A single-stage rotor-stator simulation can be performed using the Profile Transformation or Fourier Transformation pitch-change method. Compared to a time-marching simulation, a significant reduction in computational time can be achieved by performing the simulation using the Harmonic Analysis (HA) solution method, which avoids the need for solving in time over many disturbance periods in order to reach a steady-periodic state. For details, see Harmonic Analysis in the CFX-Solver Theory Guide.

A single-stage rotor-stator simulation that uses the HA methodology is set up like a single-stage rotor-stator simulation that uses the time-marching method, except:

As described above, a Flow Boundary Disturbance calculation can be performed on two blade passages using the Fourier-Transformation pitch-change method. The solution is obtained faster than for a full-wheel calculation. A significant further reduction in computational time can be achieved by performing the computation using the Harmonic Analysis (HA) (see Harmonic Analysis in the CFX-Solver Theory Guide) solution method, which avoids the need for solving in time over many disturbance periods in order to reach a steady-periodic state.

A flow boundary disturbance simulation that uses the HA methodology is set up like a flow boundary disturbance simulation that uses the time-marching method, except:

A multi-disturbance gust analysis can be modeled using the Fourier Transformation method. In this analysis, the blade is subjected to two simultaneous disturbances; one on the inlet and the other on the outlet of the blade passage. Typically, the inlet and outlet profiles are initially extracted from a steady-state stage simulation of a 1.5 stage configuration. The profiles are then imposed on the inlet and outlet of the rotor passage in a transient gust simulation. This modeling strategy can be used as an alternative for modeling a 1.5 stage compressor or turbine. It accounts for blade passing frequencies from the upstream and downstream rows on the modeled blades. The use of the Fourier Transformation method also allows for very large pitch variations such as with a fan subjected to inflow distortion.

In general, the procedure to set up a multi-disturbance gust analysis with the Fourier Transformation method closely follows the setup of an inlet disturbance problem with the Fourier Transformation method (see Fourier Transformation Disturbance Settings in the CFX-Pre User's Guide). Note the following important steps:

The fundamental input, which is the modal shape of the structure, must be computed and provided in the Ansys CFX .csv file format (see Profile Data Format in the CFX-Pre User's Guide). This file should contain the coordinates, real modal displacements, and complex modal displacements (only for cyclic symmetry models) at the structural mesh locations. For cyclic symmetry cases (usually involving complex mode shapes), the harmonic index (nodal diameter) should also be provided. A sample .csv mode shape profile is shown below. If the mode shape profile was calculated by Ansys Mechanical then System Coupling tools can be used to prepare the .csv file in the correct format. See Aerodynamic Damping Analysis in the System Coupling User's Guide.

For more information on how to export mode shape profiles from Ansys Mechanical, see the following APDL commands:

  [Name]
  normmodeshape

  [Parameters]
  Ncompt = 3
  Nnodes = 4404
  Harmonic Index = 4
  Mass = 7.25811591 [kg]
  Frequency = 1422.69639 [Hz]
  Maximum Displacement = 2.43392589 [m]

  [Spatial Fields]
  Initial X,Initial Y,Initial Z

  [Data]
  Initial X [m], Initial Y [m], Initial Z [m], node [], norm imag meshdisptot x [], norm imag meshdisptot y [], norm imag meshdisptot z [], norm meshdisptot x [], norm meshdisptot y [], norm meshdisptot z [], Sector Tag []
  2.62491941e-001, -1.55315742e-001, -2.57421546e-002, 2.48000000e+002, 3.58961582e-001, 1.76829070e-001, 7.21225142e-001, -2.96412501e-002, -2.16900989e-001, 5.16302824e-001, 1
  2.63363600e-001, -1.53833091e-001, -2.57421546e-002, 2.49000000e+002, 3.52993071e-001, 1.80246934e-001, 7.36300290e-001, -3.26568820e-002, -2.15245709e-001, 4.98975039e-001, 1
  ...
  

A typical blade flutter calculation, whether made using the symmetric sector model or a reduced model, requires modeling several blade passages, which in turn requires expansion of the structural mode shape. The latter can be accomplished using the Edit Profile Data tool provided in CFX-Pre (see Edit Profile Data in the CFX-Pre User's Guide). This tool can replicate as many copies as required by the model and can expand a cyclic symmetry modal shape using the Harmonic Index provided in the file.

Once the mode shape has been imported, it can be visualized to verify that it is aligned with the existing Ansys CFX model before using it in a boundary condition. In case the structural mode shape is not aligned with the model, it must be aligned either using a local coordinate frame once imported, or by external means.

It is highly recommended that you use the automated facility for generating boundary conditions from profiles (see Use Profile Data in the CFX-Pre User's Guide) because it will automatically fill in most of the boundary condition details from the available information in the profile. It is imperative that the settings include the intended maximum amplitude for the deformation of the boundary and the wave traveling direction. For real modal shapes, the Interblade Phase Angle (IBPA) or nodal diameter must also be provided.

While setting up a boundary condition using the structural modal shape, keep in mind that there are two ways of setting up a blade flutter case:

When running blade flutter simulations, consider doing some or all of the following:

As described above, a blade flutter or aerodynamic damping factor calculation can be performed on two blade passages for a range of nodal diameters using the Fourier-Transformation pitch-change method. The solution is obtained faster than for a full-wheel calculation. A significant further reduction in computational time can be achieved by performing the aerodynamic damping computation using the Harmonic Analysis (HA) (see Harmonic Analysis in the CFX-Solver Theory Guide) solution method, which avoids the need for solving in time over many vibration cycles in order to reach a steady-periodic state.

A blade flutter simulation that uses the HA methodology is set up like a blade flutter simulation that uses the time-marching method, except:

In this special situation, where every other row of the turbomachine has same number of blades, we have technically the configuration of a single disturbance being imposed in each passage. Therefore, there is only one pitch ratio present in the configuration. An example of such a case would be a machine in which all stator rows have the same number of blades and all rotors rows have the same number of blades, but the number of blades differs between the stator rows and the rotor rows. This special case can be modeled accurately by using TT correction on all the TRS interfaces (TT-TRS) between rows. Figure 6.8: 1.5 stage machine with equal number of blades in rows 1 & 3 but different than in row 2. Flow can be accurately modeled via a sequence of TT-corrected TRS interfaces. demonstrates an example of a 1.5 stage turbomachine with an equal number of blades in the first and third rows, but a different blade count in the middle row. TT-TRS interfaces have been applied sequentially to this configuration.

To set up this case:

Figure 6.8: 1.5 stage machine with equal number of blades in rows 1 & 3 but different than in row 2. Flow can be accurately modeled via a sequence of TT-corrected TRS interfaces.

This modeling method will allow for capturing accurate blade passing signals in all the rows as well as obtaining accurate machine performance. The limitation on acceptable pitch ratio with use of the TT method still applies to this modeling technique.

In this configuration, each row has a different pitch value or two distinct pitch ratios across the machine, as shown in the following figure. The Time Transformation TRS interface can also be applied sequentially between the IGV and Rotor, and between the Rotor and Stator. This extension of the Time Transformation TRS interface allows for capturing both upstream and downstream blade passing frequencies in the rotor passage, making this modeling method useful for forced response analysis.

Figure 6.9: 1.5 stages with differing blade numbers between rows 1, 2 & 3. Flow can be modeled via a sequence of TT-TRS interfaces.

The following procedure is necessary to model a 1.5 stage turbomachine with two different pitch ratios:

This modeling strategy will allow for obtaining accurate machine performance as well as capturing blade passing signals in rows 1 and 2. On the other hand, row 3 may contain polluted frequency content due to modeling approximation. The limitation on acceptable pitch ratio with use of the Time Transformation method still applies to this modeling technique.

Accurate multistage performance can be predicted by using a combination of Time Transformation TRS and other interface models such as Profile Transformation TRS or Stage interface (mixing-plane). For maximum accuracy, it is recommended that you use Time Transformation TRS at each rotor passage inlet, and a Profile Transformation TRS or Stage interface at the rotor passage outlet. The idea is to minimize modeling error at the TRS interfaces where flow shocks might be crossing by using Time Transformation TRS modeling at those interfaces. The next two figures illustrate two common possible combinations of interfaces for modeling multistage configurations.

Figure 6.10: Multistage modeled with a combination of TT and PT interfaces. For high speed flows, the TT-TRS interfaces should be placed where shocks are crossing between the adjacent passages. This is typically happening at the interface between the exit of the stator row and the inlet of the rotor row.

Figure 6.11: Multistage modeled with a combination of TT, PT and Stage interfaces

The following procedure is necessary to model multistage turbomachines with Time Transformation and Profile Transformation interfaces:

This modeling strategy will allow for obtaining accurate machine performance for turbines and compressors. However, this modeling strategy is not intended to be used for capturing accurate blade passing signals throughout the machine due to the modeling approximation of the Profile Transformation method. The limitation on acceptable pitch ratios using the Time Transformation method still applies to this modeling technique.

For multistage machines, an alternative to using PT-TRS or TT-TRS models on both sides of TRS interfaces is to apply the so-called single-sided TT interface (STT-TRS). The STT-TRS interface model applies the PT model to one side of the interface (typically the upstream side) and the TT model to the other side (typically the downstream side).

Also, you can reverse this correction if necessary (for example, if you want to capture signals from the downstream component). The advantage of using STT-TRS over PT-TRS is that STT-TRS maintains the upstream blade wake thickness as it convects downstream. This modeling strategy results in more accurate accounting of machine performance and losses.

For maximum accuracy, it is recommended that you use TT-TRS at the rotor passage inlet and STT-TRS on the rotor passage exit. Figure 6.12: Multistage modeled with combination of TT-TRS and STT-TRS interfaces illustrates how the STT-TRS is used with TT-TRS in modeling a multistage machine.

Figure 6.12: Multistage modeled with combination of TT-TRS and STT-TRS interfaces

It is also possible to apply sequentially several STT interfaces, as shown in Figure 6.13: Multistage modeled with several STT-TRS interfaces in sequence.

Figure 6.13: Multistage modeled with several STT-TRS interfaces in sequence

The following procedure is necessary to model multistage turbomachines with TT-TRS and STT-TRS interfaces:

This modeling strategy can improve the accuracy of machine performance predictions for both turbines and compressors. However, this type of modeling is not intended for capturing accurate blade passing signals throughout the machine due to the modeling approximation of the STT-TRS method. The limitation on acceptable pitch ratio with use of the TT method still applies to this modeling technique.

In this case, the impeller experiences a long-period disturbance due to the asymmetry of the volute.

To set up an impeller in a vaneless volute:

Note that a similar procedure could be used to set up a case involving a fan in a crosswind.

In this case, the impeller experiences two flow disturbances: one short-period disturbance due to the vanes; one long-period disturbance due to the asymmetry of the volute.

To set up an impeller in a vaned volute: