当前位置:首页 >> 数学 >>

Design Optimization Using HyperSizer


Design Optimization Using HyperSizer?
Craig Collier, Phil Yarrington, and Mark Pickenheim Collier Research Corporation 2101 Executive Dr. 6th floor Tower Box 72 Hampton, Virginia 23666 collier-research.com e-mail: info@collier-research.com

Abstract
This paper identifies an existing commercial solution that MSC users can benefit from for automated stress analysis and sizing. The HyperSizer? software is mathematically coupled with MSC/NASTRAN to provide an integrated solution for quick and accurate design optimization. Though specifically developed for the aerospace industry, the approach and methods apply to any industry. A reusable launch vehicle, which contains 7 assemblies, 21 optimization groups, and 203 structural components is used as an example. MSC/NASTRAN is used as the loads model and the entire plane is optimized using HyperSizer’s analysis methods that range from closed form, traditional hand calculations repeated every day in industry, to more advanced panel buckling algorithms. Margin-of-safety reporting for every possible failure provides the engineer with a powerful insight into the structural problem. The engineer is able to provide ‘real-world’ expertise in the optimization process by interacting with HyperSizer for designs on the fly.

? September 1998 by Collier Research & Development Corporation. Published with permission by The MacNeal-Schwendler Corporation for the 1998 MSC Americas Users’ Conference Proceedings.

Introduction
Aerospace vehicle internal load distributions are highly indeterminate and require FEA for solutions. This process is referred to in the aerospace industry as computing ‘running-loads’, publishing ‘internal loads’, or finding ‘load-paths.’ In essence, the integrated effects of flight surface pressures, temperatures, and accelerated inertia get reduced to force and moment components on panels and beams at all locations of the vehicle. In order to automate the analysis and optimization of structures, the HyperSizer structural sizing software uses the FEA computed panel and beam forces and moments for checking the many different types of failures that may occur within a structure. Some of these potential failures can be effectively predicted with traditional, hand methods. However, other failures require more rigorous methods. In general, physics based solutions are preferred over empirical or special case methods. HyperSizer is able to do discretely optimize in a manner, which guarantees structural integrity of the selected optimum design, using methods to accurately compute margins-of-safety for all potential failures. Optimization capabilities include finding minimum weight panel or beam concepts, material selections, cross sectional dimensions, thicknesses, and layups from a library of 40 different stiffened and sandwich designs and a database of composite, metallic, honeycomb, and foam materials.

About the Model
The model represents a NASA designed two-stage-to-orbit aerospace plane requiring accurate analysis capabilities to account for a complex thermo-mechanical environment. The integrated airframe/engine design contains a large volume of pressurized cryogenic fuel. Internal bulkheads serve as shape control members to maintain the vehicle's shape. The aeroshell is designed to be graphite/epoxy, hat-shaped stiffened panels. Though HyperSizer can analyze and optimize FEMs as large as one million DOFs, the choice was made to build a relatively small model of approximately 30,000 degree of freedoms (DOF) for the aerospace vehicle. This allows us to take advantage of HyperSizer's unique panel and beam stiffness formulations that achieve accuracy with coarsely meshed MSC/NASTRAN FEMs. More about this later.

2

How does HyperSizer Benefit the Aerospace Plane Design?
First, HyperSizer provides a complete and detailed analysis of the entire aircraft including concise margin-of-safety (MoS) summaries of all potential structural failure modes for all areas of the vehicle. This contour plot shows critical MoS of the aeroshell panels for all analyses performed. For instance, if local buckling of the facesheet has a lower MoS than panel buckling, then for that surface area, the MoS for local buckling is shown. This plot quickly indicates areas, which do not meet structural integrity requirements, or over-designed areas which can be made lighter. This entire vehicle, containing 7 assemblies, 21 optimization groups, and 203 structural components, is analyzed on a Pentium Workstation in two minutes. Second, HyperSizer optimizes all aspects of the aerospace plane structural design including: material selection (Gr/Ep vs. Al); panel and beam concepts (hat stiffened panel vs. honeycomb sandwich panel); and exact cross sectional dimensions (beam flange width of 1.24 vs 1.39). By optimizing all structural variables, HyperSizer will consistently reduce structural weight by 20% or more. Third, HyperSizer produces accurate structural dry weight predictions as shown in the summary tab here. It is important to accurately quantify dry weight of competing concepts early in the design process. Revolutionary designs such as the Joint Strike Fighter and Reusable Launch Vehicle typically do not have historical weights available.

3

Interaction between the engineer and the software is key to HyperSizer's design process Engineers learn within seconds the strengths and weaknesses of their structural designs from the software's interactive reporting of margins-of-safety. Interactive 3-D graphics provide visual inspection of the structural component layout, assemblies, and drawn to scale optimum panel and beam cross sections. These features are used on the aerospace plane to quickly interpret and understand design flaws. Critical design issues were identified and resolved early in the design process, allowing ample time to perform many design trade studies. This quick and highly interactive process makes the task of saving weight easy and fun. The figure below shows one of the interactive tools provided for display of HyperSizer computed data. Illustrated are the computed optimum panel unit weights on the assembly called ‘OML’.

4

How is HyperSizer used to Analyze and Optimize the Aerospace Plane?
Define Structural Components While in the FEM building process, modelers assign the same property data to a collection of elements. These collections of elements share the same PSHELL or PBAR record. HyperSizer uses these collections as structural components (components) for two primary purposes. The first purpose is to identify the smallest, practical manufacturable piece of hardware that can be sized independently. The second purpose is to be able to efficiently analyze structure with widely varying load distributions. Statistical methods are used to resolve peak loadings across structural components, and, by so doing, solve the difficult 'pulling-loads' problem that occurs for any automatic analysis procedure. In this way, analyses and optimizations are performed for the structural components, not for the finite elements of the model. The figures to the right illustrate identified structural components of the aerospace vehicle for the external aeroshell, internal panel shape control members, and the internal beam posts. The figures below represent groups comprised of any number of the components. Groups are used to assign optimization variables and bounds.

5

Identify materials, panel and beam concepts, dimension ranges and analysis methods With an infinite number of possible combinations, HyperSizer was able to make the optimization and analysis manageable by: Filtering materials HyperSizer's integrated database search engine containing metallic, composite, honeycomb and foam materials is used to filter a reduced set of materials.

6

Selecting design concepts Over 40 unique panel and beam concepts are provided such as hat and Z stiffened panels, honeycomb sandwiches, and I section beams. Fastened, bonded, and integrally machined fabrication details are included. Users simply select one or more concepts for their design.

7

Optimizing composite layups A library of over 1800 industry preferred layups makes composite optimization a snap. Layups are arranged into families based on best design practice guidelines and are displayed by a tree browser.

Selecting analysis methods HyperSizer includes over 100 potential failure modes that are interactively enabled or disabled for ultimate and limit loads. Setup Load Conditions HyperSizer couples tightly with MSC/NASTRAN to obtain 'design-to' running loads for the aerospace plane. Very accurate equivalent plate generalized stiffness terms are generated for the composite stiffened panels using exact cross sectional dimensions. Composite layups, temperature dependent properties, thermal gradients and the complex unsymmetric nature of panels leading to membrane-membrane coupling are accurately represented. Tight coupling with MSC/NASTRAN allows finite element properties and materials (PSHELL, PBAR, MAT2, MAT1) to be automatically generated and included in the FEA to obtain correct and consistent running loads.

8

The vehicle was analyzed and optimized to three distinct thermal environments: Mach 6 flight, Mach 3 flight, and takeoff/landing. In addition, six loading conditions were considered including aerodynamic pressure, thermal, landing and runway bump loads.

Visual tools provide convenient plotting of finite element loads. Any load, pressure, or temperature can be displayed. The integrated design of the interface automatically limits the view to the active component, group, or assembly. By moving through these different entities, it becomes quite easy to interpret loading magnitudes for specific vehicle areas.

9

Analyze and Optimize Designs on the Fly Optimization starts with accurate and comprehensive analysis The structure is analyzed using literally hundreds of strength and stability methods ranging from closed form, traditional hand calculations repeated every day in industry to more advanced panel buckling algorithms. Some of these methods are modern instability algorithms such as those used for unsymmetric panel buckling. Others are more traditional, simplistic hand calculations. Some of the potential failure modes are shown here for the honeycomb analysis.

Margin-of-safety reporting for every potential failure provides the engineer with a powerful insight into the structural problem.

All aspects of the structural design are optimized ? Panel and beam concepts ? Material selections ? Design dimensions, thicknesses and layups

10

?

Layups are even customizable to include odd angles and ply dropoffs using an integrated composite layup builder
P LY S E Q NUMB ER DE TA IL P LY NO. P LY O REN . WAR P/ FI LL (+ )/ ( -) PLY LAYUP TABLE: -005 WAR P DI RECT

MAKE FROM GR/BMI GR/BMI GR/BMI GR/BMI GR/BMI GR/BMI WOVEN TAPE TAPE TAPE TAPE WOVEN

2 3 4 6 7 8

P2 P3 P4 P6 P7 P2

+45? 0? 0? 0? 0? -45?

UP

DN

3 EQ SP TYP

TOOL SURFACE

14

14

SECTION R-R
PLY SCHEMATIC FOR -005 NO SCALE

E36

Design concurrently with multiple engineers HyperSizer includes a fully relational database management system which allows multiple users to work on the vehicle design concurrently across a local area network. In addition, multiple projects are stored in the same HyperSizer database meaning that archiving of the aerospace plane project and data are automatic. If temporarily pulled off of the aerospace plane project, you can come back weeks later and pick up right where you left off.

Conclusion
The commercially available HyperSizer? detailed analysis and sizing optimization program, which is integrated with MSC/NASTRAN, is described using an aerospace example. The example model is a reusable launch vehicle referred to as an aerospace plane. It contains 7 assemblies, 21 optimization groups, and 203 structural components. MSC/NASTRAN FEA is used for predicting internal loads. The entire plane is optimized for minimum weight with both composite and metallic materials. Structural integrity is ensured because of over 100 different failure analyses considered by HyperSizer that included strength, buckling, crippling, deformation, and frequency. Run times on a Pentium workstation ranged from two to ten minutes for the entire vehicle. The graphical display of analysis and design results is shown to provide the engineer with a powerful insight into the structural problem, and in so doing, allows ‘real-world’ expertise in the optimization process. The analytical methods and general approach of this integrated tool apply to MSC users in other industries.

11

References
1 HyperSizer User’s Manual, Book 1: Tutorial and Applications, Collier Research & Development Corp., Hampton, VA, September 1998 2 HyperSizer User’s Manual, Book 2: Analytical Method and Verification Examples, Collier Research & Development Corp., Hampton, VA, September 1998

Trademarks
HyperSizer? is a trademark of Collier Research and Development Corporation NASTRAN? is a registered trademark of NASA MSC/NASTRAN? is an enhanced proprietary product of The MacNeal-Schwendler Corporation

12


相关文章:
Hypersizer for blade design_图文.pdf
loadcases Companies Using HyperSizer for Lightest Composite Structure Designs ...Specializing in composite analyses and optimization, HyperSizer’s progressive ...
Parallel Optimization Strategies for MIC Heterogeneous ....pdf
With this design, the problem that the data scale of application is ...Comprehensive Performance Optimization To make full use of the MIC coprocessor...
Design of an Electron Gun Using Computer_图文.pdf
LEWIS et al.: DESIGN OF AN ELECTRON GUN USING COMPUTER OPTIMIZATION 1245 Fig. 3. Free variables for the optimal shape design problem. where is the ...
Design optimization using dynamic evaluation.pdf
Design optimization using dynamic evalua
...stell frames using Teaching Learning Based Optimization_....pdf
Design of planar stell frames using Teaching Learning Based Optimization_IT/计算机_专业资料。Design of planar stell frames using Teaching Learning Based ...
Organization Design Optimization Using Genetic Programming_....pdf
Organization Design Optimization Using Genetic Programming_专业资料。Abstract. This paper describes how we use Genetic Programming (GP) techniques to help ...
...State Machine Datapath Design, Optimization, and....pdf
Finite State Machine Datapath Design, Optimization, and Implementation_计算机...64 Figure 3.22: Datapath, FSM for Implementation using Table 3.17 ...
...Multidisciplinary Design Optimization using Desi....pdf
Aircraft Multidisciplinary Design Optimization using Design of Experiments Theory and Respo Design engineers often employ numerical optimization techniques to ...
...for Stimulation Design Optimization(脆性指数计算....pdf
A Practical Use of Shale Petrophysics for Stimulation Design Optimization(脆性指数计算_物理_自然科学_专业资料。这篇英文很好,介绍了页岩气相关的岩石力学参数的...
Practical 3D aerodynamic design and optimization using ....pdf
Practical 3D aerodynamic design and optimization using unstructured meshes_专业资料。A methodology for performing optimization on 3D unstructured grids based on ...
Optimization Method for Broadband Modem FIR Filter Design ....pdf
Optimization Method for Broadband Modem FIR Filter Design using Common Subexpression Elimin_专业资料。Abstract- An approach for broadband modem FIR filter ...
...of Curves and Surfaces Using Multi-Objective Optimization_....pdf
Design of Curves and Surfaces Using Multi-Objective Optimization_专业资料。Abstract. Design by optimization of curves and surfaces is a powerful design ...
A study of suspension design using optimization technique and....pdf
A study of suspension design using optimization technique and DOE_自然科学_专业资料。A study of suspension design using optimization technique and DOE ...
Conceptual Design Optimization of a Cis-Lunar Trans....pdf
Conceptual Design Optimization of a Cis-Lunar Transportation Architecture Using Genetic Alg Most of the costs of a space mission are defined at the ...
...design optimization of brushless PM motor using 3D EMCN_....pdf
Multiobjective design optimization of brushless PM motor using 3D EMCN_电力/水利_工程科技_专业资料。无刷永磁电机的等效磁网络建模及多目标优化论文 ...
...BASED DESIGN STUDY ON CONSTRAINED OPTIMIZATION USING THE ....pdf
A FEATURE-BASED DESIGN STUDY ON CONSTRAINED OPTIMIZATION USING THE SIMULATED ANNEALING ALGO_专业资料。In this paper, we study a design problem in which ...
Optimization of Mixed DiscreteContinuous Design Variable ....pdf
Optimization of Mixed DiscreteContinuous Design Variable Systems Using Neural Networks_专业资料。Emerging multidisciplinary design techniques must be capable of ...
2009.1_Optimization_of_Die_Design_for_Forging_of_a_Turbo-....pdf
2009.1_Optimization_of_Die_Design_for_Forging_of_a_Turbo-Charger_Impeller_and_a_Ring_Gear_Using_Proc_机械/仪表_工程科技_专业资料。锻造仿真软件simufact....
Design of fractional delay filters using convex optimization_....pdf
Design of fractional delay filters using convex optimization_信息与通信_工程科技_专业资料。Design of Fractional Delay Filters Using Convex Optimization WiZZiu...
lsopt_optimization_class.pdf
Design Optimization (RBDO) Incorporates Reliability and Robustness into design ...derivatives Spurious derivatives computed using small intervals are due to: ?...