3.4. Modeling Hints and Examples

3.4.1. Adding a Stationary Part

The stationary portion of your model could be a housing, a fixed support, or a flange. To add a stationary part, first create the part mesh. Since the rotational velocity is applied only to the rotating part of the structure, you need to create a component based on the elements of the rotating parts.

An example input to create a rotating component and apply the component rotational velocity using the CMOMEGA command follows:

! create the model

! create the rotating component
esel,,type,,1,2
cm,RotatingPart,elem
allsel

! apply rotational velocity to rotating component only
cmomega,RotatingPart,1000.

3.4.2. Transforming Non-Axisymmetric Parts into Equivalent Axisymmetric Mass

If your model is composed of a non-axisymmetric part, you can transform it into an equivalent axisymmetric mass using the following procedure.

  • First select the non-axisymmetric part volumes using VSEL command

  • Enter the VSUM command to printout global mass properties of these volumes.

  • Delete all the volumes.

  • Define a new mass element (MASS21) on a node located at the center of gravity of the volumes. Real constants are the calculated mass and rotary inertia properties. These characteristics are approximated to obtain the axisymmetry. For example if the rotational velocity axis is along X, then the mass in Y and Z directions, along with the rotary inertia YY and ZZ are equal.

  • Define a rigid region between the mass element node and the rest of the structure using the CERIG command.

You can obtain more precise mass, center of mass and moments of inertia by using inertia relief calculations. For more information, see Mass Moments of Inertia in the Mechanical APDL Theory Reference.

3.4.3. Defining Multiple Spools

To define several rotating parts, first create the part meshes. Since each part has a different rotational velocity, you need to define each part as a component based on the elements.

An example input to create two rotating components and apply the component rotational velocities using the CMOMEGA command follows:

! create the model

! create the first rotating component
esel,,type,,1,2
cm,RotatingPart1,elem

! create the second rotating component
esel,inve
cm,RotatingPart2,elem
allsel

! apply rotational velocities to rotating components
cmomega,RotatingPart1,1000.
cmomega,RotatingPart2,3000.

3.4.4. Using Component Mode Synthesis (CMS) for Rotating Parts

The component mode synthesis (CMS) procedure can be used to model rotating parts in a rotordynamics analysis.

When the CMS procedure is used, outputs specific to rotating structure dynamics are limited to:

  • The animation of the whirl (ANHARM) which is available before and after the CMS expansion pass

  • The Campbell diagram (PRCAMP and PLCAMP) which is available only before the expansion pass

Orbits (PRORB and PLORB) are not available.

3.4.4.1. Generation Pass (Creating the Superelement)

In this step, the rotor is reduced to two CMS superelements. The following example input illustrates the generation pass:

! Generation Pass (Creating the Superelement)

/filname,part
/solu
antype,substr
seopt,part,3,1
cmsopt,fix,40 	


cmomega,comp1,1000     ! Rotational velocity of component ‘comp1’ (previously defined) is 1000rd/s about global X-Dir 
coriolis,on,,,on       ! Coriolis on, in a stationary reference frame

cmsel,s,comp1   
cmsel,s,interface2
m,all,all        
nsle      
     
solve     
allsel,all,all 
save     
fini

You can specify the rotational velocity of the rotating superelement using the OMEGA and CMOMEGA commands. The Coriolis effect is included using the CORIOLIS command in a stationary reference frame (CORIOLIS,ON,,,ON). In the generation pass, only gyroscopic damping can be used. Other types of damping such as element damping and Rayleigh damping (DMPRAT, ALPHAD, BETAD, MP,ALPD, MP,BETD, TB,SDAMP,,,,ALPD, and TB,SDAMP,,,,BETD) are not supported. Damping is defined in the use pass step.

3.4.4.2. Use Pass

Define a superelement by reading in the superelement matrices using the SE command to build the full model and perform the desired analysis.

The following example input illustrates reading superelement matrices in the use pass, and a modal analysis procedure:

! Use Pass

/filname,use
/prep7

et,1,matrix50
type,1       
se,part

save     
fini

! Perform Modal Analysis
/solu

antype,modal
modopt,QRDAMP,20,,,ON
mxpand,20,,,yes

solve
fini

In the use pass, you can specify damping using DMPRAT, ALPHAD, BETAD, MP,ALPD and MP,BETD. If non-rotating parts like bearings and supporting structure are present, they are modeled during the use pass. The CMOMEGA command is used to specify which parts are rotating and which are not.

The following example input illustrates modeling of a rotating superelement and a non-superelement in the use pass:

! Generation Pass

/filname,part
/solu
antype,substr
seopt,part,3,1
cmsopt,fix,Nmode 	

cmomega,comp1,1             ! Rotational velocity of ‘comp1’ (previously defined) is 1rd/s about global X-Dir (definition of the direction only)  
coriolis,on,,,on            ! Coriolis on, in a stationary reference frame

cmsel,s,comp1   
cmsel,s,interface2
m,all,all        
nsle      
     
solve     
allsel,all,all 
save     
fini

! Use Pass

/filname,use
/prep7

!Superelement 
et,1,matrix50
type,1       
se,part

! Non-superelements mp,ex,1,2.0e11
mp,dens,1,7800
mp,nuxy,1,0.3

et,2,188
type,2
secnum,2
sectype,2,beam,ctube
secdata,0.20,0.30,1,16

k,3,3,0,0
k,4,7,0,0

l,3,4
lesize,1,,,8

type,2
secnum,2
mat,1
lmesh,1
allsel,all,all

et,4,214                    ! Creation of symmetric Bearing
keyopt,4,2,1
keyopt,4,3,0
real,4
r,4,1.0e6,1.4e5,,,10,10
rmore,,,,

n,1000,5,0.22,0

type,4
real,4
e,node(5,0,0),1000
allsel,all

nsel,s,loc,x,0
nsel,a,loc,x,10
d,all,all,0
nsel,all

d,1000,all,0
allsel,all

cpintf,all
esel,s,type,,2
cm,comp2,elem               ! Creation of non-superelement component
allsel,all,all

save
fini

! Perform Modal Analysis
/solu
antype,modal
modopt,qrdamp,20,,,on
mxpand, 20,,,yes
cmomega,comp1,1000,0,0      ! Rotational velocity of ‘comp1’ is 1000 rd/s about global X-Dir 

cmomega,comp2,1000,0,0      ! Rotational velocity of ‘comp2’ is 1000 rd/s about global X-Dir 
coriolis,on,,,on

solve
fini

In general, you should specify a unit rotational velocity vector in the generation pass and then specify the true rotational velocity during the use pass. The Coriolis effect must be activated during both the generation and use passes in this case (CORIOLIS,ON,,,ON). Internally, the reduced gyroscopic matrix from the generation pass is scaled with the amplitude of the rotational velocity vector defined during the use pass. This procedure is used in particular for the generation of a Campbell diagram where modal analyses are performed for different rotational velocities.

The following example input illustrates the scaling of the gyroscopic damping matrix to print Campbell diagram data:

! Generation Pass (Creating the Superelement)

/filname,part
/solu
antype,substr
seopt,part,3,1
cmsopt,fix,40 	


cmomega,comp1,1.0           ! Component ‘comp1’ (defined previously) is rotating about global X-Dir
coriolis,on,,,on            ! Coriolis on, in a stationary reference frame

cmsel,s,comp1   
cmsel,s,interface2
m,all,all        
nsle      
     
solve     
allsel,all,all 
save     
fini

!Use Pass

/filname,use
/prep7

et,1,matrix50
type,1       
se,part

esel,s,type,,1
cm,rot1,elem
allsel,all,all

save     
fini

!Perform Modal Analysis
/solu

antype,modal
modopt,qrdamp,20,,,ON
mxpand,20,,,yes

coriolis,on,,,on            ! Coriolis on, in a stationary reference frame

! First rotational velocity 
cmomega,rot1,1000           ! Component ‘rot1’  is rotating at 1000rd/s about global X-Dir (Scaling)
solve

! Second rotational velocity
cmomega,rot1,2000           ! Component ‘rot1’  is rotating at 2000rd/s about global X-Dir (Scaling) 
solve

fini

/post1

prcamp,,,,,rot1             ! Print Campbell diagram information for component ‘rot1’
fini

If the true rotational velocity is specified in both the generation and use passes, the resulting gyroscopic effect will be based on the square of the rotational velocity, due to an internal scaling of the gyroscopic matrix. Make sure you use unit rotational velocity during the generation pass to ensure the correct scaling.

To compute and postprocess quantities using nodal velocities and nodal accelerations (damping force, inertial force, kinetic energy, etc.) resulting from the gyroscopic effect, the OUTRES command with DSUBres = ALL must be issued in the first load step of the use pass. These quantities are not computed if:

  • At least one superelement load vector is applied in the use pass (SFE,,,SELV)

  • The use pass analysis is a mode-superposition method (TRNOPT,MSUP and HROPT,MSUP)

For more information, see Component Mode Synthesis in the Substructuring Analysis Guide.

3.4.4.3. Expansion Pass

This step is the same as described in The CMS Use and Expansion Passes in the Substructuring Analysis Guide.

3.4.5. Verifying Your Model with the Rotor Mass Summary Printout

You can verify the rotating parts modeling using the rotor mass summary printout activated via RotMass = ON on the CORIOLIS command. It is provided for line element models only. This summary contains the following information for each rotating part:

  • Number of nodes

  • Spin direction

  • Mass summary, including the details of inertia properties at nodes

  • Center of gravity

To see an example of the rotor mass summary printout, see Example: Quasi-Static Analysis of a Multi-Rotor System.