This page describes the structure and usage of CAEBAT, building instructions, how to run an example, and how it is related to NiCE.
Computer-Aided Engineering for Electric Drive Vehicle Batteries (CAEBAT) is an open-architecture framework that integrates multi-physics and multi-scale battery models. After experimental validation, these simulations can be used for research, design, and manufacturing.
OASFramework
The OAS modeling framework integrates all the physics necessary for development of realistic and predictive battery performance and safety models. This framework uses a Python scripting language, and controls the battery state through component and state adapters.
The component adapters interact with the components by preparing the necessary inputs to run the components and by scheduling the component runs. The state adapters interact with the battery state file by updating all the necessary information about the battery state and the methods for coupling the components. The Battery state file is the core for passing data between components.
VIBE
This is the Virtual Integrated Battery Environment. The OAS framework, along with physics and support components and the adapters constitute the VIBE. The current components are electrochemistry, thermal, electrical, and cost model.
Other
The Data for CAEBAT is in a tar.gz file available upon request.
The Battery Markup Language (Battery ML) provides standardized format for the definition of all the necessary information.
The BatteryML Schema controls the structure of the output, and it uses a relational data model to stay flexible and avoid data duplication.
OASFramework
To create the OASFramework :
1. Create a directory
CAEBAT
and cd to that directory
2. Check out the source code from the repository using following command:
svn checkout http://svn.code.sf.net/p/ipsframework/code/trunk OASFramework
3. To build the library, use the following terminal commands:
cd OASFramework
mkdir build
cd build
ccmake ..
/PATH/TO/OASFramework/install
ccmake
make
make install
VIBE
1. Pre-install HDF5 and 32bit CGNS
For CGNS: in ccmake enable Fortran, HDF5 and tests. Also enable the tools if you wish to query the output.
If you want to perform optimization or parameter sweep, you need to also install DAKOTA (Design Analysis Kit for Optimization and Terascale Applications).
2. Untar the VIBE compressed file (
tar xzvf VIBE.tar.gz
) into the CAEBAT directory.
cd VIBE/trunk/components/config/
ln -sf makeconfig.gnu makeconfig.local
cd ../
(go back to
components
directory)
* make
/usr/bin/install
and remember not to end with a "/"
* make install
(this will install everything in the bin directory under components)
**DATA **
1. Untar the DATA directory (
tar xzvf CAEBAT-DATA.tar.gz
) into the
/CAEBAT directory
.
In order to run this, make a new directory in CAEBAT called RunExample. When going through these steps remember not to have "/" on the end, and that "Library" must be spelled correctly.
Create the following directories in RunExample:
Move the following files into RunExample:
Move the following files into RunExample/Data/dualfoil/
Configure the "workstation.conf"
Next, Configure the "thermal_chartran_cell_twoway.conf"
Next, do sourcing to use the cgns library, by using the following commands:
Now, to actually run the example use the following command while in /RunExample.
/CAEBAT/OASFramework/install/bin/ips.py --simulation=thermal_chartran_cell_twoway.conf --log=temp.log --platform=workstation.conf -a
In order to run it again, the /RunExample/Results/work directory must be emptied each time between runs.
We plan to integrate CAEBAT into NiCE as a plugin so we have a tool for simulating battery codes. NiCE will simply use its current tools to run CAEBAT the way that is mentioned above, behind the scenes.
This is a list of all the values that the user may like to change.
In workstation.conf:
HOST = allu1, which is an arbitrary user set description of the host.
MPIRUN = eval, which is the only allowed value. "mpirun" is a shell script used to launch a job with mpi.
**Resource detection method**
NODE_DETECTION = manual
Description: This determines whether the user chooses the finer options in the manual allocation description section, or simply chooses a preset node detection option.
Allowed Values: checkjob | qstat | pbs_env | slurm_env [Default Value = manual]
**Manual allocation description**
TOTAL_PROCS = 2, NODES = 1, PROCS_PER_NODE = 2
Description: This only applies when NODE_DETECTION = manual, and allows the user to choose how they want to process on their nodes.
Allowed Values: These are shown with their default values, and can have all integer values up to the platform's capabilities.
**Node topology description**
CORES_PER_NODE = 2 and SOCKETS_PER_NODE = 1
Description: A node is a group of sockets, which are groups of cores that are single processing units capable of performing computations.
Allowed Values: These are shown with their default values, and can have all integer values up to the platform's capabilities.
**Framework for setting for node allocation**
NODE_ALLOCATION_MODE
Description: This must adhere to the platform's capabilities, and describes the number of processes per node.
Allowed Values:
In thermal_chartran_cell_twoway.conf:
These will be shown with their default values, and allowed values are arbitrary strings unless otherwise stated.
RUN_ID = case1
Indentifier for this simulation run CoupledCellElectroChemThermalTransport.
SIM_NAME = ${RUN_ID}
Name of current simulation, the default value automatically finds the name.
PLASMA_STATE_WORK_DIR = $SIM_ROOT/work/battery_state
Where to put state files as the simulation evolves, allowed values will be any existing directory.
RUN_COMMENT = Coupled CAEBAT Module simulation
A run comment picked up by the portal.
TAG = ElectroChemThermal
A tag that enables related runs to be retrieved together.
USER = ul2
An optional parameter, if missing the unix username is used.
NUM_ZONES = 15 and NUM_CHARTRAN_ZONES = 15
These are the number of zones for ---?
SIMULATION_MODE = NORMAL
The Simulation mode setting, with the allowed values of ---?
LOG_LEVEL = WARNING
This changes what will appear in the log, with the allowed values of ERROR or WARNING.
LOG_FILE = $SIM_ROOT/${SIM_NAME}.log
This determines the name and location of the created log file, and the allowed values are any path to an existing directory followed by ".log".
The PORTS
NAMES = INIT DRIVER CHARTRAN THERMAL
The DRIVER is a required value in the port section.
IMPLEMENTATION = CHARTRAN_THERMAL_DRIVER
This tells how the simulation initialized, and will generate the first state if needed.
INIT is a required value in the port section.
For THERMAL,
IMPLEMENTATION = AMPERES.
For CHARTRAN,
IMPLEMENTATION = DUALFOIL.
These IMPLEMENTATIONs and DRIVERs are described below.
**Individual configuration sections**
Component specification, with entries similar for all components. NAME entry must match the name of the python class that implements the component. This section is where the user can fine tune the simulation to their liking. The user can determine which scripts to execute, the input and output files, and the classes to run for these main three types of simulations.
CHARTRAN_THERMAL_DRIVER
CLASS = DRIVERS
This is a class from the initialized list in the PORTS section.
SUB_CLASS = CHARTRAN_THERMAL
This is an implementation for the above CLASS?
NAME = Driver
The name of the python class implementing the component.
NPROC = 1
This is the number of processes that can run at once?
BIN_PATH = $CAEBAT_ROOT/bin
This provides the location of the bin directory.
INPUT_DIR = $DATA_ROOT/
This provides the location of the input files.
INPUT_FILES =
This provides each required input file by giving their names.
OUTPUT_FILES = $CURRENT_STATE
This writes the output files to a certain file...?
SCRIPT = $BIN_PATH/thermal_chartran_driver_n.py
This tells which script to execute.
AMPERES
CLASS = THERMAL
SUB_CLASS =
NAME = Amperes
NPROC = 1
ELECTRICAL = 0
BIN_PATH = $CAEBAT_ROOT/bin
INPUT_DIR = $DATA_ROOT/amperes_thermal
INPUT_FILES = 'Battery-zones1.e'
OUTPUT_FILES = $CURRENT_STATE
SCRIPT = $BIN_PATH/amperes_thermal.py
DUALFOIL
CLASS = CHARTRAN
SUB_CLASS =
NAME = DualFoil
NPROC = 1
BIN_PATH = $CAEBAT_ROOT/bin
INPUT_DIR = $DATA_ROOT/dualfoil
INPUT_FILES = 'dualfoil5.in' , 'li-ion-ebar.in'
OUTPUT_FILES = 'df_caebat.out'
SCRIPT = $BIN_PATH/dualfoil_chartran.py
**Time loop sepecification**
The two MODEs:
Below is shown the default values:
Overall structure of NiCE Plugin
These are a compiled effort based on the meeting of 3/21/2013
We will design CAEBAT into three different pieces:
1.) Model
2.) Launcher
3.) Analysis (tentative)
1.) Model:
There are multiple pieces to this particular module and will probably be the most complex. The CAEBAT program will require 3 types of input files in varying degrees of format and complexity. They are listed below:
A.) The workstation.config. This will probably be pre-generated under the hood by NiCE or another script and not be visible in the model builder.
B.) "Chartran".config file. This is a configuration setup file for the simulation. This will be pre-generated with NiCE's XML format for creating the model, and the input model will be generated in the listed format (a combination of key/value pairs with bracketing). Part of the k/v pair parsing is in NiCE, but will need to investigate K/V parsing for tables and MasterDetailsPairs.
C.) The "model" BatML.file. The schema will be provided by the CAEBAT Folks where we will investigate a way to generically handle the schema file in order to produce the GUI layout in the model, and then transform the values inputted into the schema enforced xml file.
2.) Launcher:
The basic launcher in development will suffice, but will investigate looking into adding Entries into the JobLauncher to handle multiple input files.
3.) Analysis ("Pretty Pictures"):
AUTOCAD or other visualization tools to show the results of the simulation.
The standing basis for the model will be CAEBAT CASE 4 model.