk-5C-290K.dat
ntgk-3C-310K.dat
ntgk-1C-290K.dat
ntgk-2C-300K.dat
ntgk-5C-300K.dat
ntgk-4C-290K.dat
ntgk-1C-300K.dat
 file name for curve (wildcards "*" and "?" can be used within the file names) [] ()
Battery capacity (Ah) [14.6] 
Number of DOD-levels [20] 
Include Capacity Fade Effect? [yes] no

For more information, see Inputs for the NTGK/DCIR Model.

To fit the ECM model parameters, you need to have battery's Hybrid Pulse Power Characterization (HPPC) testing data at different SOC levels. Similar to the NTGK model, you can use testing data at different temperatures.

To use the parameter estimation tool for the ECM model:

If you provide testing data files at different temperatures, the parameter estimation tool will perform the fitting for each temperature level. Ansys Fluent will report the fitted results in the console and automatically pass them in a table data type to the parameter data table dialog boxes. The Table data type will be automatically selected in the Battery Model dialog box (Model Parameters tab). This will allow you to consider the temperature-dependent effect.

The following is an example of using the parameter estimation tool for fitting the ECM model parameters.

Parameter Estimation for Model:
	1: NTGK Model
	2: ECM Model
	3: Thermal Abuse Model
Model option:  [2] 2

-- Make sure every input file has this format --
       SOC       0.6
       I         3.153
       Temperature   300.0 
       time_1    voltage_1
       time_2    voltage_2
       ...       ...
------------------------------------------------
  where SOC: soc level
          I: pulse current

 file name for curve (wildcards "*" and "?" can be used within the file names) [] soc??-???k.dat
The following 30 new curves have been added:
soc02-310k.dat
soc09-290k.dat
soc10-290k.dat
soc05-290k.dat
soc02-290k.dat
soc06-300k.dat
soc01-290k.dat
soc05-300k.dat
soc02-300k.dat
soc08-300k.dat
soc03-310k.dat
soc05-310k.dat
soc03-290k.dat
soc04-290k.dat
soc10-310k.dat
soc03-300k.dat
soc07-310k.dat
soc04-310k.dat
soc08-290k.dat
soc09-310k.dat
soc01-300k.dat
soc10-300k.dat
soc01-310k.dat
soc07-300k.dat
soc06-290k.dat
soc06-310k.dat
soc04-300k.dat
soc08-310k.dat
soc07-290k.dat
soc09-300k.dat
 file name for curve (wildcards "*" and "?" can be used within the file names) [] ()
Battery capacity (Ah) [14.6] 
Circuit Model:
	1: 4P parameter model (one RC loop)
	2: 6P parameter model (two RC loops)
 Circuit model option:  [2] 
Fitting Method:
	1: Jiang-Hu Method (JH)  
	2: Levenberg-Marquardt Method (LM) 
 Fitting method option:  [2] 
Include Capacity Fade Effect? [yes] no

For more information, see Inputs for the Equivalent Circuit Model.

Kinetics data used in the thermal abuse model are battery specific. Reaction kinetics parameters in Equation 19–33 can be fitted by using testing data (for example, Oven test or ARC). Temperature vs. time data collected from experiment for a battery cell under thermal abuse condition can be used as input in the parameter estimation tool.

The following thermal balance equation is used to describe the heating process and in the kinetics parameter fitting of the temperature-time data:

(31–1)

where:
	= volume
	= external area
	= material density
	= specific heat
	= heat transfer coefficient
	= Stefan-Boltzmann constant
	= surface emissivity
	= heat of reaction
	= pre-exponential factor
	= activation energy
	= reaction progress variable
	= temperature
	= ambient temperature
	= enclosure temperature for radiation
	and = reaction order parameters

The equation will be printed in the console for your reference.

You will be prompted to specify file names for temperature testing data and values for the variables used in Equation 31–1. For good fitting results, it is recommended that you use multiple sets of testing data obtained under different heating conditions, such as different oven temperature or different battery initial temperature.

After running the parameter estimation tool for the one-equation thermal abuse model, Ansys Fluent will obtain and pass the following parameter to the Battery Model dialog box (Advanced Options tab): A, E, HW, m, and n. These parameters will also be printed in the console.