A Method to Integrate Drive System Design - Georgia tech
1 like629 views
The document presents a study on topology optimization of gear webs to enhance weight savings in aerospace applications, revealing that up to 30% weight reduction can be achieved with non-standard gear designs. It discusses the balance between analysis fidelity and design lifecycle stages, along with the methodology for optimizing gear designs using finite element analysis (FEA). The results indicate a 12% weight saving after optimization, although further research is needed to refine the processes and criteria for future applications.
A Method to Integrate Drive System Design - Georgia tech
- 1. Topology Optimization of Gear Web using OptiStruct Sylvester Ashok and Fabian Zender Integrated Product Lifecycle Engineering Laboratory Daniel Guggenheim School of Aerospace Engineering Georgia Institute of Technology Content in this presentation may not be duplicated, copied, or reproduced, without the permission of the authors or Altair
- 2. Overview • Background • Introduction • Problem Setup • Results • Conclusion
- 3. Background • A framework has been developed to study and optimize rotorcraft drive systems
- 4. Background • One of the biggest problems a designer is faced with is understanding what fidelity of analysis is required to obtain the detail of information needed. • The balance between fidelity and detail at different stages of the design life-cycle, among different disciplines, is one of the keenly researched topics in MDAO. • Weight saving is critical in aerospace applications • How should topology optimization be treated in drive system design?
- 5. Introduction • Gear web and non-standard gear designs can produce up to 30% weight saving. • Gear webs have been designed in detailed stages through bench tests so far. • Topology Optimization is a method that can be used to study this using FEA.
- 6. Introduction σb*= f(x) weight Design B weight Design A Design A Design B σb * σb * x x Analytical sizing result Topology optimization result Since the design space is discrete, if after topology optimization, design B weighs less than design A, topology optimization cannot be considered a design refinement, instead must be used to select the optimum design.
- 7. Problem Setup Parameter Symbol Unit Baseline Design 2 Power P HP 700 RPM - RPM 5000 Gear ratio mg - 3.3 Diametral pitch Pd in-1 6 Face width F in 2.4 Pressure angle φ deg 20 Material - - AISI 9310 Number of teeth Np - 23 24 Tangential load Pt lbf 4616.7 4424.4 Radial load Pn lbf 1680.4 1610.3 Bending Stress (AGMA) σb psi 35000 32951 Weight W lb 7.233 7.885
- 8. Problem Setup • Radioss run for calibrating load • Gear model brought in from CATIA • Load for FEA application was calibrated to obtain analytical bending stress value
- 9. Problem Setup – Topology Optimization • Components • Gear Teeth (blue) • Design Area (red) • Shafthole (blue) • Mesh Setup • Shafthole and Design Area: Edge Deviation, 75 elements per edge • Gear Teeth: QI Optimize, element size 0.06 • 3D Solid Map, size 0.06 • Load Setup • Load was distributed across a single line of nodes at pitch circle • Inside of Shafthole constrained for 6 DoF • Calibrated Load Parameter Symbol Unit Baseline Design 2 Tangential load Pt lbf 6463.4 6194.2 Radial load Pn lbf 2352.6 2254.4
- 10. Problem Setup – Topology Optimization Objective: min : F ( x) Volumedesign Subject to: x X x R n | gi ( x) 0, i 1,..., m Where, g1 ( x) ( f design ) * g 2 ( x) manufacturing constraint (draw) g3 ( x) manufacturing constraint (cyclic and symmetry) Design Non-design
- 11. Problem Setup – Topology Optimization • σdesign< 20ksi • Cyclic Pattern, plane normal to axis of the gear, node (0,0,1.2) and node (0,0,0). Instances = multiple of teeth • Ensure cyclic symmetry since single tooth loading is simulated • Split Draw Constraint, draw direction given by line through center of gear (0,0,1.2) and center of front face (0,0,0) • Ensure symmetry through the thickness for simpler manufacturing and ease of assembly
- 12. Problem Setup – Topology Optimization • Optimized result only maintains structural integrity within design area • Optimized result would lead to increased bending stress in gear teeth
- 13. Problem Setup – Topology Optimization Bending Region Objective: min : F ( x) Volumedesign Subject to: x X x R n | gi ( x) 0, i 1,..., m Where, g1 ( x) ( f design ) * g 2 ( x) b b * g3 ( x) manufacturing constraint (draw) g 4 ( x) manufacturing constraint (cyclic and symmetry) Design Non-design
- 14. Problem Setup – Topology Optimization • Optimization Responses • Volume – Design region • σb - Tooth bending stress • Optimization Constraint • σb < 35ksi • Optimization Objective • Minimize volume (design region) • Design Variable Constraints • σdesign < 20ksi • Cyclic constraint (number of instances = number of teeth) • Split draw constraint
- 15. Problem Setup – Initial Results N = 23 Cyclic instances = 23
- 16. Problem Setup – Initial Results N = 24 Cyclic instances = 24
- 17. Problem Setup – Further Refinement • Uniform gear web is desired for manufacturing and operation purposes • Increase number of instance = i * number of teeth, i = 1,2,3…..n • Number of instances for Baseline: 8 * number of teeth = 184 • Number of instances for Design 2: 4 * number of teeth = 96 • Eliminates holes through thickness
- 18. Results - Baseline N = 23 Cyclic instances = 184
- 19. Results – Design 2 N = 24 Cyclic instances = 96
- 20. Results Baseline Design 2 Baseline Design 2 Weight Before After % Difference Before After % Difference Gear Teeth 3.777 3.777 - 3.972 3.972 - Web 3.243 2.674 18% 3.691 2.731 26% Shaft 0.213 0.213 - 0.222 0.222 - Total 7.233 6.664 8% 7.885 6.925 12%
- 21. Conclusion • Topology Optimization was used to estimate material removal in the gear web. • For a gear designed to carry a specific torque, 12% weight saving was obtained. • An overdesigned gear does not translate to a lighter gear, post-topology optimization, based on preliminary studies. • Further work is required to understand feature manipulation, application for helical gears, and implementation of stiffness criteria.
- 22. Questions THANK YOU Sylvester Ashok sylvester_ashok@gatech.edu Fabian Zender fzender@gatech.edu