Rula M. Coroneos

COMPARATIVE EVALUATION OF DIFFERENT OPTIMIZATION ALGORITHMS FOR STRUCTURAL DESIGN APPLICATIONS

Non-linear programming algorithms play an important role in structural design optimization. Fortunately, several algorithms with computer codes are available. At NASA Lewis Research Centre, a project was initiated to assess the performance of eight different optimizers through the development of a computer code CometBoards. This paper summarizes the conclusions of that research. CometBoards was employed to solve sets of small, medium and large structural problems, using the eight different optimizers on a Cray-YMP8E/8128 computer. The reliability and efficiency of the optimizers were determined from the performance of these problems. For small problems, the performance of most of the optimizers could be considered adequate. For large problems, however, three optimizers (two sequential quadratic programming routines, DNCONG of IMSL and SQP of IDESIGN, along with Sequential Unconstrained Minimizations Technique SUMT) outperformed others. At optimum, most optimizers captured an identical number of active displacement and frequency constraints but the number of active stress constraints differed among the optimizers. This discrepancy can be attributed to singularity conditions in the optimization and the alleviation of this discrepancy can improve the efficiency of optimizers.

Structural optimization with approximate sensitivities

Computational efficiency in structural optimization can be enhanced if the intensive computations associated with the calculation of the sensitivities, that is, gradients of the behavior constraints, are reduced. Approximation to gradients of the behavior constraints that can be generated with a small amount of numerical calculations is proposed. Structural optimization with these approximate sensitivities produced correct optimum solution. Approximate gradients performed well for different nonlinear programming methods, such as the sequence of unconstrained minimization technique, method of feasible directions, sequence of quadratic programming and sequence of linear programming. Structural optimization with approximate gradients can reduce by one third the CPU time that would otherwise be required to solve the problem with explicit closed-form gradients. The proposed gradient approximation shows potential to reduce intensive computation that has been associated with traditional structural optimization.

A CASCADE OPTIMIZATION STRATEGY FOR SOLUTION OF DIFFICULT DESIGN PROBLEMS

A research project to evaluate comparatively ten different non-linear optimization algorithms was completed recently. A conclusion was that no single optimizer could successfully solve all the 40 structural design problems in the test-bed, even though most optimizers successfully solved at least one-third of the problems. We realized that improvements to search directions and step lengths, available in the ten optimizers compared, were not likely to alleviate the convergence difficulties. For the solution of those difficult problems we have devised an alternate approach called, the cascade optimization strategy. The strategy utilizes several optimizers, one followed by another in a specified sequence, to solve a problem. A pseudo-random dumping scheme perturbs the design variables between the optimizers. The cascade strategy has been tested out successfully in the design of supersonic and subsonic aircraft configurations and air breathing engines for high-speed civil transport applications. These problems could not be successfully solved by an individual optimizer. The cascade optimization strategy, however, generated feasible optimum solutions for both aircraft and engine problems. This paper presents the cascade strategy, solution of aircraft and engine problems along with discussions and conclusions.

Recent advances in the method of forces: integrated force method of structural analysis

Stress that can be induced in an elastic continuum can be determined directly through the simultaneous application of the equilibrium equations and the compatibility conditions. In the literature, this direct stress formulation is referred to as the integrated force method. This method, which uses forces as the primary unknowns, complements the popular equilibrium-based stiffness method, which considers displacements as the unknowns. The integrated force method produces accurate stress, displacement, and frequency results even for modest finite element models. This version of the force method should be developed as an alternative to the stiffness method because the latter method, which has been researched for the past several decades, may have entered its developmental plateau. Stress plays a primary role in the development of aerospace and other products, and its analysis is difficult. Therefore, it is advisable to use both methods to calculate stress and eliminate errors through comparison. This paper examines the role of the integrated force method in analysis, animation and design.

Compatibility conditions of structural mechanics

The theory of elasticity has camouflaged a glitch in the compatibility formulation since 1860. In structures the ad hoc compatibility conditions through virtual ‘cuts’ and closing ‘gaps’ is not parallel to the strain formulation in elasticity. The deficiency in the compatibility condition has prevented the development of direct stress determination method in structures and in elasticity. We have addressed and attempted to unify the theory of compatibility. This has lead to the development of the Integrated Force Method for structures and the Completed Beltrami–Michells Formulation in elasticity. The utilization of the compatibility conditions allows mapping of variables and facile movement among different structural analysis formulations. This paper reviews and illustrates the requirement of compatibility in structures and in elasticity. Addresses the generation of the conditions and quantifies the benefit of their use. The traditional analysis methods and available solutions should be verified for compliance of the compatibility conditions. Published in 2000 by John Wiley & Sons, Ltd.

DYNAMIC ANIMATION OF STRESS MODES VIA THE INTEGRATED FORCE METHOD OF STRUCTURAL ANALYSIS

Dynamic animation of stresses and displacements, which complement each other, can be a useful tool in the analysis and design of structural components. At the present time only displacement-mode animation is available through the popular stiffness formulation. This paper attempts to complete this valuable visualization tool by augmenting stress-mode animation to the existing art. The reformulated method of forces, which in the literature is known as the Integrated Force Method (IFM), became the analyser of choice for the development of stress-mode animation because stresses are the primary unknowns of its dynamic analysis. Animation of stresses and displacements, which have been developed successfully through the IFM analyzers, is illustrated in several examples along with a brief introduction to IFM dynamic analysis. The usefulness of animation in design optimization is illustrated considering the spacer structure component of the International Space Station as an example. An overview of the integrated force method analysis code (IFM/ANALYSERS) is provided in the appendix.

Lessons Learned During Solutions of Multidisciplinary Design Optimization Problems

Abstract Optimization research at NASA Glenn Research Center has addressed the design of structures, aircraft and airbreathing propulsion engines. During solution of the multidisciplinary problems several issues were encountered. This paper lists four issues and discusses the strategies adapted for their resolution. (a) The optimization process can lead to an inefficient local solution. This deficiency was encountered during design of an engine component. The limitation was overcome through an augmentation of animation into optimization. (b) Optimum solutions obtained were infeasible for aircraft and air-breathing propulsion engine problems. Alleviation of this deficiency required a cascading of multiple algorithms. (c) Profile optimization of a beam produced an irregular shape. Engineering intuition restored the regular shape for the beam. (d) The solution obtained for a cylindrical shell by a sub-problem strategy converged to a design that can be difficult to manufacture. Resolution of this issue remains a challenge. The issues and resolutions are illustrated through six problems: (1) design of an engine component, (2) synthesis of a subsonic aircraft, (3) operation optimization of a supersonic engine, (4) design of a wave-rotor-topping device, (5) profile optimization of a cantilever beam, and (6) design of a cylindrical shell. The combined effort of designers and researchers can bring the optimization method from academia to industry.

A Cascade Optimization Strategy for Solution of Difficult Multidisciplinary Design Problems

A research project to comparatively evaluate 10 nonlinear optimization algorithms was recently completed. A conclusion was that no single optimizer could successfully solve all 40 problems in the test bed, even though most optimizers successfully solved at least one-third of the problems. We realized that improved search directions and step lengths, available in the 10 optimizers compared, were not likely to alleviate the convergence difficulties. For the solution of those difficult problems we have devised an alternative approach called cascade optimization strategy. The cascade strategy uses several optimizers, one followed by another in a specified sequence, to solve a problem. A pseudorandom scheme perturbs design variables between the optimizers. The cascade strategy has been tested successfully in the design of supersonic and subsonic aircraft configurations and air-breathing engines for high-speed civil transport applications. These problems could not be successfully solved by an individual optimizer. The cascade optimization strategy, however, generated feasible optimum solutions for both aircraft and engine problems. This paper presents the cascade strategy and solutions to a number of these problems.

Reliability-Based Design Optimization of a Composite Airframe Component

*† ‡ A stochastic design optimization methodology (SDO) has been developed to design components of an airframe structure that can be made of metallic and composite materials. The design is obtained as a function of the risk level, or reliability, p. The design method treats uncertainties in load, strength, and material properties as distribution functions, which are defined with mean values and standard deviations. A design constraint or a failure mode is specified as a function of reliability p. Solution to stochastic optimization yields the weight of a structure as a function of reliability p. Optimum weight versus reliability p traced out an inverted-S-shaped graph. The center of the inverted-S graph corresponded to 50 percent (p = 0.5) probability of success. A heavy design with weight approaching infinity could be produced for a near-zero rate of failure that corresponds to unity for reliability p (or p = 1). Weight can be reduced to a small value for the most failureprone design with a reliability that approaches zero (p = 0). Reliability can be changed for different components of an airframe structure. For example, the landing gear can be designed for a very high reliability, whereas it can be reduced to a small extent for a raked wingtip. The SDO capability is obtained by combining three codes: (1) The MSC/Nastran code was the deterministic analysis tool, (2) The fast probabilistic integrator, or the FPI module of the NESSUS software, was the probabilistic calculator, and (3) NASA Glenn Research Center’s optimization testbed CometBoards became the optimizer. The SDO capability requires a finite element structural model, a material model, a load model, and a design model. The stochastic optimization concept is illustrated considering an academic example and a real-life raked wingtip structure of the Boeing 767–400 extended range airliner made of metallic and composite materials. ngineers have recognized the existence of uncertainty in material properties, in load, and in structural analysis as well as in design constraints. Consider for example the yield strength of a steel that is required to design a steel structure. Strength is measured in the laboratory from tests conducted on standard coupons. It is commonly observed that repeated tests yield different values for the strength of steel. The test data can be processed to obtain a nominal or mean value and a dispersion range or a standard deviation. The nominal strength along with a safety factor is used to define allowable strength for traditional deterministic design calculations. Alternatively, strength can be considered as a random variable with a mean value and a standard deviation. The experimental data can be processed to obtain a probability distribution function such as, for example, the commonly used normal distribution function that is defined by a mean value and a standard deviation. This concept for strength can be extended to Young’s modulus, Poisson’s ratio, density, and so forth, and a probabilistic material model can be generated. The procedure can be repeated and a probabilistic load model can be developed for mechanical, thermal and initial deformation loads. Likewise, a probabilistic design model can be developed for sizing variables like depth and thickness of a beam.

Subsonic Aircraft Design Optimization with Neural Network and Regression Approximators

The Flight-Optimization-System (FLOPS) code encountered difficulty in analyzing a subsonic aircraft. The limitation made the design optimization problematic. The deficiencies have been alleviated through use of neural network and regression approximations. The insight gained from using the approximators is discussed in this paper. The FLOPS code is reviewed. Analysis models are developed and validated for each approximator. The regression method appears to hug the data points, while the neural network approximation follows a mean path. For an analysis cycle, the approximate model required milliseconds of central processing unit (CPU) time versus seconds by the FLOPS code. Performance of the approximators was satisfactory for aircraft analysis. A design optimization capability has been created by coupling the derived analyzers to the optimization test bed CometBoards. The approximators were efficient reanalysis tools in the aircraft design optimization. Instability encountered in the FLOPS analyzer was eliminated. The convergence characteristics were improved for the design optimization. The CPU time required to calculate the optimum solution, measured in hours with the FLOPS code was reduced to minutes with the neural network approximation and to seconds with the regression method. Generation of the approximators required the manipulation of a very large quantity of data. Design sensitivity with respect to the bounds of aircraft constraints is easily generated.