2008 international ansys · pdf file2008 international ansys conference ... step 3: abaqus...
TRANSCRIPT
© 2008 ANSYS, Inc. All rights reserved. 1 ANSYS, Inc. Proprietary
2008 International
ANSYS Conference
Approved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
NUMERICAL OPTIMIZATION UTILIZING FLUID,
THERMAL AND STRUCTURAL MODELING OF
THE 155 MM NLOS-C MUZZLE BRAKE
Robert Carson – RDECOM/ARDEC/WSEC – Benet LabsJeffrey Greer – RDECOM/ARDEC/WSEC – Benet LabsMark Witherell – RDECOM/ARDEC/WSEC – Benet Labs
© 2008 ANSYS, Inc. All rights reserved. 2 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Non-Line-of-Sight Cannon
• FCS NLOS-C
• 155mm Self-Propelled Howitzer
• 6 Rounds/Min
• 24 Round Magazine
© 2008 ANSYS, Inc. All rights reserved. 3 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
NLOS-C Muzzle Brake
• 3.5 Caliber Length Optimized Muzzle Brake
• Improved Efficiency over M284
• Designed for Maximum Recoil Reduction
• Compatibility with all 155mm Ammunition
© 2008 ANSYS, Inc. All rights reserved. 4 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Problem Statement
• The high rate of fire steady state temperature for the NLOS-C muzzle brake is unknown. Therefore, the muzzle brake was structurally modified to increase material where stress concentrations were highest as shown in previous ambient condition analyses as to ensure survivability at extreme high temperatures. A coupled thermal/structural analysis will provide accurate temperatures allowing for the optimization of the muzzle brake potentially reducing weight by approximately 5-10%.
© 2008 ANSYS, Inc. All rights reserved. 5 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Outline
• One-Way Coupling Structural-Thermal Analysis Approach
– Design of Experiment Setup
– Fluent Steady-State Convection Heat Transfer Analysis
– ABAQUS Unsteady Thermal Analysis
– Fluent Unsteady Structural Loading Analysis
– ABAQUS Unsteady Structural Analysis
– ABAQUS Weight Reduction and Optimization
• Conclusions
© 2008 ANSYS, Inc. All rights reserved. 6 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Fluent to ABAQUS One-Way Coupling
Steady Heat Fluxes Unsteady Temperature
Fluent ABAQUS
Unsteady PressuresDynamic & Static
Stress Analyses
© 2008 ANSYS, Inc. All rights reserved. 7 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
One-Way Coupling Structural-Thermal
Analysis Approach
Steady-State Fluent
CFD Surface
Convection Heat Flux
at Various Wall
Temperature and Inlet
Pressure Conditions.
Fluid Only Modeled.
Polynomial Models of
Surface-Average
Heat Flux
Constructed Using
Designed
Experiments For 33
Different Surface
Locations.
Polynomial Models Utilized
in Unsteady ABAQUS
Thermal Analysis Model of
Solid to Determine End
Temperature Condition after
Firing 96 Rounds. Natural
Convection Assumed
Between Rounds.
Un-steady Fluent
CFD Analysis to
Determine Surface-
Average Pressure
vs. Time Loading
of Muzzle Brake at
33 Different
Surface Locations
Unsteady ABAQUS
Structural Model to
Determine Stress vs.
Time on the Muzzle
Brake. Unsteady
Surface Average
Input from Fluent
Utilized for Loads.
Weight Reduction
Based on
Removing Material
in Areas of Low
Stress. Rerun
Unsteady Thermal
and Structural
ABAQUS Models.
Step 1 Step 2 Step 3
Step 4Step 5Step 6
© 2008 ANSYS, Inc. All rights reserved. 8 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Design Of Experiment (DOE) Setup
• Determining our Design
• Our design will contain curvature,
therefore a Response Surface Method
(RSM) is chosen to approximate the
shape of the surface with a
polynomial.
• 2 Factors are of interest that will have
an impact on heat flux.
• 33 Responses are chosen. Each
represents heat flux area-weighted
averages.
PROCESS
Controllable Factors
Noise Factors
Responses
(4 step process)
Analysis
Conjecture
Design
Experiment
Iterative Experimentation
Optimization
Contour Plots
ANOVA
Empirical Models (Polynomials)
Responses
Process
Factors
Subject Matter Knowledge
RSM Flow Chart
© 2008 ANSYS, Inc. All rights reserved. 9 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Design of Experiment (DOE) Setup
• Identifying Response Surfaces
– The heat flux was averaged over the segregated
surfaces.
– Each surface was a response for the given factors.
– In the D-Optimal architecture, 24 runs populated the
DOE
© 2008 ANSYS, Inc. All rights reserved. 10 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
DOE Populated CFD Runs
• Two Factors
– Wall Temperature
– Static Pressure
© 2008 ANSYS, Inc. All rights reserved. 11 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
CFD Heat Transfer Setup
• Gambit
– One-eighth Section
– 1,112,178 Tet-Cells
• 11 row boundary layer applied.
• 1st Row Height = 0.059 mm
• Fluent
– Density-based, Explicit, Steady,
Node-based
– Propellant Modeled, Volumetric
Reactions Deactivated
– k-epsilon Turbulence Model
© 2008 ANSYS, Inc. All rights reserved. 12 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 1: CFD Heat Transfer Analysis
• Velocity Magnitude
shows excellent
shock structure in the
critical vane areas.
• Velocity vectors show
good flow movement
and turning in the
vanes as well.
© 2008 ANSYS, Inc. All rights reserved. 13 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 1: CFD Heat Transfer Analysis
• Total Temperature is the temperature at the thermodynamic state that would exist if the fluid were brought to zero velocity.
• Shows relative temperature distribution on the different sections of the brake.
© 2008 ANSYS, Inc. All rights reserved. 14 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 2: DOE Analysis
• Significant Model: F-value of 8951 with p-value <
0.0001.
• R-Squared, Adj R-Squared and Pred R-Squared:
Pred R-Squared of 0.9995 is in reasonable
agreement with Adj R-Squared of 0.9992.
• Adeq Precision: Measures signal to noise ratio. A
ratio greater than 4 is desired. Our ratio is 297.
F1-u
© 2008 ANSYS, Inc. All rights reserved. 15 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 2: DOE Analysis
• Residuals should follow a straight
line which, on this specially
scaled graph, indicates a normal
distribution
• Predicted vs. Actual shows how
the model predicts over the range
of data
– Random scatter should occur
about the 45 degree line
• Clusters below or above indicate
under or over prediction
Internally Studentized Residuals
No
rma
l %
Pro
ba
bili
ty
Normal Plot of Residuals
-2.06 -1.04 -0.03 0.98 2.00
1
51020305070809095
99
Actual
Pre
dic
ted
Predicted vs. Actual
-1.00E+08
-7.25E+07
-4.50E+07
-1.75E+07
1.00E+07
-9.96E+07 -7.46E+07 -4.96E+07 -2.46E+07 3.88E+05
© 2008 ANSYS, Inc. All rights reserved. 16 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
DOE Export
• Model Graph shows non-linearity
• The output used is the polynomial of the two factors
• ABAQUS will use these 33 polynomials for the heat flux
-45
216
477
738
1000
1892088
39875886
7785
-1E+8
-7.25E+7
-4.5E+7
-1.75E+7
1E+7
f1-u
© 2008 ANSYS, Inc. All rights reserved. 17 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Maximum Muzzle Brake High Rate of Fire Operating Temperature
– Firing Sequence:
• Fire 1 full magazine
• Rate: 6 shots per minute
• Magazine reload of 12 minutes
• Repeat firing process for 4 magazines
– Boundary Conditions• Worst-Case high temperature conditions
– Temperature dependant natural convection (no wind) h ≈ 7.5
– Solar radiation heat flux ≈ 1,100
– Radiation to ambient (exterior facing surfaces only): Emissivity ≈ 0.88
– Assuming perfect conduction between gun tube and muzzle brake threads and pilots
– Ambient air temperature ≈ 54ºC
– Mesh: utilized linear tetrahedron elements
Step 3: ABAQUS Thermal FEA (Gen 4)
Cm
W2
2m
W
© 2008 ANSYS, Inc. All rights reserved. 18 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 3: ABAQUS Thermal FEA (Gen 4)
Maximum Muzzle Brake High Rate of Fire Operating Temperature
– Thermal Loading:
• 33 surfaces were loaded with the pressure and wall temperature dependent heat fluxes provided from the DoE results
– Example: Surface F1L
• A polynomial equation for pressure vs. time was used to calculate pressures during the firing heating steps
0
1000
2000
3000
4000
5000
6000
7000
8000
9000
0 0.01 0.02 0.03 0.04 0.05 0.06
Time (s)
Sta
tic
Pre
ss
ure
(p
si)
32)( dxcxbxaxf
© 2008 ANSYS, Inc. All rights reserved. 19 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 3: ABAQUS Thermal FEA (Gen 4)
Max Temp after
95th shot = 467 C
Maximum Muzzle Brake High Rate of Fire Operating Temperature
– The maximum temperature of the Gen 4 muzzle brake 10 seconds after the 95th shot = 467ºC = 872ºF
© 2008 ANSYS, Inc. All rights reserved. 20 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 4: Fluent Unsteady Loading Model
• Temperature and pressure patched into gun
barrel based on projectile ready to enter muzzle
brake.
• Flow allowed to expand using unsteady, coupled-
explicit inviscid model.
• Surface average pressure vs. time recorded
during run for multiple surfaces.
• Used as input for unsteady ABAQUS Model
© 2008 ANSYS, Inc. All rights reserved. 21 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
FEA Boundary Conditions
– Mesh: Same mesh as Thermal FEA. Added “mid-side” nodes to create quadratic elements (allows for import of 95th shot temperature field)
– Modeled with temperature dependant material properties
• Modulus of Elasticity
• Poisson’s Ratio
• Yield Strength
• Ultimate Strength
• Specific Heat
• Thermal Conductivity
Step 5: ABAQUS Structural FEA (Gen 4)
© 2008 ANSYS, Inc. All rights reserved. 22 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Maximum Muzzle Brake High Rate of Fire Operating Temperature
– Collar Boundary Condition (simulated twice)
• Modeled statically
• Modeled including gun system dynamics
– Pressure Loading:
• Utilized 3-D CFD transient CFD results for the 33 surfaces
Step 5: ABAQUS Structural FEA (Gen 4)
Collar BC
Muzzle Brake Accel
Gun Tube Accel
© 2008 ANSYS, Inc. All rights reserved. 23 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 5: ABAQUS Structural FEA (Gen 4)
Lowest Factor
of Safety = 1.53 Max Stress
Results: Von Mises Stress Results: Factor of Safety
Notes:
- FOS based upon Temp, Mises Stress, and Yield Strength
- All Gray areas have a FOS > 3.0
© 2008 ANSYS, Inc. All rights reserved. 24 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 6: Weight Reduction Initiative (Gen 5)
Examples of Changes from Gen 4 (~12 in all)
1.) Reduce Lock Key Boss Height
Gen 4 Gen 5
2.) Reduced in Supports on the last 2 Vanes (increased ID of support)
Gen 4 Gen 5
© 2008 ANSYS, Inc. All rights reserved. 25 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 7: ABAQUS Thermal FEA (Gen 5)
Maximum Muzzle Brake High Rate of Fire Operating Temperature
– The maximum temperature of the Gen 5 muzzle brake 10 seconds after the 95th shot = 475ºC = 888ºF
Max Temp after
95th shot = 475 C
© 2008 ANSYS, Inc. All rights reserved. 26 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Step 8: ABAQUS Structural FEA (Gen 5)
Results: Von Mises Stress Results: Factor of Safety
Reference: Stress = 73% Max Stress
Max Stress
Notes:
- FOS based upon Temp, Mises Stress, and Yield Strength
- All Gray areas have a FOS > 3.0
Lowest Factor of
Safety = 1.39
© 2008 ANSYS, Inc. All rights reserved. 27 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
One-Way Coupling Structural-Thermal
Analysis Approach
Steady-State Fluent
CFD Surface
Convection Heat Flux
at Various Wall
Temperature and Inlet
Pressure Conditions.
Fluid Only Modeled.
Polynomial Models of
Surface-Average
Heat Flux
Constructed Using
Designed
Experiments For 33
Different Surface
Locations.
Polynomial Models Utilized
in Unsteady ABAQUS
Thermal Analysis Model of
Solid to Determine End
Temperature Condition after
Firing 96 Rounds. Natural
Convection Assumed
Between Rounds.
Un-steady Fluent
CFD Analysis to
Determine Surface-
Average Pressure
vs. Time Loading
of Muzzle Brake at
33 Different
Surface Locations
Unsteady ABAQUS
Structural Model to
Determine Stress vs.
Time on the Muzzle
Brake. Unsteady
Surface Average
Input from Fluent
Utilized for Loads.
Weight Reduction
Based on
Removing Material
in Areas of Low
Stress. Rerun
Unsteady Thermal
and Structural
ABAQUS Models.
Step 1 Step 2 Step 3
Step 4Step 5Step 6
© 2008 ANSYS, Inc. All rights reserved. 28 ANSYS, Inc. ProprietaryApproved for public release; distribution is unlimited. Case GOVT 08-8120. 7 August 2008
Conclusions
• Seven percent (7%) Reduction of Weight on the Gen 5 Muzzle Brake
• Structural viability of both Gen 4 and Gen 5 verified with FOS measurements.
– Gen 4 – Lowest FOS 1.53
– Gen 5 – Lowest FOS 1.39
• Applicability of the procedure to other components is possible and encouraged.
– Current work on another platform for a tube analysis is underway.