Advanced Turbomachinery Simulation using STAR-CCM+ Chad Custer, PhD - - PowerPoint PPT Presentation

Advanced Turbomachinery Simulation using STAR-CCM+

Chad Custer, PhD Technical Specialist

STAR-CCM+ is a robust tool well suited for many types of turbomachinery simulation Today’s talk will focus on just a few key objectives and capabilities Key Objectives

– Conjugate heat transfer – Aeroelastic response – Performance mapping

Key Capabilities

– Complex geometry handling – Conformal polyhedral meshing – Pipelined workflow – Harmonic balance – Advanced post-processing

Outline

Key Capabilities Geometry handling

– Direct CAD import – 3D CAD editing

Meshing

– Polyhedral cells – Conformal interfaces – Automatic prism layer generation

Conjugate Heat Transfer

Key Capabilities Geometry handling

– Direct CAD import – 3D CAD editing

Cooled Turbine Blade

External and cooling air volumes generated using 3D CAD Direct import of CAD solid geometry

Key Capabilities Meshing

– Automatic mesh generation

Cooled Turbine Blade

• Pipelined meshing
• Simple global size settings
• Local refinement control
• Automatic solution interpolation

Key Capabilities Meshing

– Automatic mesh generation – Polyhedral cells

Cooled Turbine Blade

Fewer cells required

Key Capabilities Meshing

– Automatic mesh generation – Polyhedral cells

Cooled Turbine Blade

Good for swirling flow such as tip vortices

Polyhedral cell faces are orthogonal to the flow regardless of flow direction

Key Capabilities Meshing

– Automatic mesh generation – Polyhedral cells

Cooled Turbine Blade

High quality cells, even with complex geometry

Key Capabilities Meshing

– Automatic mesh generation – Polyhedral cells – Conformal interfaces – Automatic prism layer generation

Cooled Turbine Blade

Cells are one-to-one connected on the solid/fluid interface Fluid-side prism layers are automatically generated

CHT Validation: NASA C3X Cooled Vane

The C3X is commonly used to validate heat transfer simulation Structured and unstructured (polyhedral) grids with and without transition modeling are analyzed Other work has been performed on the C3X using STAR-CCM+ by

– Solar Turbines (GT2006-91109) – Honeywell (GT2012-68861)

Polyhedral Grid

1.0 million polyhedral cells Extruded cell topology in the span- wise direction y+ of less than one

Structured Grid

1.0 million cells y+ less than one

Heat transfer coefficient is sensitive to inlet turbulent viscosity ratio

HTC sensitivity to Inlet Turbulent Viscosity Ratio

TVR VR=1 =10 TVR VR=4 =40 TVR=100 100 TVR VR=7 =70

Solutions produced with STAR-CCM+ correlate more closely to experiment than reference simulations

Structured Grid: HTC for γ-Reθ Transition model

Hylton, L.D., Mihelc, M.S., Turner, E.R., Nealy, D.A., York, R.E., “Analytical and Experimental Evaluation of the hHeat Transfer Distribution over the Surfaces of Turbines Blades”, NASA CR 168015, May 1983

Structured and polyhedral mesh solutions correlate well

Comparison of Mesh Solutions

Polyhedr hedral al Struc uctur ured ed

Comparison of Heat Transfer Coefficient: No Transition

Polyhedral mesh correlates more closely with experiment when transition is not considered

Polyhedr hedral al Struc uctur ured ed Polyhedr hedral al Struc uctur ured ed

Polyhedral mesh and structured grid produce comparable results when transition is considered

Comparison of Heat Transfer Coefficient: With Transition

Polyhedr hedral al Struc uctur ured ed

STAR-CCM+ accurately predicts surface pressure and heat transfer Transition modeling is important in accurately modeling heat transfer Polyhedral mesh shown to correlate more closely with experiment than structured grid Polyhedral meshing technology allows conformal meshing of complex geometries

C3X Conclusions

GE Energy Efficient Engine

• Simulations are being performed on the GE

Energy efficient engine

• All Cooling holes and internal geometry is

modeled

Traditional simulation methods present many challenges Aeroelastic analysis must be run unsteady Traditional unsteady simulation is challenging

– Very long run times – Must mesh the entire machine – Hard to specify blade vibration – Hard to extract stability information

Aeroelastic Response

• Harmonic balance method in STAR-CCM+

resolves each of these challenges

• The HB method is not available in any other

commercial package

The harmonic balance method takes advantage of the periodic nature of a turbomachine Solves a set of equations that converge to the periodic, unsteady solution Full non-linear solver All unsteady interactions captured

Harmonic Balance Basics

Rapid calculation of unsteady solution Unsteady simulation must be run for many time steps to converge HB simulation converges to the unsteady solution 10x faster

Harmonic Balance Key Benefits

Red: Time Domain Blue: Harmonic Balance

Single blade passage mesh All blades must be meshed for an unsteady simulation Only one blade passage must be meshed for a HB simulation, however the solution is calculated for all blades

Harmonic Balance Key Benefits

Time Domain Harmonic Balance

Specify blade vibration

– The vibration of each blade is staggered. This is known as the “Interblade phase angle” – To determine stability a simulation must be run for each phase angle – Traditional unsteady solvers require manual set up of motion for each phase angle – HB solver takes the inter-blade phase angle as a simple parameter

Harmonic Balance Key Benefits

Small vane motion results in large unsteady response

Example: D2 Vane Flutter

Unsteady Pressure (Pa)

Simulation run for many inter-blade phase angles If the work done on the blade is negative for all inter-blade phase angles, the vane is dynamically stable

Example: D2 Vane Flutter

Key Benefits Complex geometry handling Polyhedral cells High quality mesh Prism layer generation Harmonic balance solver Grid sequencing initialization Efficiency optimization with Optimate+ Turbomachinery specific post-processing

Performance Mapping

Already discussed

Key Benefits Grid sequencing initialization

Performance Mapping

• Drastically reduce run time
• Reduce need for ramping
• Increased simulation

robustness Initialization Converged Solution

Time to initialization: 80 seconds

Key Benefits Efficiency optimization with Optimate+

Performance Mapping

Key Benefits Turbomachinery specific post-processing

Performance Mapping

Blade-to-blade projection

Key Benefits Turbomachinery specific post-processing

Performance Mapping

Meridional projection

Key Benefits Turbomachinery specific post-processing

Performance Mapping

Circumferential Averaging

Aachen turbine is a 1.5 stage cold-flow turbine Simulations will be performed using the harmonic balance method implemented within STAR-CCM+ Analysis will be performed for two different vane clocking positions of +1 degree and -3 degrees

Example: Aachen Turbine Performance Analysis

• 3o

Computational Domain

Structured HOH mesh containing 1.85 million cells Near-wall cell thickness of 0.03 mm 8 cells resolve 0.4 mm tip gap

HB solver converges to the unsteady solution

Residuals

Wakes resolved across interfaces

Mid-span Entropy

• 3o

Time domain and harmonic balance solutions correlate well

Unsteady Rotor Blade Loading

Unsteady wake interaction captured

Velocity Magnitude

Local Efficiency

3

• 3o

Local Efficiency

3

+1 +1o

Time Averaged Local Efficiency

3

• 3o

+1 +1o

• Circ. Averaged Local Efficiency

3

+1 +1o

• 3o

Aachen Turbine Conclusions

HB solver able to calculate the unsteady solution in 1/60th the compute time as a time domain trial (GT2012-69690) Allows for clocking studies to be performed:

– Efficiency of +1 degree clocking: 84.609% – Efficiency of -3 degree clocking: 84.663%

Data can be visualized easily in the time domain or frequency domain

STAR-CCM+ is a robust tool well suited for many types of turbomachinery simulation Today’s talk focused on just a few key objectives and capabilities Key Objectives

– Conjugate heat transfer – Aeroelastic response – Performance mapping

Key Capabilities

– Complex geometry handling – Conformal polyhedral meshing – Pipelined workflow – Harmonic balance – Advanced post-processing

Overview