FAQ

Check our FAQ before adding questions.

Assessing Results

Yes, Code Leo can write solution files in Plot3D, Tecplot, Fieldview and VTK format for use with 3rd party post-processing / visualization packages.  

For Plot3D, Tecplot and Fieldview, select the appropriate IPLOT parameter in the "Configure Solver Settings" section of the configuration wizard.

IPLOToptions

 

For VTK (ParaView), right click on a completed restart file, select "Conversion" and "Convert to VTK".

ConvertToVTK

The .LOAD file contains flow properties on the airfoil surface at five different spanwise locations.

The .INOUT file contains mass-averaged information about flow variables at the inlet and exit of the domain.

The .SURFDATA file is used to create three-dimensional contour plots that are limited to the airfoil surface in order to avoid the file sizes necessary to save the entire fluid domain.

In addition to an updated restart file, Cloud Leo produces several turbomachinery-specific output files for your convenience:

Output File Type

Description

.CONVERGENCE Summary of solver settings used in a given simulation, as well as a convergence history of the residuals.
.OVERALL Contains history of performance data from the simulation as well as information regarding conservation
.MEANVALUE Summary of performance data for a given row at the last iteration conducted
.PSUF Contains a time history of pressure loading on the surface of the airfoil
.MIXEDAVG Summary of circumferentially averaged radial profiles at the inlet, exit, leading edge and trailing edge of a given airfoil row
.FORCES History of the forces acting on the airfoil as well as the power being added to or extracted from the system
.STATION Summary of flow data mass-averaged over I-planes up and downstream of the airfoil
.SURFDATA File for creating three-dimensional contour plots that are limited to the airfoil surface
.INOUT Summary of mass-averaged information about flow variables at the inlet and exit of the analysis domain
.LOAD Summary of flow properties on the airfoil surface at five different spanwise locations

 

For multi-row cases, two additional files are provided:

Output File Type

Description

-MPI-.OVERALL History of performance data for the entire multi-row simulation
-MPI-.MEANVALUE Summary of performance data for the last iteration of the entire multi-row simulation

Cloud Leo

Here are six tips to getting the most out of the automated speedline capabilities in the ADS Workbench:

  1. Begin with an operating point that is close to predicted peak efficiency for each speedline.  While the automated speedline generator will function with any converged operating point between numerical stall and choke, it works most efficiently when the operating point is close to predicted peak efficiency. This will reduce the number of points required to construct the speedline, and as a result, the turnaround time and computational cost. 
  2.  

  3. Make sure the initial operating point is properly defined and converged.  Take extra care to ensure you have properly captured key design features and aerodynamic conditions.  Garbage in = garbage out!

     

  4.  

  5. Double check your runtime preferences to ensure maximum throughput. Be sure to double check your runtime preferences in the Window>Preferences>Run section of the workbench to make sure the speedline generator can take advantage of as many solvers as you have licensed.
  6.  

     

  7. Let the workbench do its thing.  Once the automated compressor speedline flow has been initiated, give it time to do its thing.  It will first initiate simulations at 95%, 98%, 100%, 102% and 105% of exit static pressure, then three points at a time to work its way towards numerical stall and choke.  The process usually requires 3-5 additional iterations, resulting in 12-20 operating points per speedline.
  8.  

  9. Use the "Extend Stall" and "Extend Choke" options to creep closer to numerical stall and choke.  These options will add three points between the last known converged and unconverged points near stall and choke respectively.
  10.  

    toptips02

     

  11. Use the "Find Target Mass Flow" option to obtain predictions for a targeted mass flow rate.  In Code Leo, the mass flow rate is a dependent variable, not an independent variable.  With this option, the ADS Workbench attempts to pinpoint the exit static pressure value that produces the targeted mass flow rate.  Three operating points are configured and queued for execution based on the targeted mass flow rate and derived relationship between exit static pressure and mass flow rate.  

toptips03

 

The number of cloud instances is controlled by the max concurrent cpu usage setting.  This can be changed by going into Window->Preferences->Run Preferences and changing the maximum concurrent cpu usage value.  

It is important to note that the number of cpus per instance depends on the instance type used.  If your instance type is cc1.4xlarge (which equates to 8 cpus), a value between 1 and 8 will boot 1 instance, a value between 9 and 16 will boot 2 instances, and a value between 17 and 24 will boot 3 instances, and so on.

It is not necessary to have enough cpus to run your case.  If a smaller cluster is booted than needed, the additional LEO processes will time-share the cpus in your cluster.   However, this will result in slower turnaround times.

Fast turnaround time is achieved at several levels in Cloud Leo.

At a solver level, Code Leo applies a number of techniques.  Local time stepping is employed, as well as pre-conditioning with gauge pressure to speed up convergence for low speed flow problems.  We also employ two convergence acceleration techniques: a residual propagation method for unstructured mesh, and a multi-grid scheme for structured mesh.  For time accurate simulations that employ wall integration, we employ dual time stepping to facilitate convergence.  Finally, for highly skewed meshes, our cell-vertex numerical scheme produces accurate results with fewer elements compared to cell-centered schemes.  The need for fewer elements under this condition also results in improved turnaround time.

At an execution level, we support parallel execution over an MPICH2 ring, and meshes may be partitioned into O and H mesh components for even faster throughput.

At a hardware infrastructure level, we employ only the Amazon EC2 cluster compute instances, which currently include 8 and 16 core machines designed specifically to handle our class of engineering/scientific application.  For more information about Amazon EC2 instance types click here.

Cloud Leo supports the following analysis types:

  • 2D cascade analysis for airfoil section design (steady or unsteady)
  • 3D single row steady analysis
  • 3D multi-row steady analysis for multi-stage compressor and turbine design
  • 3D time accurate analysis for the study of blade row interaction effects, endwall secondary flow and rotating stall

In addition, Cloud Leo supports automated speedline generation to enable you to easily construct compressor and turbine maps with no manual intervention. 

General

Our subscription plans will be comprised of two components:

  • A base monthly fee, which covers the cost of the workbench, the mesh generator Code Wand and technical support
  • A usage fee, which is based on the number of solver-hours you've utilized for the month, less any hours that might be included as part of your subscription plan

In addition, a 30 day free trial is included with every subscription.

 

Cloud Leo is designed to put the full force of world-class turbomachinery CFD in the hands of compressor and turbine aero designers.  

We aim to address three primary challenges:

  • Deliver field-proven, aerospace-class CFD you can count on for turbomachinery aero
  • Remove arbitrary capacity constraints that limit your ability to conduct aero design
  • Dramatically lower the upfront costs required to use CFD

Using Cloud Leo, you’ll be able to put massive CFD analysis capacity to work on your behalf on a pay as you go basis.  No capacity constraints.  No cluster to manage.  No exorbitant upfront fees.  

We take security very seriously and work with Amazon Web Services and our credit card processor to assure its integrity in Cloud Leo:

  • Access to Cloud Leo resources is controlled. Access to Cloud Leo is gained through a custom license manager built specifically for you, using information provided at registration and secret keys you have with Amazon, to ensure that you and only those you authorize can use the service.
  • All data transmissions between the workbench and Cloud Leo are authenticated and encrypted.  All transmissions of data between the workbench and Cloud Leo are authenticated via SSH and encrypted using 256-bit encryption.
  • Your data does not persist in Cloud Leo.  All simulation data is purged from Cloud Leo instances provisioned for a simulation and the instances themselves are deleted once simulation results have been successfully downloaded back to the workbench for post-processing.
  • Your credit card information is handled by a PCI compliant card processor. No credit card data is stored on our web servers.  All credit card information provided during registration is managed by our credit card processor, Stripe.  Stripe is PCI compliant and certified as a Level 1 Service Provider, the most stringent level of certification available.  Click here for more information about Stripe's security procedures. 
  • Your cloud computing infrastructure provider is SAS 70 Type II certified, ISO 27001 certified, and PCI compliant.  Cloud Leo runs on cloud computing infrastructure provided by Amazon Web Services (AWS).  Amazon has many years of experience designing large data centers and has comprehensive policies in place to ensure physical security, secure services and data privacy.  Click here for more information about AWS security. 

The workbench for Cloud Leo can be installed on Windows XP, Vista, Windows 8, and Red Hat Linux 5.2+.

Introduction to Cloud Leo

This tutorial provides a quick overview of Cloud Leo and how it can be used to conduct compressor and turbine aero analysis with ease.  Cloud Leo workflow and solutioncomponents are introduced to the viewer as It introduces the components and the workflow used to easily configure, execute and post-process.  Time: 4:54.  Watch

Screen Shot 2013-01-23 at 5.24.20 PM

Imagine the possibilities when you have massive aerospace-class CFD capacity at your fingertips, with the ability to scale your use of it based on need and pay for it on a pay-as-you-go basis.

For savvy independent consultants and small design shops, Cloud Leo delivers the following benefits:

  • Greater design robustness
  • Agility
  • Competitive differentiation
  • No clusters to manage/maintain
  • No exorbitant upfront license fees to pay in advance of revenue
  • The ability to pay simply for what you use

Cloud Leo is ideal for independent consultants and small-medium sized design shops with the savvy to take full advantage of CFD but have been constrained from its use due to large upfront license fees, limited analysis capacity and infrastructure costs.

By using Cloud Leo, these individuals and organizations gain the ability to apply massive CFD capacity on-demand to carry out aero design quickly and cost effectively.  They will be able to gain design robustness, agility and the means to differentiate themselves from the competition in the process.

Cloud Leo is accessed through a graphical workbench that you download to your local desktop/laptop.  The workbench allows you to import section data, generate meshes and configure your simulation.  

When you click on “Start Run”, the case files are uploaded securely to Cloud Leo, where one or more CFD analysis instances are dynamically provisioned on your behalf and initiated to carry out your simulation to completion.  You’ll be able to monitor your cases from the workbench or simply walk away.  

Upon completion, the output files from the simulation are downloaded securely back to the workbench for post-processing and analysis. 

Meshing

Yes, bleed and leakage flows may be modeled and analyzed in Code Leo.  They are set at the inner or outer diameter endwalls upstream or downstream of the airfoil.

Film cooling holes may be modeled in Code Leo.  The model allows you to specify local cooling flow injection by row of uniform holes, or by individual hole locadtions.

Recirculation ports may be modeled in Code Leo.  They may be set at upstream and downstream port locations  and flow is dictated by pressure gradient between the port locations.

Cooling flow may be modeled in Code Leo.  The model supports the injection of cooling flow uniformly across a plane for bookkeeping purposes.

Debug output is available in the MESH_DIAGNOSTIC.DATA file to assist in debugging mesh problems.  Additionally, troubleshooting guides are available on the support portal.

Mesh visualization can be done through the workbench using the "View Mesh in Paraview" option.

No, tip treatment meshing is not currently supported by Cloud Leo.

Yes, Code Wand meshes tip and hub clearances, which are defined by distance from endwall at the leading edge, mid-chord and trailing edge.

Yes, Code Wand can mesh radial impellers with and without splitters using standard section data input in a variety of formats.  We recommend at least five sections per airfoil.

Yes, Code Wand can support fillet meshing on pressure or suction side at the hub or tip.  We support constant or variable radius fillets, and we support variable radius fillets with cutoff trailing edges, which is useful for radial impellers.

Yes. Adjacent rows can be offset by a user specified angle to model clocking effects.

Yes, Code Wand can mesh axial rotors and vanes using standard section data input in a variety of formats.  We recommend at least five sections per airfoil.

Yes. For cases with cantilevered stators and hub clearance, the inner drum can be set to rotate at a user specific speed and direction.

Code Wand does not support the meshing of non-axisymmetric components such as volutes.  A third party mesh converted into the restart file format employed by Cloud Leo may be used instead for analysis.  Please contact technical support for more information.

Code Wand does not support the meshing of non-axisymmetric casing treatments.  A third party mesh converted into the restart file format employed by Cloud Leo may be used instead for analysis.  Please contact technical support for more information.

Code Wand supports multiple mesh topologies:

  • OHH mesh for viscous flow analysis with two types of clearance meshing
  • OH mesh for viscous flow analysis with two types of clearance meshing
  • Hybrid H mesh for inviscid and viscous analysis
  • Sheared H mesh for inviscid flow analysis

Though we support several mesh topologies, we strongly recommend using an OHH mesh topology with Type 2 clearance meshing.  This topology surrounds the airfoil with an O-mesh and places H-meshes on the pressure and suction sides of the airfoil as well as at the leading and trailing edges in order to improve resolution of wake capture.  Our innovative tip clearance meshing approach provides both excellent resolution and high efficiency, resulting in a 30% improvement in turnaround time vs. traditional approaches.

Advanced users have the option of directly specifying mesh distribution through the section data input.

You will need the following:

  • Airfoil section data for each airfoil you wish to analyze
  • Endwall geometry in (x,r) format for the flow path
  • Upstream and downstream boundary location sin (x,r) format
  • Aerodynamic conditions
  • Hub and/or tip clearance values

Preparing Geometry

Preparing airfoil geometry for your first case

A key to success with Cloud Leo -- or any CFD for that matter -- is making sure the airfoil geometries are suitable for mesh generation.  This tutorial provides essential tips and guidelines for ensuring your mesh gets done right the first time.   Time: 6:28.  Watch

Screen Shot 2013-01-23 at 5.27.16 PM

The following airfoil section data formats are supported by Code Wand.

 

Supported Input Format

Comments

(x,rθ,r) Defined trailing edge to trailing edge
(x1,y1,z1,x2,y2,z2) Defined leading edge to trailing edge
CFX format (x,y,z) Defined leading edge to leading edge
Suction(x,y,z) & Pressure(x,y,z) Defined leading edge to trailing edge
(x1,y1,y2) Defined leading edge to trailing edge at a given spanwise location
(x,y) at constant radius Defined trailing edge to trailing edge at a given spanwise location
(x1,rθ1,x2,rθ2) Defined leading edge to trailing edge at a given spanwise location
(x1,rθ1,x2,rθ2,r1,r2) Defined leading edge to trailing edge
(x1,t1,r1,x2,t2,r2) Defined leading edge to trailing edge
(-y1,x1,-y2,x2) Defined leading edge to trailing edge at a given spanwise location
(x1,rθ1 all) Read spanwise location, then all x1 values then all rθ1 values

Tutorials

Preparing airfoil geometry for your first case

A key to success with Cloud Leo -- or any CFD for that matter -- is making sure the airfoil geometries are suitable for mesh generation.  This tutorial provides essential tips and guidelines for ensuring your mesh gets done right the first time.   Time: 6:28.  Watch

Screen Shot 2013-01-23 at 5.27.16 PM

Introduction to Cloud Leo

This tutorial provides a quick overview of Cloud Leo and how it can be used to conduct compressor and turbine aero analysis with ease.  Cloud Leo workflow and solutioncomponents are introduced to the viewer as It introduces the components and the workflow used to easily configure, execute and post-process.  Time: 4:54.  Watch

Screen Shot 2013-01-23 at 5.24.20 PM