Post on 28-Feb-2020
transcript
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
1
Centrifuge Model
The centrifuge model demonstrates how to perform a rotational spinning centrifugal load analysis. It
also demonstrates how to group different parts with/without the intended initial bonding for contact
analysis, and to demonstrate the tolerance in solid model with initial touching contact.
Using only the quarter model with proper symmetry boundary condition, the result is the same as
analyzing the whole model, but can be solved faster.
This model applies the automatic contact condition so the upper and the lower portion of the centrifuge
will interlock together. In many legacy FEA tools, when contact analysis is involved, it requires the
loading/constraint be applied in steps of very small increment. This model demonstrates the robustness
of the automatic contact ability, and will automatically apply non-penetrating contact condition without
any user intervention.
Suggested what-if analysis: change the contact friction; use half model rather than quarter model;
change rotation speed.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
2
Engine Block Model
The small 2-stroke hobby engine model demonstrates how to do a thermal-stress coupled analysis, and
show how to set the proper thermal boundary condition to predict the thermal induced stress
deformation.
The combustion heat generation effect is modeled using an equivalent heat flux in the interior of the
cylinder wall area. It is analyzed by applying a known seating temperature at the rear end of the engine
block. The only stress boundary condition is the fixed bolting constraint. It automatically computes both
the temperature distribution and the induced deformation, mostly around the engine block.
In many legacy FEA tools, when thermal and stress coupled analysis is desired, usually a thermal heat
transfer analysis has to be done first, and then overlaying the temperature result on the stress model to
get the thermal deformation and stress. Not only such segregated method is not accurate, it may never
converge or give accurate result if material properties have temperature dependency. IronCad
Multiphysics simulates these types of coupled problem directly by solving the heat transfer and stress
equilibrium together, and realistically simulate the physics as it happens.
Suggested what-if analysis: change the inner combustion heat flux intensity; apply surface convective
heat loss on the engine block fins to see how it can decrease the steady-state temperature.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
3
Tuning Fork Model
The tuning fork model demonstrates how to perform modal analysis (harmonic vibration mode),
buckling analysis, and frequency domain analysis. These three analysis functions are the most
commonly used in structural dynamic design and in instability checking.
Modal analysis extracts the
fundamental vibration modes of the
model. These are the frequency a part
should avoid during operations as it
will create resonance responses.
Frequency domain analysis can analyze
a load or constraint at the given range
of frequencies to determine the
dynamic amplification load effect, and
as demonstrate, if you excite a part
over a range of frequencies containing
the natural vibration modes, it will
produce higher dynamic amplification
response and could possibly damage
the part. To see the effect, in the
frequency domain Result section,
select the Node inquiry, select a node
around the loading source, and plot
the displacement magnitude versus
frequency to see the peak behavior.
Buckling analysis is a commonly used
design instability analysis. It requires
that you place a loading or constraint
on the model, then compute the bucking factor(s) based on the loading status. The buckling factor
(shown as “LoadFactor” in Result) is the factor scale of the applied loading/constraint. It informs the
user that under the indicated factor of load/constraint, the system can go into the mode of instability
buckling, and such factored loading state should be avoided or suppressed by adding more constrain to
prevent such buckling instability.
Suggested what-if analysis: change the stiffness or mass to see the sensitive of the modal frequency;
apply lateral constrain to the buckling analysis to see how it changes the buckling factor; use finer
frequency increment in the frequency domain analysis to see how the displacement changes around the
computed vibration mode.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
4
Rubber Squeezing Model
The rubber squeezing model shows how to use the incompressible hyperelastic material model. The
model simulates an experiment of squeezing a hollow rubber tubing against an incompressible rubber
block of the same material using a rigid steel bar.
To use the hyperelastic material, the user can enter the experimental stress-strain curve, and the
hyperelastic material model input tool will automatically compute the needed the general 9-coefficient
Mooney-Rivlin model coefficients. In this case, we have entered the coefficient from an experiment
based on a Neo-Hookean single coefficient model.
Since the final step of squeezing will compress element shape a lot and increase the nonlinearity, we
refine the load step to half the initial increment at the half way into squeezing. We also show how
different mesh refinement can be used to increase the simulation accuracy. In many situations, if you
refine the mesh, it also requires to reduce the time/load step size.
Suggested what-if analysis: change the mesh size to see the sensitivity of the result; Increase/decrease
the step size to see how program succeeds and fails after automatic step size cut down;
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
5
Taylor Bar Impact Model
The Taylor bar impact model demonstrate how to perform a high speed dynamic impact analysis using a rigid contact plane with the automatic contact boundary condition. It is a seem-to-be simple but advanced nonlinear large deformation plasticity benchmark. The bar has an initial length of 32.4mm with an initial radius of 3.2 mm, Young’s modulus=1.17e11 Pa, Poisson's ratio=0.35, initial yield strength=4.0e8 Pa, hardening modulus=1.0e8 Pa, and mass density of 8.93e3 Kg/m3. The bar has an initial impact speed of 227m/sec to a rigid frictionless wall. Initial time step of 5.0e-7 seconds is taken for 200 steps.
Suggested what-if analysis: change the mesh size to see the sensitivity of the result; Increase/decrease
the step size to see how program succeeds and fails after automatic step size cut down; Change the
default Sefea Formulation and use the standard formulation; use the 2nd order TET10 element and check
the much longer runtime and the increased iterations/instability.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
6
Pool Game Model
The pool game simulation is based on an actual children’s pool table so the size is smaller than the actual one to speed up the analysis. It demonstrates the automatic contact feature in IronCad Multiphysics simulation, and it also demonstrates the gravitational loading effect, multi-material setup, grouping, and the use of foam material for the pool table sides. The model is setup to run as realistic as possible, but with the least amount of size. In the dynamic analysis, the traditional Courant step size criterion requires that the time step size must be small enough so the stress wave propagation (or dynamic impact wave) should not be traveling beyond the smallest mesh size, and the element traveling speed should also follow the similar rule. The stress wave speed is roughly equal to the squared root of the stiffness/mass ratio, and the theoretical allowable time step size will be very small. Based on advanced algorithm and formulations, the demonstrated step size of 0.005 is much larger than the theoretical limit for practical design analysis purpose. In order to achieve more realistic behavior, the pool balls must have some frictional effect for realistic responses, and Coulomb friction coefficient of 0.02 is used. The pool table and the foam sides also have been modeled with a 1% damping to simulate the damping behavior happens in real game due to the pool table/side cloth surface. To reduce the result output file, the model only outputs the time step result at every other step (in Analysis/Advanced Optional controls).
Suggested what-if analysis: change the cue ball initial direction and see the response (till a ball is
dropped and fall to the ground); increase the time step size and see how the program automatically cut
down the time step to the allowable minimum step size; increase the pool ball mesh size so you can
increase the time step
to see how fast a one-
second pool game
motion can be achieved;
change the damping
ratio of the table and
the side to see how it
impact the behavior.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
7
Ball in Channel Flow Model (MPIC 2016) The moving cylinder in a confined channel flow problem is a popular benchmark for mass conservation in Computational Fluid Dynamics (CFD). The benchmark focuses on comparing the total flow rate passing through each cross section. The mass conservation error at the narrowest flow section is computed by integrating the total flow passing the cross section plane. Automatic meshing with an average mesh size of 0.1 is used for all types of meshes with finer mesh of 0.04 across the cross section. Using only the half model with proper symmetry boundary condition, the result is the same as analyzing
the whole model, but can be solved faster. Actual water property is used for this benchmark. Mass
conservation error is without the specified error tolerance of 5% (about 0.05 of the theoretical
2.25/2=1.125 m^3/sec). The user is encouraged to check the velocity integration over the cutting plan
for the consistent flow rate across different cross section.
The drag force on the ball can also be computed by summing the “Nodal force in X direction” on the ball
surface nodes by selecting the ball surface.
This model demonstrates the
general CFD application of the
MPIC advanced LSFEA
incompressible fluid formulation
with Sefea technology.
Suggested what-if analysis:
change the ball speed; use quarter
model rather than half model;
change flow BC to advance the ball
rather than moving the channel;
change mesh density to check
sensitivity
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
8
Natural Convection in Cavity (MPIC 2016)
The natural convection is critical CFD applications as it apply to applications in CVD (chemical vapor deposition), heat tube thermal exchanger and many other designs. The driving force of the flow is due to the thermal expansion of the incompressible fluid under different temperature, and Boussinesq assumption is generally applied in engineering applications such the density of the fluid varies due to the volumetric contraction/expansion, and thus the differential density is the driving force due to the gravitational field. The difficult is many legacy CFD methods is because of the numerical smoothing and stabilization formulation usually produce noise much larger than the minute natural convection driving force. The Least-squares based formulation does no use any of these, and the proven SPD (symmetric-positive-definitive) nature of the formulation ensures the existence of the solution. The thermal equations are directly coupled with the Navier-Stoke equations, and produce realistic simulation. Two examples are presented. The first one is the basic
of the thermal convection at the differential
temperature, but it is performed in 3D to demonstrate
the simple and effectiveness of the Sefea CFD
technology. This is the basic of heat tube thermal
exchanger design. In such analysis, the laminar layer on
the surface could be very thin due to the high ratio of
the buoyancy to viscous force, and there must be
enough mesh point in the laminar zone to capture the
physics. Due to the Sefea
velocity/pressure/vorticity/temperature formulation
with much wider local footprint, much less mesh density
can generally be used.
The second example is the adapted 3D Rayleigh Bernard
convection benchmark. In the modified example, a
center rod with a fixed temperature is inserted in the
sealed square box with the outside wall kept at a lower
temperature, and the recirculation of the fluid inside the
box is computed using both water and air to
demonstrate the accuracy and convergence superiority
of LSFEA.
Suggested what-if analysis: change the box size; change
the prescribed temperature; change the fluid type;
change the boundary layer mesh density.
Mutltiphysics for Ironcad Demonstration Models (Rev. 1.1)
9
Impeller/Pump Flow Rate and Housing FSI Analysis (MPIC 2016) Impeller, pump and fan blade designs are an important part of CFD application, and this is an realistic model from an actual model. There are two parts of the design analyses that the customer is interested in. The first part is the flow rate versus the impeller rotation speed as the customer is interested in optimize the design. The second part is the performance of the pump housing in operation as they are interested in knowing whether it will leak at specified flow speed. The pump impeller is rotating at about 3900 rpm with water as the designed fluid and the inflow/outflow opening of 34mm. The fluid on the impeller surface is prescribed to assume the impeller’s rotation speed using the rigid fluid rotational boundary condition*. For the flow rate calculation, only the fluid region is used in the calculation, and the results show consistent inflow and outflow rate at the prescribed impeller RPM. To determine whether the pump will leak under load, fluid-structure-interaction computation is needed by including the pump housing in the analysis. The fluid momentum and pressure are seamless integrated with the continuum stress portion of the housing in the integrated FSI formulation. The pump upper and the lower housing opening is computed with/without contact BC, and if metal material is used, the opening is less than 0.0015mm with the maximum housing stress of around 110 MPa and fluid pressure of 18650 Pascal and -75400 Pascal in intake suction.
In both analyses, the fluid region should have finer mesh on
the flow boundary, and it is refined for the flow rate
calculation. Due to the Sefea
velocity/pressure/vorticity/temperature formulation with
much wider local footprint, much less mesh refinement can
generally be used. Even in the refine mesh, the flow rate did
not change much, and it demonstrates the superior of Sefea
LSFEA technology.
*For MPIC 2016 beta release, due to the rigid fluid rotation BC interface is not
completed yet, it is modeled using a tangential flow based on the cylindrical
coordinate local axis. Also, the FSI case should not be re-meshed till IronCAD patch
is available.
Suggested what-if analysis: change the boundary layer
mesh density; change the fluid type