Scribd
Upload
72 views5 pages

Multidisciplinary Design Optimization Approach and Its Application To Aerospace Engineering

The document discusses multidisciplinary design optimization (MDO) approaches and their application to aerospace engineering. It describes several MDO frameworks including all-at-once, multidisciplinary feasible, individual disciplinary feasible, concurrent subspace optimization, and bi-level integrated system synthesis. It also discusses the use of design of experiments, surrogate models, and handling of uncertainty in MDO approaches. The interaction between aerodynamics and structures is common in aerospace engineering and should be decoupled and solved using MDO for conceptual design.

Uploaded by

Saravanan R
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
72 views5 pages

Multidisciplinary Design Optimization Approach and Its Application To Aerospace Engineering

The document discusses multidisciplinary design optimization (MDO) approaches and their application to aerospace engineering. It describes several MDO frameworks including all-at-once, multidisciplinary feasible, individual disciplinary feasible, concurrent subspace optimization, and bi-level integrated system synthesis. It also discusses the use of design of experiments, surrogate models, and handling of uncertainty in MDO approaches. The interaction between aerodynamics and structures is common in aerospace engineering and should be decoupled and solved using MDO for conceptual design.

Uploaded by

Saravanan R
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
You are on page 1/ 5

Multidisciplinary design optimization approach and its

application to aerospace engineering


Zhenguo Wang • Wei Huang • Li Yan

DOI 10.1007/s11434-014-0671-1

Design of experiments
Design of experiments is one of the application methods in mathematical statistics, and it can
enable a highly efficient investigation where only the samples needed are collected and
analyzed. It is one of the crucial techniques for the high-fidelity optimization approach, and this
is induced by the increasing computational cost for the computational fluid dynamics (CFD) and
the computational structure mechanics (CSM). The design of experiments applied to the MDO
framework can reduce the experimental time compared to the conventional one. Recently, the
multi-fidelity strategy has been widely employed to cut the computational cost and improve the
optimization efficiency, namely the proper combination of the low-fidelity and high-fidelity
optimization approaches. A hierarchical multi-fidelity design approach to correct the
shortcomings of the low-fidelity models.

Surrogate model
Surrogate model has been widely used for design evaluation and optimization in the aerospace
engineering to reduce the computational cost in the MDO framework, especially for the high-
fidelity analysis process. The response surface models and the Kriging models have been
applied to the multidisciplinary design and analysis of an aerospike nozzle, and the optimization
problem of the aerospike nozzle consists of a computational fluid dynamics model and a finite
element analysis model. The response surface model can provide an explicit functional
representation of the sampled data, and its computational cost is lower than the Kriging model.
Recently, a double-stage surrogate model has been proposed to analyze and optimize the airfoil
and wing configurations, and this model has been constructed by integrating advantages of
both interpolation surrogate model and regression surrogate model. The obtained results show
that the double-stage model has higher fidelity, and its approximation error is lower than each
type of the surrogate model, a novel multi response and multistage metamodeling approach,
and it is more efficient than the one stage metamodeling and the sequential metamodeling.
Optimization approach
Recently, the uncertainty in the design variables has been considered in the MDO approach, as
well as the uncertainty in the disciplines, and its purpose is to avoid inefficient design and then
costly time over runs due to redesign. Two major kinds of uncertainty-based design problems,
namely robust design problems and reliability-based design problems, have been proposed,
and they both play an important role in the conceptual design in the aerospace engineering. A
reliability-based MDO approach to analyze and optimize a two-stage solid propellant
expendable launch vehicle, and the Latin hypercube sampling method has been employed to
choose the sample values for simulation runs. At the same time, the surrogate models coupled
with the multi-level approaches have been developed to reduce the computational effort while
handling high-fidelity discipline level simulation models, i.e., a surrogate model-based CO
approach.

Multidisciplinary design optimization framework


The MDO framework can be classified as the common and special ones. The common MDO
framework owns many advantages. This framework includes a graphical user interface, a state-
of-the-art probabilistic analysis engine, and approaches to rapidly integrate legacy and in-house
applications as well as third party commercial applications for multidisciplinary analysis. The
web services-based MDO framework can implement any type of the MDO strategy via the
process designer at any phase of the product development cycle, and it can be applied to a
broad range of engineering system design work. A novel framework for the flexible
computational design studies at the preliminary design stage has been developed, and it
combines a workflow management device with many advanced numerical treatments, namely
multi objective optimization, sensitivity analysis and uncertainty management. Recently, a
conceptual prototypical multi-fidelity, multi strategy and MDO environment, to optimize an air
vehicle wing, and it can optimize the configurations with a choice of the level of fidelity desired,
the choice of a single-step or multi-step optimization strategy and a combination of a series of
different structural and aerodynamic analyses.
Applications
All-at-once approach (AAO)

The AAO approach is the most basic MDO technique, and it is restricted to small design
problems. The AAO approach performs neither the system analysis nor the individual analysis
for each discipline. The AAO-based MDO approach, which incorporates sparse nonlinear
programming and surrogate models, has been utilized to analyze a horizontal takeoff SSTO
vehicle powered by airbreathing engines and rocket engines, and the engine–airframe
integration and its effect on rigid body characteristics such as trim and stability have been taken
into consideration. the AAO-based MDO framework and its application to a SSTO vehicle.
The relationships among the airframe– engine integration, rigid body characteristics and
payload transportation capability have been comprehensively explored.

Multidisciplinary feasible approach (MDF)


The MDF approach can be viewed as the most traditional MDO approach, and this is because
that an optimizer has been placed over a multidisciplinary analysis module [120]. However, its
cost is large for the system and subsystem analysis, and it cannot be applied to the large-scale
engineering application as well as the AAO method. The MDF approach has been employed to
optimize a suborbital reusable launch vehicle (SRLV), and seven disciplines have been taken into
consideration, namely geometry, aerodynamics, structure, propulsion, trajectory, aero-heating
and thermal conduction/TPS design. The calculation module for each discipline has been
integrated in the software framework FIPER/iSight-FD, and the takeoff weight has been
decreased by about 2.4 %, to analyze and optimize an aerospike nozzle as well, and its
advantage is obvious when comparing with the single discipline design strategy.

Individual disciplinary feasible approach (IDF)


The individual disciplinary feasible approach can be viewed as a decoupled version of MDF, and
it can utilize the mature discipline analysis code directly. Instead of enforcing multidisciplinary
feasibility at each design point, the IDF method only enforces discipline feasibility. Thus, its
organizing difficulty is reduced, and the complex multidiscipline iterative analysis is avoided.
However, the number of the optimization variables is large, and it has been rarely applied to
the aerospace engineering to the best of the authors’ knowledge.
Concurrent subspace optimization approach (CSSO)
The CSSO approach is one of the two primary architectures developed to provide a
collaborative multidisciplinary design environment through the use of distributed design
optimizations. In the CSSO method, each subspace optimization minimizes the system objective
function subject to its own constraints as well as constraints from the other subspaces
concurrently, with a system-level coordination optimization procedure whose goal is to update
the coupling parameters. The number of auxiliary design variables employed in CSSO only
depends on the number of disciplines, and this would substantially reduces the complexity of
the optimization problem within each disciplinary. A response surface-based CSSO algorithm
has been implemented within iMOODS to solve an aircraft concept sizing problem [125], and
three disciplines have been considered in the framework, namely aerodynamics, weight and
performance. In the CSSO approach, the design task has been divided into several disciplinary
related subsystems, and the subsystems are responsible for the satisfaction of the system
constraints.

Bi-level integrated system synthesis approach (BLISS)


The BLISS framework was originally developed by Prof. Sobieszczanski-Sobieski and his co-
workers at NASA, Langley. The first version of BLISS was developed in 1998 [139] with the most
recent derivative, BLISS-2000 [140], proposed in 2002. Brown and Olds [141] have analyzed and
optimized a reusable launch vehicle (see Fig. 18) by using the BLISS-2000 approach, and the
optimized results have been compared with those obtained by the AAO, CO and MCO
approaches. The obtained results show that the BLISS method may be the most promising
multi-level MDO technique available. Recently, the global optimization strategy and uncertainty
analysis strategy have been introduced into the BLISS-2000 framework, proposed as GBLISS-
2000 and UBLISS-2000 by Xu [143], and a bi-level system reliability of the multidisciplinary
design synthesis optimization (RBMDOSB) algorithm has been developed for largescale
uncertainty multidisciplinary systems

Summary and conclusions


The interaction between the aerodynamics and the structures is more common in the
aerospace engineering, and it should be decoupled and solved by the MDO approach in order
to carry out the conceptual design. At the same time, the wing of the aircraft is the simplest
configuration to verify the efficiency of the MDO approach, and it constructs the foundation for
the design of the vehicle.

The decision based on the optimization results is a crucial step in the overall design of the
aircraft, but the high dimensional design space would limit its application. Thus, the data mining
method is a very important supplement for the MDO process, and it can reduce the multi-
dimensional optimization results to two-dimensional ones and provide more information for
the designers. At the same time, the further relationship between the design variables and the
objective function can be explored by the data mining method. The MDO approach coupled
with the data mining techniques is a very promising research field for the aerospace
engineering, especially on the hypersonic air-breathing propulsion system.

You might also like