Step-by-Step Tutorial from Core to Basic
Preconditioning
Installing PyRolL Basic
Note
If you already have a working installation of pyroll-basic, you can safely skip this section.
The first requirement to follow this tutorial is to have a working Python installation with Python’s package
manager pip.
Then, install the required packages by running the following in a terminal resp. command prompt:
pip install pyroll-basic
This will install the PyRolL core together with a bunch of PyRolL plugins and extensions as well as their dependencies. Afterward, the command line interface (CLI) of PyRolL shall be available. Check this by running.
pyroll --help
You shall see a help message similar to the following:
Usage: pyroll [OPTIONS] COMMAND1 [ARGS]... [COMMAND2 [ARGS]...]...
Options:
-c, --config-file FILE The configuration TOML file.
-C, --global-config / -nC, --no-global-config
Whether use the global config file.
-p, --plugin TEXT
-d, --dir DIRECTORY Change the working directory to the
specified one. [default: .]
--help Show this message and exit.
Commands:
create-config Creates a standard config in FILE that can be used...
create-input-py Creates a sample input script in FILE that can be...
export-csv Exports the simulation results to JSON and writes them...
export-json Exports the simulation results to JSON and writes them...
export-toml Exports the simulation results to TOML and writes them...
export-yaml Exports the simulation results to YAML and writes them...
input-py Reads input data from the Python script FILE.
new Creates a new PyRoll simulation project in the...
report Generates a HTML report from the simulation results...
solve Runs the solution procedure on all loaded roll passes.
Initializing Your Workspace
It is recommended to create a new empty directory somewhere. Open a terminal resp. a command prompt in this directory to be able to run the input files of the following sections with the CLI. Download and save the input files of the following sections in this directory.
Using Just the Core
The following code block shows the contents of the file input_01.py.
It contains the absolute basic information needed to run a simulation in PyRolL using just the core package for one roll
pass.
The contents will be explained in the following in detail.
input_01.pyimport pyroll.core as pr
in_profile = pr.Profile.round(
diameter=8e-3,
flow_stress=100e6,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
The first line contains the import statement of the pyroll core package.
The alias pr is given to the package to shorten the access to the core classes of PyRolL.
import pyroll.core as pr
After that, the input profile is defined.
The input profile is an instance of the class Profile.
The name of the variable in_profile is mandatory for the CLI to recognize it as the input profile.
Here, a round shaped profile of 8 mm diameter is created using the Profile.round() factory method.
A constant flow stress of 100 MPa is given to the profile to enable basic roll force calculation.
in_profile = pr.Profile.round(
diameter=8e-3,
flow_stress=100e6,
)
Now, the pass sequence is defined, although it consists for now of only one roll pass.
The pass sequence is an instance of the class PassSequence.
Its only argument is usually the list of roll passes and transports it shall consist of.
For now, were are only adding one roll pass, nevertheless we need the pass sequence instance.
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
A roll pass is represented by an instance of the RollPass class.
For now, we need only two arguments to the roll pass, an object representing the working rolls (only one, since they are
equal) and the size of the roll gap.
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
A roll is represented by an instance of the Roll class.
The most important properties are here the nominal radius of the roll (radius without the groove indent) and the groove
geometry.
The rotational frequency of the roll is here used to determine the kinetics.
The groove geometry is defined by a groove object.
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
The PyRolL core includes a bunch of classes representing all common elongation groove types.
Here a circular oval (also known as 2-radii-oval) is used.
Here, the geometry is defined using the two radii r1 and r2 and the groove depth depth.
There are several other possibilities how the geometry can be constrained, see
the docs of the circular oval groove.
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
With this the first simulation can be run with the CLI, if the file input_01.py resides in the current working
directory of the shell:
pyroll input-py -f input_01.py solve report
The command loads the input script named input_01.py, runs the simulation procedure and creates a report page.
A new file named report.html should appear in the directory.
Open it in your favorite web browser and inspect the simulation results.
The results will generally be quite far from reality, since the core only includes very rough approximations of the physical relations. In the following we will go step-by-step to improve our simulation by adding additional model approaches for partial problems of the process model.
Adding a Spreading Model
See also
You can observe, that the outgoing profile of our roll pass takes the ideal shape of the groove, meaning that it has exactly the same width as the usable width of the groove (filling ratio is 1). This is the default behavior of the core, if no further models are provides, that may describe the material flow in the gap. This may also lead to the strange behavior, that the outgoing profile has a larger cross-section as the incoming one. Here, we will include Wusatowki’s spreading model, which will calculate the outgoing width of the profile by using an equivalent flass pass approach and an empirical spreading function.
We have nothing more to do as loading the plugin module with an import statement.
input_02.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
in_profile = pr.Profile.round(
diameter=8e-3,
flow_stress=100e6,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
If you run the simulation again, you will notice, that the profile does not fill the whole roll pass cross-section anymore. We are now approximately simulating the material flow in the pass.
Adding a Flow Curve
See also
The next improvement will be to include an empirical flow curve instead of the constant flow stress given before. This is done by importing the Freiberg flow stress plugin and giving some material state data. The flow stress functions needs the mean workpiece temperature and a set of material coefficients. Luckily, the plugin comes with some predefined material parameter sets. We can define the name of the material the workpiece is made of and the flow stress plugin will check its database for appropriate coefficients.
input_03.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
If you run the simulation again, you will notice that the flow stress of the incoming and outgoing profile changed and are now different due to the strain applied in the roll pass.
Adding Influence of Friction and Shear
See also
Currently, the simulation does not respect any influences of process kinetics, friction and shear aside from pure equivalent strain rate in the flow stress. The exact consideration of those is quite complicated and high in computational effort. But there are simplified analytical solutions to that as the one by Lippmann and Mahrenholz. We can include these considerations by importing the respective module.
input_04.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
import pyroll.lippmann_mahrenholz_force_torque
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
If you run the simulation again, you will notice that roll force and roll torque changed to higher values.
Adding More Precise Contact Areas
See also
Another huge influence on the calculation of roll forces and torques is the estimation of the contact area between rolls and workpiece. Since the geometry of this can be rather complex, often approximations are used. The core assumes just a trapezoidal or a triangle depending on the roll pass type. Zouhar developed a set of empirical equations to correct these rather rough approximations depending on the roll pass type. We can use these corrections by just loading the module, no further information must be given.
input_05.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
import pyroll.lippmann_mahrenholz_force_torque
import pyroll.zouhar_contact
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
])
If you run the simulation again, you will notice that the contact area changed and therefore also roll force and roll torque. The used correction coefficients are also listed in the report.
Adding Thermal Calculation
The last model we want to add for now is the calculation of temperature change due to deformation and heat flow through
the roll contact.
Therefore, we import the pyroll.integral_thermal module, which gives an integral heat flow and generation balance in
the roll pass.
We must provide the density and the specific heat capacity of the workpiece material to be able to
use this plugin.
Additionally, we may change the mean surface temperature of the roll as shown.
input_06.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
import pyroll.lippmann_mahrenholz_force_torque
import pyroll.zouhar_contact
import pyroll.integral_thermal
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
density=7.5e3,
specific_heat_capacity=465,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
temperature=293.1,
),
gap=1e-3,
),
])
If you run the simulation again, you will notice that the temperature of the profile changes now within the roll pass. The respective contributions to the temperature change are also listed in the report.
Adding a Second Roll Pass
Simulating one roll pass alone is not appropriate for developing a rolling technology. So let us add a second roll pass with a round groove of 7 mm diameter.
input_07.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
import pyroll.lippmann_mahrenholz_force_torque
import pyroll.zouhar_contact
import pyroll.integral_thermal
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
density=7.5e3,
specific_heat_capacity=465,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
pr.RollPass(
roll=pr.Roll(
groove=pr.RoundGroove(
r1=0.5e-3,
r2=3.5e-3,
depth=3e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
)
])
If you run the simulation again, you will notice that the second roll pass now appears in the report and the outgoing profile of the whole sequence is taken as the one of the second pass.
Adding a Transport
If you look at the input profile of the second pass, you will notice that it is assumed to have the strain and the temperature of the output profile of the first one. In reality, roll passes are not connected directly, but there is some space between the stands and/or some time lasts until the next pass is performed. In this time the material may recrystallize and the temperature goes down due to heat transfer by convection and radiation. This can be described by a transport unit, here assumed with a duration of 1 s.
input_08.pyimport pyroll.core as pr
import pyroll.wusatowski_spreading
import pyroll.freiberg_flow_stress
import pyroll.lippmann_mahrenholz_force_torque
import pyroll.zouhar_contact
import pyroll.integral_thermal
in_profile = pr.Profile.round(
diameter=8e-3,
material="S355J2",
temperature=1000 + 273.15,
density=7.5e3,
specific_heat_capacity=465,
)
sequence = pr.PassSequence([
pr.RollPass(
roll=pr.Roll(
groove=pr.CircularOvalGroove(
r1=2e-3,
r2=10e-3,
depth=2e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
),
pr.Transport(
duration=1,
),
pr.RollPass(
roll=pr.Roll(
groove=pr.RoundGroove(
r1=0.5e-3,
r2=3.5e-3,
depth=3e-3,
),
nominal_radius=100e-3,
rotational_frequency=1,
),
gap=1e-3,
)
])
If you run the simulation again, you will notice that the strain is reset to 0 in the transport and the workpiece will cool down a bit. The first appears since the core assumes complete softening in each transport. By using appropriate model plugins this behavior can be changed, but this is out of scope for this tutorial. Search in the plugins list for appropriate plugins.