The following topics related to applying settings for an acoustic analysis solution are available:
To specify the modal analysis type, issue the ANTYPE,MODAL command.
The eigenvalue solution can include damping effects from
the impedance boundary
(SF,Nlist,IMPD), the absorbing element
(FLUID130), or an PML/IPML absorbing boundary. The absorbing element and PML/IPML are
used for termination of the infinite acoustic domain.
The sloshing effect can be modeled
(SF,Nlist,FREE and
ACEL commands) in an acoustic modal analysis.
Excitation sources are ignored.
To specify solver options for the modal analysis, issue the MODOPTcommand. Following are the eigen equations and solvers available in an acoustic modal analysis.
Table 10.1: Acoustic Eigen Equations and Solvers
| Eigen Matrices | Damping | FSI Coupling |
MODOPT,Method
Option | KEYOPT(2) |
|---|---|---|---|---|
| Unsymmetric | NO | YES | UNSYM | 0 |
| Unsymmetric | YES | YES | DAMP | 0 |
| Symmetric | NO | NO | LANB,SUBSP | 1 |
| Symmetric | YES | NO | DAMP | 1 |
When damping is present, the eigensolutions are complex. For more information, see Complex Eigensolutions.
Specifying a proper frequency range results in efficient and accurate eigenvalue
calculations. Input a lower-end frequency just below the anticipated frequency
(FREQB on the MODOPT command),
then specify an upper-end frequency (FREQE). Specify
NMODE to request the number of modes to extract.
To enable modal solution viewing, and to perform other postprocessing options, specify the number of modes to expand (for example, calculate and write the element solution to the results file). Expanding the modes is required if you intend to postprocess the element data. To specify the number of modes to expand, issue the MXPAND command (valid only within the first load step).
The participation factor table lists participation factors, mode coefficients, and mass distribution percentages for each mode extracted. See Participation Factor Table Output in the Structural Analysis Guide for details. For FSI applications, the effective masses are calculated when all the following conditions apply:
They reflect both the mass of the structure and the mass of the fluid.
When either PML or IPML is used to terminate the infinite domain in a modal solution, one buffer element between the PML region and the resonant structure is recommended to avoid the spurious modes. You should evaluate the modal patterns to ensure the correct modes are obtained.
To specify the harmonic analysis type, issue the following command:
ANTYPE,HARMIC,Status,LDSTEP,SUBSTEP,Action
Two harmonic analysis methods are available:
For more information, see Full Harmonic Analysis in the Structural Analysis Guide.
The following topics related to harmonic acoustic analysis settings are available:
Set the frequency using the following command:
HARFRQ,FREQB,FREQE,--,LogOpt
|
To perform a frequency sweep, specify the frequency
range via the HARFRQ command. You can perform a harmonic
analysis over an nth-octave band or
general-frequency band with a logarithm frequency span (the
HARFRQ command with
LogOpt). Using the logarithm frequency span may
result in unexpected octave band sound pressure level (SPL). The uniform
frequency increment should be defined for band SPL.
The octave band is defined based on 
and 
, where:
f0 = central frequency of the octave band (f0 = 16, 31.5, 63, 125, 250, 500, 1000, 2000, 4000, 8000, 16000 Hz for the specified 11 octave bands)
n = 1 (octave band), 2 (1/2 octave band), 3 (1/3 octave band), 6 (1/6 octave band), 12 (1/12 octave band), and 24 (1/24 octave band)
For more information, see Logarithm Frequency Spacing in a Harmonic Analysis in the Mechanical APDL Theory Reference.
Example 10.1: Harmonic Analysis Over a 1/3 Octave Band
The central frequency is 1000 Hz and 10 frequencies are specified:
harfrq,1000,,,OB3 nsubst,10
You cannot restart a harmonic analysis. If you want to apply a different set of harmonic loads, you must perform a new analysis each time.
Specify harmonic analysis options using the following command.
HROPT,Method,MAXMODE,MINMODE,MCFwrite,Damp,
MCkey
The following solvers are available (EQSLV) for a full harmonic analysis:
Sparse direct solver (SPARSE) (default)
Quasi-Minimal Residual iterative solver (QMR)
Incomplete Cholesky Conjugate Gradient iterative solver (ICCG)
Jacobi Conjugate Gradient iterative solver (JCG)
For a relatively small problem (500,000 degrees of freedom [DOFs] or less), the default sparse solver is recommended. For the number of DOFs per element, see Table 2.1: Acoustic Element Properties.
For larger problems (1 million or more DOFs), consider using the sparse direct solver, and ensure that the solution is running in-core for optimal performance. (For more information, see Direct (Sparse) Solver Memory Usage in the Performance Guide.) If the model is too large to run in the sparse direct solver in-core mode with the available hardware RAM, consider using one of the following iterative solvers: QMR (for symmetric matrices); ICCG or JCG (for unsymmetric matrices). (If the iterative solutions diverge, however, you must use the sparse direct solver.)
Using the symmetric formulation in a harmonic analysis reduces computational requirements.
Select the symmetric matrix equation for the coupled problem in the preprocessor (/PREP7) using either the ET or KEYOPT command.
For more information, see Coupled Acoustic Fluid-Structural System with Symmetric Matrix Equation for Full Harmonic Analysis in the Mechanical APDL Theory Reference.
Both the total and scattered pressure formulation are available for analytic incident wave sources.
Activate the pure scattered formulation via the ASOL command.
To control the sound pressure field output for an acoustic scattering
analysis, issue the ASCRES,Opt
command. You can output either the total sound pressure
(Opt = TOTAL) or the scattered pressure
(Opt = SCAT).
To calculate the far-field parameter, define the equivalent source surface
that encloses the scatter
(SF,,MXWF).
Example 10.3: Specifying the Scattered Pressure Formulation
block,0,xs,0,ys,0,zs ! geometry of model … awave,1,dipole,pres,ext,p0,ang,-xs,-ys,-zs ! incident wave outside of model asol,scat,on ! activate scattered formulation ascres,total ! output total pressure
For more information, see Pure Scattered Pressure Formulation in the Mechanical APDL Theory Reference.
An RPM value can be defined via the MRPM command at each load step during a multi-load solution. These RMP values are used during post-processing.
Example 10.4: Multi-Load Solution with Different RPMs
/solu … mrpm,rpm_1 ! first rpm harfrq,FreqB_1,FreQE_1 ! first frequency range nsub,nsub_1 ! number of substeps solve ! first solution … mrpm,rpm_2 ! second rpm harfrq,FreqB_2,FreQE_2 ! second frequency range nsub,nsub_1 ! number of substeps solve ! second solution … finish
A complete description of the mode-superposition analysis procedure is available in Mode-Superposition Harmonic Analysis in the Structural Analysis Guide.
The following conditions apply specifically to acoustic analysis:
When the symmetric formulation is used, global system matrices are badly conditioned and the eigensolver may not succeed in extracting all eigensolutions accurately. In this case, the unsymmetric (default) formulation along with the unsymmetric eigensolver (MODOPT,UNSYM with
ModType= BOTH) is recommended. The mode-superposition method is well adapted for FSI analysis with structural loads and/or the following fluid loads: F,,FLOW and SF,,FREE (sloshing).The QR damped eigensolver method does not apply to FSI analysis because unsymmetrical coupling is ignored with calculating the undamped modes used to obtain the damped modes.
The mode-superposition method is not supported if damping is present. In particular, if acoustic damped boundary conditions are specified (SF with
Lab= IMPD, ATTN, INF, or BLI) or if fluid dynamics viscosity (MP,VISC), perforated material (TB withLab= PERF or AFDM), PML or IPML absorbing condition, or the absorbing boundary elements (FLUID129 or FLUID130) are defined.Only modal damping (MDAMP or DMPRAT) is supported.
Note: Unlike in a FULL harmonic analysis where the constant structural damping (DMPRAT) is only applied to structural degrees of freedom, in a mode-superposition harmonic analysis it is applied to the whole reduced matrix (both structural and acoustic degrees of freedom). To be consistent, you should run a damped modal analysis with constant structural damping beforehand to extract the modal damping ratios, which can then be used in the mode-superposition harmonic analysis using the MDAMP command.
The residual response (RESVEC with
KeyResp= ON) is supported when requesting real solutions (Cpxmod= REAL) and both left and right eigenvectors (ModType= BOTH).The residual vector ( RESVEC with
KeyVect= ON) and the enforced motion (MODCONT withEnforcedKey= ON) methods are not supported by the unsymmetric eigensolver. As a consequence of the second limitation, non-zero pressure and velocity (D withLab= PRES, SF withLab= SHLD, AWAVE, and BF withLab= VELO) are ignored.Use a sufficient number of modes to obtain an accurate pressure solution. The upper frequency times two for the modal base may be insufficient. In general, as the number of modes increases, the convergence of the pressure solution is slower than that of the displacement solution, especially far from the resonance frequencies. Using the residual response method may help the convergence.
For an example, see VM282
in the
Mechanical APDL Verification Manual.
To specify the transient analysis type, issue the ANTYPE,TRANS command.
To obtain an acceptable solution in an acoustic transient analysis, the time increment Δt is determined by Δt =1/(2fmax). Estimate the maximum operating frequency to determine the mesh size in the model. (See Specifying Acoustic Analysis Region Attributes and Meshing.)
Consider using the sparse direct equation solver (EQSLV) with automatic time stepping disabled. For linear analyses, matrix factorization need only be performed once (for purely acoustic) or twice (for coupled vibro-acoustic). The remaining time steps do not require matrix factorization, resulting in very efficient solutions.
The spatial distribution of the pressure field must also be taken into account for meshing. The mesh should be fine enough to resolve the spatial variation of the pressure.
For example, if a sound dipole with a 1 m separation radiates sound in air at 20 Hz, setting the mesh size as 1.7 m for low-order elements (10 elements/per wavelength) is unreasonable. If a 10-degree increment is used to discretize the circumference around the dipole, the minimum mesh should be 0.087 m. Both the maximum operating frequency and the pressure spatial variation must be taken into account.
An acoustic transient analysis performs a standard transient linear solution.
Example 10.5: Defining an Acoustic Transient Solution
num_timestep = 8
frq_step = 2000
dim_waveleng = 1500/frq_step
time_end = dim_distance/1500*3
time_step = 1/frq_step/num_timestep
num_steps = time_end/time_step
num_elements = 8
dim_esize = dim_waveleng/num_elements
*dim,load_time,table,4,,,time
load_time(1,0)=0,1/frq_step,2/frq_step,time_end
load_time(1,1)=0,1,0,0
…
et,1,220,,1 ! second-order fluid220
et,2,130,2 ! second-order fluid130
…
/solu
antype,trans
trnopt,full
autots,off
time,time_end
deltim,time_step
outres,nsol,all
! time varying load
f,node(0,0,0),flow,%load_time%
solve
finish
For a transient analysis that includes an acoustic free surface (sloshing effect) and a ramped gravity acceleration, it is necessary to use a uniform time step. In addition, as the acoustic sloshing mass matrix depends upon gravity (ACEL), the mass matrix must be updated (KUSE,-1) in the first load step, when gravity is applied. For subsequent load steps, the factorized matrices can be reused (KUSE,0).
If no pressure boundary condition is prescribed, a zero frequency fluid mode exists. In this case, weak pressure spring-dampers (COMBIN14 with KEYOPT(2) = 7) can be added on the free surface nodes to constrain and possibly damp the fluid vibrations (see constants k and CV1 in Equation 13–9 and Equation 13–10 of the Mechanical APDL Theory Reference).
By default, a transient solution uses the pressure formulation which requires either acceleration or mass source rate excitation. As an alternative, you can specify the velocity potential formulation (KEYOPT(1) = 4) which solves a transient with either the velocity or mass source excitation.
When PML or IPML is used to absorb the outgoing sound wave in a transient analysis, additional auxiliary variables (degree-of-freedom labels: VX, VY, VZ and ENKE for 3D acoustic elements; VX, VY and ENKE for 2D acoustic elements) are introduced on the nodes of the PML/IPML elements. That is, there are five degrees of freedom per node for PML and IPML elements. The number of PML elements should be carefully controlled so that a large number of degrees of freedom are not added to the solution. The program applies constraints for the additional variables on the PML/IPML elements.
A complete description of the spectrum analysis procedures (single point response spectrum, multiple-point response spectrum, and random vibration) is available in Spectrum Analysis in the Structural Analysis Guide.
The conditions listed under Mode-Superposition Harmonic Analysis apply for spectrum analysis except that fluid loads are not supported. Additional limitations are listed below:
The solution from the modal analysis must be real to perform a subsequent spectrum analysis.
For an example of a single-point response spectrum (SPRS) analysis, see Example: Spectrum Analysis of a Cylindrical Tank Filled with Water.
A steady-state analysis can be used to model room acoustics. To specify the steady-state analysis type, issue the ANTYPE,STATIC command.