A Paradigm Shift In Complex System Design

Enabling Technologies for Strategic Decision Making of

Advanced Design Concepts
By
Prof. Dimitri Mavris
Director

Aerospace Systems Design Laboratory

General Electric University Strategic Alliance
Boeing Professor in Advanced Aerospace Systems Analysis
School of Aerospace Engineering
Georgia Institute of Technology

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Quality Issues to be Addressed

Successful Utilization of Concurrent Engineering (CE) Approaches
U.S. Company
Japanese Company

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

+3
Months

Job #1

1-3
Months

14-17
Months

90%
Total Japanese
Changes Complete

20-24
Months

Number of Engineering Product

Changes Processed

by the Japanese Automotive Manufacturers

Motivation for PhysicsPhysics-based Conceptual Design

Subsonic Transports

Supersonic Aircraft
Rotorcraft

Uninhabited Air Vehicles

Personal Air Vehicles

New Generation of Vehicles can
not be modeled accurately in the
absence of historical data
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Extreme STOL

Traditional Development Process

Advanced Design
Requirements

Conceptual
Design
Optimization
Parametric
1st Level Analysis

General Arrangement/Performance
Representative Configurations
General Internal Layout

Conceptual
Baseline

Preliminary
Baseline
Sophisticated Analysis
Problem Decomposition
Multidisciplinary Optimization

Project Design

System Specifications
Detailed Subsystems
Internal Arrangements
Process Design

Allocated
Baseline

Detailed
Design

Problems with not foreseeing design flaws

Cannot rely on historical data
Communication between manufacturing and engineers is poor

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Production
Baseline

Production &
Support

Design Stages
Requirements Definition understanding the
requirements posed by the customer/market
Conceptual Design initial formulation,
interpretation based on experience/background
knowledge
Preliminary Design transforming the concept so
that the product will work and/or make money
Detailed Design testing and fine-tuning

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Uneven Distribution of Knowledge Effects

100%

100%

100%

Design
Freedom

1. Aerodynamics
2

3. Structures

4. Controls

5. Manufacturing

6
7

Conceptual

Preliminary

Time into the Design Process

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

2. Propulsion

Knowledge
about design

6. Supportability
7

Detailed

7. Cost

Traditional, PointPoint-Design Philosophy

May be characterized as a manual, deterministic, data driven, serial or
parallel, disciplinary-centric, point design process
Design requirements, and technology assumptions are usually fixed
and a design space exploration is performed around one or a handful
of concepts (point solutions)
As organizations strive to decrease costs and reduce operational
overhead, the number of personnel available for given activities is
decreasing
At the same time, the demands on the organization for more in depth
analysis at the conceptual and preliminary stages is increasing
As a result, a paradigm shift is required to reduce design cycle time,
allow for more iterations, and increase fidelity
Traditional organizations can be supported and enhanced by several
enabling technologies, to be presented here, that allow for this
transformation to take place in a practical fashion
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Acquisition Process
Short concept design phase with unequal distribution of
disciplines does not allow use of design freedom to
improve quality and integrate disciplines for
optimization
Uneven distribution of knowledge and effort
Need better representation of all disciplines in earlier stages
(conceptual, preliminary)

If data is in the historical database, it is pointless to use

physics based analysis
uses too many assumptions

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Phases in Acquisition Process

Pre-Milestone 0
Determination of Mission need and deficiencies

Phase 0
Concept exploration

Phase I
Program definition and risk evaluation

Phase II
Engineering and manufacturing development

Phase III
Production, development, and operations support
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Affordability - Making the Right Decisions Early

Cost Committed

$$
Pre-milestone
0
Determination of
Mission Need and
Identification
Deficiencies

Phase
0

Phase
I

Phase
II

Phase
III

Concept
Exploration

Program Definition
and Risk
Reduction

Engineering &
Manufacturing
Development

Production,
Deployment, and
Operational
Support

AoA I

AoA II

Milestone 0

Milestone I

Approval to
Conduct
Concept
Studies

Approval to
Begin a New
Acquisition
Program

AoA III
Milestone II

Milestone III

Approval to
Enter
Engineering
and
Manufacturing
Development

Actual Cost
Expenditure

Production or
Deployment
Approval

LRIP
Approval
Program Initiation

Decision-Makers Need New Methods

Decisionto Make the Right Trades !!
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Acquisition Timeline
Emphasis of Affordability Initiative

Capability based/Affordability Paradigm Shift

A paradigm shift is underway that challenges the manner in
which complex systems are being designed
Emphasis has shifted from design for performance to design
for affordability to design for overall capability
There is a need for a multidisciplinary approach to the problem
based on more sophisticated, higher fidelity tools
There is a need for forecasting the economic viability of a
system with a high probability of success
Long-term goal: Creation of a virtual engineering environment
for simulation-based acquisition
Academia is reacting to this paradigm shift and is trying to change its
own culture in an attempt to meet future research needs and take
advantage of new funding opportunities
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Definition of Affordability in Our Context

Affordability: The balance of benefits provided or gained from the system
to the cost of achieving those benefits. In a probabilistic, Modeling &
Simulation approach, Risk is inherent in these estimates.

S & T ROI

Weapon System
System Effectiven
Effectiveness
Weapon
ess
InvestmenttotoAchieve
AchieveThis
thisEffectiven
Effectiveness
Investment
ess

Weapon System Effectiveness- Aircraft Example

Cost
Survivability

Ope ration cost

Susce ptibility

Maintainability

Performance

Maintenance cost

Vulnerability

Inhe rent availability

Mane uv er ability

Air craft re place me nt

Re liability

Cre w re placement
training
RDT&E Cost

Logistics support

Satisfying mission
re quir eme nts

Readiness

Capability

Le thality

CDF

. . .

Dependability
Safe ty
Re liability
Maintenance
de fe cts
De sign de fe cts
Oper ations

Effectiveness = k1(Capability)+ k2(Survivability)+ k3(Readiness)+ k4(Dependability)

+ k5(Life Cycle Cost)

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

. . .

The Affordability
Balance

Operational Effectiveness

Acquisition cost

CDF

Survivability
Maneuverability
Availability Safety
Capability

RDTE Cost
O & S Cost
Acquisition Cost

Physics--based Conceptual Design - A Paradigm Shift

Physics
Acquisition Timeline
Pre-milestone 0

Phase 0

Phase I

Phase II

Phase III

Determination of
Mission Need and
Deficiencies

Concept
Exploration

Program Definition
and Risk
Reduction

Engineering &
Manufacturing
Development

Production,
Deployment, and
Operation Support

100 %
Design Freedom

Knowledge
becomes available
when time to make
decision

Cost Committed
Today
Future

Knowledge

0%

Requirements
Definition

Conceptual
Design

Preliminary
Design

Detail Design + Manufacturing

Design Timeline
Mavris, D.N., DeLaurentis, D.A., Bandte, O., Hale, M.A., "A Stochastic Approach to Multi-disciplinary Aircraft Analysis and Design", AIAA 98-0912.
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

A Basis from Which to Begin:

Generic IPPD DecisionDecision-Making Process
COMPUTER-INTEGRATED ENVIRONMENT
QUALITY
ENGINEERING METHODS

PROCESS DESIGN DRIVEN

ESTABLISH
THE NEED

DEFINE THE PROBLEM

SYSTEMS
ENGINEERING METHODS
REQUIREMENTS
& FUNCTIONAL
ANALYSIS

SYSTEM DECOMPOSITION
&
FUNCTIONAL ALLOCATION

ESTABLISH
VALUE

ROBUST DESIGN
ASSESSMENT &
OPTIMIZATION

GENERATE FEASIBLE
ALTERNATIVES

SYSTEM SYNTHESIS
THROUGH MDO

PRODUCT DESIGN DRIVEN

7 M&P TOOLS AND

QUALITY FUNCTION
DEPLOYMENT (QFD)

TOP-DOWN DESIGN
DECISION SUPPORT PROCESS

EVALUATE
ALTERNATIVE
ON-LINE QUALITY
ENGINEERING &
STATISTICAL
PROCESS

MAKE DECISION

SYSTEM ANALYSIS
&
CONTROL

Schrage, D.P., Mavris, D.N., "Technology for Affordability - How to Define, Measure, Evaluate, and Implement It?",
50th National Forum of the American Helicopter Society, Washington, D.C., May 11-13, 1994.
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

What is needed for the Paradigm Shift to occur?

occur?
Transition from single-discipline to multi-disciplinary analysis,
design and optimization
Automation of the resultant integrated design process
Transition from a reliance on historical data to physics-based
formulations, especially true for unconventional concepts
Means to perform requirements exploration, technology infusion
trade-offs and concept down selections during the early design
phases (conceptual design) using physics-based methods
Methods which will allow us to move from deterministic, serial,
single-point designs to dynamic parametric trade environments
Incorporation of probabilistic methods to quantify, assess risk
Transition from single-objective to multi-objective optimization
Need to speed up computation to allow for the inclusion of
variable fidelity tools so as to improve accuracy
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Elements needed to enable this Paradigm Shift

Advances in MDA/MDO methods and techniques to encompass the
holistic nature of the problem, emphasis on uncertainty associated with the
early design phases
Creation of computational architecture frameworks to allow for easy
integration and automation of sometimes organizationally dispersed tools
Emergence of commercially available frameworks will further expedite
the usefulness of the proposed approaches
Creation of physics-based approximation models (surrogate or metamodels) to replace the higher fidelity tools which are usually described as
too slow for use in the design process, cryptic in their use of inputs,
interfaces and logic, and non-transparent (lack of proper documentation,
legacy)
Use of probability theory in conjunction with these meta-models will
enable us to quantify, assess risk and to explore huge combinatorial spaces
In fact it will enable us to uncover trends, solutions never before examined
in a very transparent, visual, interactive manner
Use of Multi-attribute decision making techniques, pareto optimality,
genetic algorithms to account for multiple, conflicting objectives and for
discrete settings
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Varying Fidelity M&S Initiative

Safety
Safety

Aerodynamics
Aerodynamics

Ge ometry

Economics
Economics

Synthesis & Sizing

Structures

Mission

Structures

Manufacturing

Integrated Routines
Table Lookup
S&C

Approximating Functions
Direct Coupling of Analyses

Conceptual Design Tools

Increasing
Sophistication and
Complexity

(First-Order Methods)
Propulsion

S&C

Performance

Preliminary Design Tools

(Higher-Order Methods)

Propulsion

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Performance

Manufacturing

Key Enabler Surrogate Models

Reliance on meta-models or surrogate models as a means to:

speed up processes,
protect proprietary nature of codes used,
overcome organizational barriers (protectionism of tools and data),
allow for the framework to be tool independent (no need for direct
integrations of codes; also enables our desire for variable tool fidelity
formulations),
allow the designer to perform requirements exploration, technology
infusion trade-offs, and concept down selections during the early design
phases (conceptual design) using physics-based methods

Surrogate models can also be used at the integrated system level

to determine responses at that level. This will allow us to move
from deterministic, serial, single-point designs to dynamic
parametric trade environments.
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Response Surface Methodology (RSM)

RSM is a multivariate regression technique developed to model the
response of a complex system using a simplified equation
Regression data is obtained intelligently through the Design of
Experiments (DoE) techniques
RSM is based on the design of experiments methodology which gives
the maximum power for a given amount of experimental effort
Typically, the response is modeled using a second-order quadratic
equation of the form:
R

R bo

k
i 1

bi xi

k
i 1

bii xi2

k 1

bij xi x j

i 1 j i 1

Where,
bi are regression coefficients for the first degree terms
bii are coefficients for the pure quadratic terms
bij are the coefficients for the cross-product terms
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Design of Experiments
Design of
Experiments
Full Factorial
Central
Composite
Box-Behnken
D-Optimal
Design

Run
1
2
3
4
5
6
7
8
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

For 7
Variables
2,187
143

For 12
Variables
531,441
4,121

Equation

62
36

2,187
91

(n+1)(n+2)/2

X1
-1
+1
-1
+1
-1
+1
-1
+1

Factors
X2
-1
-1
+1
+1
-1
-1
+1
+1

X3
-1
-1
-1
-1
+1
+1
+1
+1

3n
2 n+2n+1

Response
y1
y2
y3
y4
y5
y6
y7
y8

Enabling Tools and Techniques

Established Techniques

Response Surface Method (Biology; Ops Research)

Design of Experiments (Agriculture, Manuf.)
Quality Function Deployment, Pugh Diagram (Automotive)
Morphological Matrix (Forecasting)
MADM techniques (U.S Army, DoD)
Uncertainty/Risk Analysis (Control Theory; Finance)
Technology Readiness Levels (NASA)

ASDL Innovation

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Feasibility/Viability Identification
Robust Design Simulation (RDS)
Technology Identification, Evaluation, Selection (TIES)
Joint Probabilistic Decision Making (JPDM)
Unified Trade-off Environment (UTE)
Virtual Integrated Stochastic System Technology
Assessment (VISSTA)

Point Design Identifies a Single, Feasible Design

A point design is a single point on the thrust/wing area plot

This point will not satisfy evolving mission requirements
A parametric design environment would allow movement
around this space
Constraints could also be changed in real time and the
impact on the design could be assessed

Point Design for

A notional Concept

21420

TOWOD

Vapp
Ps

Alternate
Range

Turn Radius
Turn Rate
TOGW

Thrust
(lbs.)

O&S
14535
380
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Area (ft^2)

520

Integrated Design: Reduction in Cycle Time Through Automation

Performing an integrated design involves linking conceptual and
preliminary design tools in a computational environment that
automatically passes information between design codes
Enablers:
Computational environment such as ModelCenter or iSIGHT
Design codes with simple inputs/outputs without hard coding of design
variables or internal optimizations that may skew results

Integrated design provides tremendous advantages in design

cycle time by eliminating the re-keying of information from
output files to input files.
The next slide shows a missile design environment. As an
integrated suite of codes, it takes 35 seconds to perform a
design. If the codes were not linked, it would take approximately
45 minutes to pass the information back and forth and check for
errors!
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Example: Integrated Missile Design Tool in the ModelCenter Environment

Linked Computer Codes

Weights/
Sizing
Propulsion
Aero
Trajectory
Plume
OPS

Cost
Reliab/
Safety

Design Variables

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Varying Fidelity M&S Initiative

Safety
Safety

Aerodynamics
Aerodynamics

Ge ometry

Economics
Economics

Synthesis & Sizing

Structures

Mission

Structures

Manufacturing

Integrated Routines
Table Lookup
S&C

Approximating Functions
Direct Coupling of Analyses

Conceptual Design Tools

Increasing
Sophistication and
Complexity

(First-Order Methods)
Propulsion

S&C

Performance

Preliminary Design Tools

(Higher-Order Methods)

Propulsion

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Performance

Manufacturing

Physics--Based Modeling and Simulation Environment

Physics
Robust Design Simulation
Subject to

Technology
Infusion
PhysicsBased
Modeling
Activity and
ProcessBased
Modeling

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Decision Making
(MADM)

Design & Environmental

Constraints

Objectives:
Synthesis
& Sizing

Simulation

Operational
Environment

Economic
Life-Cycle
Analysis

Attribute 1
(e.g. Cost)

Attribute 2
(e.g. Performance)

Attribute 3
...

Economic &
Discipline
Uncertainties

Impact of New
TechnologiesPerformance &
Schedule Risk

Customer
Satisfaction

VIRTUAL INTEGRATED STOCHASTIC SYSTEM AND TECHNOLOGY ASSESSMENT (VISSTA) ENVIRONMENT

Computer Integrated Environment

Environmental,
Operational
Maintenance
Model

Integration Methodology
Simplified Analysis
Historical-Based

Variability

Proposed
Module Integration
Physics-based
Simulation

Transparent,
Seamless
Integration

Geometry
Solid Mechanics
Fluid Mechanics

Parametric Definition
Probabilistic FEM
Virtual Wind Tunnel

Stability &
Controls

Flight Simulation

Safety

Virtual Operation
Environment

Propulsion
Subsystems

VIPER-CAT

Integration Environment
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Process

Constrained
Probabilistic
Optimization

Sizing
Synthesis

Product (Physics-Based)

Activity-Based Costing
Process-Based Models

Neural Networks
Fuzzy Logic Response Surfaces
Knowledge-Based Systems
Agents Expert Systems

Economics

Uncertainty

Comprehensive Life
Cycle Customer
Requirements

Risk/Benefit
Analysis

Systems Engineering Methods

Virtual Manufacturing

Probability
Fuzzy Logic
Distributions

Constraints

Simulation
Environment

Reduced Variability

Manufacturing
Re- Manufacturing

Technology
Readiness/
Risk Library

Probabilistic
Assessment

Current
Module Integration

Quality Engineering Methods

System Level Objectives

Fidelity
Uncertainty

Product Family
Design,
Enterprise Design

Design Guidance
Knowledge
Based System

Decision Support

Decision Making
Processes

Numerical
Optimization
(MDO)

Why Do We Need a CapabilityCapability-Based Design Approach?

Noting that schedule slips have become ubiquitous in the acquisition of
complex systems, the Air Force is pursuing techniques which will facilitate
accelerated acquisition (also known as agile acquisition.)
The paradigm shift in systems design advocates moving knowledge forward.
We now want to move the ability to examine capabilities to the conceptual
design phase
Assists future military planners
Identifies solutions which may be non-optimal in and of themselves, but
maximize a macro-level performance function
Improve interoperability of weapons systems and platforms through more
rigorous interoperability evaluation in a replicated battlefield environment
Identify technologies for systems and subsystems in the presence of changing
requirements and evolving threats

Facilitate Shift to Capability-Based Acquisition and Planning

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Enabling CapabilityCapability-Based Design

There is an overall desire to select systems and architectures based on their
overall capability
Because these architectures rely on multiple, interoperable, heterogeneous
systems, an integrated design environment is needed
Collaboration is required because an architecture is comprised of different
elements belonging to various entities
To perform trade studies between requirements, design criteria, and technologies,
rapid parametric analysis capabilities are needed

Collaboration is hindered by competition and intellectual property issues

Surrogate Models may be viewed as an enabler for capability-based design
If processes can be sped up to the point where they are not a computational
burden, the mapping of capabilities to candidate designs is trivial
An integrated, parametric modeling and simulation environment facilitates
bottom-up trade studies

Probabilistics, coupled with surrogate models, enables large-scale design

studies where top-level capabilities can be mapped to systems and any
variable can be treated as an independent variable
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Collaborative Design Aided by Surrogate Models

IT issues and intellectual property concerns frequently limit collaborative activities

Surrogate models can be traded as a currency for exchanging information
Generated using the tools specific to a collaborative partner
Proprietary concerns are mitigated since the surrogates are made for a specific problem
(cannot be reverse engineered)
Brings the disciplinary experts into the conceptual design process as they generate the
surrogates
Equations are not operating system or platform-specific
Shields Intellectual Property
Provide intelligence to assets in an agent-based framework

Multi-Site Collaboration

Integrated Suite of Tools

Surrogate Models
www.engineous.com

www.phoenix-int.com

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Realizing the Vision for CapabilityCapability-Based Design

Strategic Challenges
Scenario Modeling Assumptions

Physical Assumptions

Political Climate

Doctrine

Friend or Foe, no-fly zones

Guiding principles for actions

Geography
Range where?

Basing Options
Asset Allocation

Specify tactics
Execution of actions to fulfill doctrine

System-of-Systems Level

Establish heuristics, behaviors,

and actions for assets

Campaign/Theater/
Engagement Analysis

Map heuristics, behaviors, and

actions to the environment

Deployment Status

System Level

Propulsion, Avionics, Structures

Technologies and Design Variables

Campaign Level
Warfighter View

MoEs become MoPs

Technologies

MoPs

Responses

Metrics/Objectives

Design Vars

Subsystem Level (Avionics)

Design Vars

Technologies

Technologies

MoPs

MoPs

Responses

MoPs

Capability Options

System Level (Missile)

Requirements

Requirements

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Engagement Model

Constraints

Requirements
Asset Families

Mission Level
MoPs of vehicle become variables for next level

System Level (Platform)

Requirements

Subsystem Level (Sensors)

Requirements

Design Vars

Technologies

Design Vars

MoPs

(force/force or one/one)

System-of-Systems-Level

Metrics/Objectives

Simulation method

Subsystem Level

Constraints

Scenarios/Missions/Threats

MoEs

Design Team

Weapons and Platforms

Many Heterogeneous Assets Interoperating

MoEs

Assumptions

Technologies

Subsystem Level (Propulsion)

Requirements

Design Vars

Technologies

Capability Based Design - System of Systems Affordability

National
Security

National
Level

Economic
Security

Requirements
Doctrine
Missions

Campaign
Level

Needs
Probabilistic
Matching

Asset
Level

System
Effectiveness

Attributes

Dependability

Survivability

Capability

Lethality

Total Own. Cost

Systems

maneuverability

speed

payload

$ RDTE

$ O&S

range

susceptibility

Capabilities
Discipline
Level
Technologies
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

L/D

SFC

EW

IR/RCS

DOC/Sortie
S&T $

Surrogate Modeling Enables MultiMulti-Level Trade Studies

MoEs

System-of-Systems-Level

MoEs become MoPs

MoEs

Mission Level
Engagement Model

Requirements

Design Vars

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Technologies

System Level (Platform)

Requirements

Subsystem Level (Sensors)

Requirements

Design Vars

Technologies

Design Vars

MoPs

Subsystem Level (Avionics)

Technologies

MoPs

Design Vars

Metrics/Objectives

MoPs

Responses

MoPs of vehicle become variables for next level

Constraints

Metrics/Objectives

Warfighter View

System Level (Missile)

Requirements

MoPs

Campaign Level

Constraints

MoPs

Responses

Enabler to perform
trades between
dissimilar systems
(eg: satellites vs.
stealth UAVs) with
MoEs at multiple
levels

Environment
allows flow-up
and flow-down

Technologies

Subsystem Level (Propulsion)

Requirements

Design Vars

Technologies

Parametric Design: Using an Integrated Design on a LargeLarge-Scale

The integrated design environment is an enabler for a parametric
design study
Instead of passing in a series of input variables, a parametric
design can take a distribution of inputs.
In this manner, an entire design space can be explored, rather
than small perturbations around a single point design
Large design spaces may take too long to explore by traditional
means
The integrated design environment above can be used to generate
metamodels of the design process
These metamodels, custom made for a given range of inputs, can be
evaluated in a spreadsheet hundreds of times per second
Metamodels represent another order of magnitude in reduction for design
cycle time.
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

1
Define
the
Problem

Define
Concept
Space

Modeling
and
Simulation

Investigate
Design
Space

Identify
Technologies

Feasible
or
Viable?

Problem Definition:

Evaluate
Technologies

Select
Technologies

HSCT concept

Societal Need:
Next generation supersonic aircraft
Increased commercial traffic growth
Increased comfort, safety, and affordability

Potential concept:
High Speed Civil Transport*
6. Descent

5. Cruise

67,000 ft.

M=2.4

9. Reserve
M=0.6

50,000 ft.

4. Climb
3. Cruise
35,000 ft.

M=0.9

10. Land
F.L.= 11,000 ft.

7. Loiter
M=0.6

2. Climb
1. Taxi & T.O.

* Potential concept is actually

established in the following step

8. Abort
3000 ft.

F.L.=11,000 ft.

50 nm

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

750 nm

5,000 nm

100 nm

200 nm

1
Define
the
Problem

Define
Concept
Space

Modeling
and
Simulation

Investigate
Design
Space

5
Feasible
or
Viable?

6
Identify
Technologies

7
Evaluate
Technologies

Define Concept Space:

8
Select
Technologies

Morphological Matrix

Purpose: Establish the concept space that may fulfill the customer requirements and establish a
datum point for the feasibility investigation
Performed with the aid of the Morphological Matrix technique
Procedure:

Alternatives

Functionally decompose the existing

system into contributing
characteristics
For each characteristic, list all the
possible ways in which it might be
satisfied
Select a datum point; permutations
are concept alternatives

Define Design Space

Further decompose the system from
the Alternatives Space to elementary
attributes, such as geometric and
propulsive characteristics

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Struct Aero Propulsion Mission Config

Define Alternatives Space

Characteristics
Vehicle
Fuselage
Pilot Visibility
Range (nmi)
Passengers
Mach Number
Type
Materials
Combustor
Nozzle
Low Speed
High Speed
Materials
Process

3
4
Wing,
Tail
&
Wing & Tail Wing & Canard
Wing
Canard
Cylindrical
Area Ruled
Oval
Synthetic Vision Conventional Conventional &
Nose Droop
5000
6000
6500
250
300
320
2
2.2
2.4
2.7
Mid
Tandem
MFTF
Turbine Bypass
Flade
Fan
Conventional
High T Comp
Conventional
RQL
LPP
Internal
Conventional
Mixed Ejector Mixer Ejector &
Acoustic Liner
Flow Alteration
Conventional Conventional
CC
Flaps
Flaps & Slots
Active Control
Conventional
NLFC
HLFC
High Temp.
Aluminum
Titanium
Composite
Spanwise
Integrally
Monocoque
Hybrid
Stiffened
Stiffened

Example of a Parametric Design Exercise for a Supersonic Business Jet

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Example of a Parametric Design Exercise for a Supersonic Business Jet

Each aircraft to the left is an

example of a complete design.
Parametric design provides the
user with the power to test
hundreds or thousands of designs,
where previously, time permitted
a single design point only.
Each aircraft to the left has
A complete analysis of the
propulsion system
An aerodynamic analysis to
calculate accurate drag polars

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

They have all been sized for the

mission requirements, which are
ALSO parametrically scalable. A
change in desired range will regenerate this matrix of designs.
The creation of a single one of
these aircraft designs can take less
than a minute or up to a day,
depending on the desired fidelity
of the design tools.

Man in the loop Genetic Algorithm

Sonic Boom Profiles for Various SBJ Configurations

Conventional Baseline
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Swept Configuration

Highly Swept Configuration

w/ Long VTail

Unconventional Joined
Wing Design

1
Define
the
Problem

Define
Concept
Space

Modeling
and
Simulation

Investigate
Design
Space

5
Feasible
or
Viable?

6
Identify
Technologies

7
Evaluate
Technologies

Note: The geometric and

propulsive parameters may
vary in the ranges defined
with the same likelihood
since at the outset, there
should be no preference of
values. Hence, uniform
distributions are assigned
to each parameter.

X2,Y2

Define Design Space

Variable

X4
X3

Define Concept Space:

8
Select
Technologies

X5

X6

SW
TWR
TIT
FPR
OPR
CLdes
X2
X3
X4
X5
X6
Y2
t/c_root
t/c_tip
SHref
SVref

Minimum

Maximum

Units

7500
0.29
3000
3.5
18
0.08
1.54
2.1
2.4
2.19
2.18
0.44
3
2
400
350

9000
0.33
3400
4.5
21
0.12
1.69
2.36
2.58
2.37
2.5
0.58
5
4
700
550

ft
~
o
R
~
~
~
~
~
~
~
~
~
%
%
2
ft
2
ft

Description
Wing area
Thrust-to-weight ratio
Turbine Inlet Temperature
Fan Pressure Ratio
Overall Pressure Ratio
Design lift coefficient
LE kink x-location*
LE tip x-location*
TE tip x-location*
TE kink x-location*
TE root x-location*
LE kink y-location*
Wing root t/c ratio
Wing tip t/c ratio
Horizontal Tail area
Vertical Tail area

* Variables Nondimensionalized by wing semi-span

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Parametric Description of a Wing Planform

Xwing

Other Design Variables

for the Aerodynamic Screening
xwing
t/c at root
t/c at tip
Nacelle Scaling
Horizontal Tail Area
CL Design
Root Airfoil (loc. max. thickness)
Tip Airfoil (loc. max. thickness)
Nacelle X-location
Wing Reference Area

0, 0

Planform Variables
(Normalized by Span)

(X1, Y1)

naY2 naY1
X2

X3
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

X-axis

(X4, Y1)

X5, 0

Y-axis

Possible Wing Planforms

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Parametric Technology Space:

Family of Designs

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

3
Modeling
and
Simulation

4
Investigate
Design
Space

5
Feasible
or
Viable?

6
Identify
Technologies

7
Evaluate
Technologies

Modeling and Simulation:

8
Select
Technologies

Vehicle Modeling

M&S environment:
Relates responses to inputs via a physics-based
M&S environment
Metamodels are employed to facilitate the use of
higher-fidelity analysis for unconventional
configurations

Design

or

Tech. (k)

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

C
L

DoE
Output Responses

Response Data

Response = f (design variables), or

= f (technology k factors)

Aero RSEs=f(design)

CD

FLOPS/ALCCA

Variables & Distributions

ALCCA (Aircraft Life-Cycle Cost Analysis): Developed by

NASA-Ames and enhanced by ASDL; calculates life-cycle costs
and airline economics for transport aircraft.
FLOPS (Flight Optimization System): A NASA-Langley
vehicle synthesis and sizing code, well-suited for the conceptual
and preliminary design of subsonic transport aircraft.

FLOPS/
ALCCA

Define
Concept
Space

Input Variables

1
Define
the
Problem

Creation of Modeling and Simulation Environment

Fidelity
Multipliers

Economic
Assumptions
Mission
Requirements

Market
Requirements

Airframe Fixed

Thrust
Available
Engine
Architectures
A

Technology
Setting

Multi-Disciplinary DOE

Given Engine Architecture

NEPP

FLOPS

Engine Performance
Program

Flight Optimization
Code
Vehicle Size

WATE

ALCCA

Weight Analysis of
Turbine Engines Code

Aircraft Life Cycle

Cost Analysis Code

Emissions
Modules

Vehicle
Performance
Thrust
Required

Vehicle
Economics

Aircraft Needs
NOx
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

CO2

NOISE

ASDL Probabilistic Methods Process

FPI

CDF
Customer Driven Requirements
(Concept & Technology Set Specific)

100%

Aero

Architecture B

Aspiration Space
Technology
Insertion
Impact

x
x

x
x

0%

Structures

Architecture A

Thrust

Probability

DISCIPLINARY RSEs

Objective

Design Point

x
Physics
Driven
Growth

x
Growth Spurs

Engine Weight

Competitive Assessment
Strategic Decision Making

Weights
Requirements
Space

RSEs

Dynamic
Contour
Plots

Constraints
%$/RPM

TOFLmod
SLNmod

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Technology
Space
Metrics/Objectives

Responses

Metrics/Objectives

Responses

Concept Space

Constraints

Metrics/Objectives
Constraints

SYNTHESIS & SIZING

Responses

Etc.

Viewing RSEsRSEs- Prediction Profiles

Prediction Profile: This displays prediction traces for each X variable. A prediction trace is the
predicted response as one variable is changed while the others are held constant at the current
values. The Prediction Profile can recompute the traces as you vary the value of an X variable.*
Calculated Value

Prediction Trace

Uses of Prediction Profile

Responses

Hairlines

1) Debugging: Review each sensitivity,

checking for those that dont make intuitive
sense: investigate
2) Fidelity: Adjusting the regressor variables
to investigate the strength of their impact on
responses

AREA

Input Value
Variable Limits
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Regressor Variables

3) Life/Technology: Model the impact of

new technologies (or the degradation of
current systems) by using metric-factors as
regressors.

Dynamic Interactive Design Space TradeTrade-off Environment for an SST

Metric
Responses

Upper/Lower bounds of
the design space

Influence of parameter on response

(either or or no influence)
The larger the slope, the greater the influence

Hairlines move and update

responses in real-time
Optimization can be performed

1450671

TOGW 837264.1

T/W

TIT

X2

X4

Lower bound values of SLN and $/RPM

indicates NO feasible space
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

X5

X6

Y2

t/c
root

t/c
tip

HT
Area

-1 min value of Y2
0 current value of Y2
1 max value of Y2
All are in a non-dimensional space

-1

1
-1

-1

-1

-1

-1

1
-1

-1

-1

-1

-1

X3

-1

-1

CL
design

-1

OPR

-1

FPR

-1

-1

0.24

-1

-1

Wing
Area

-0.33

Design
Variables

-0.3301

-1

765601.0
18545
TOFL 10327.29
8980
14787
LDGFL 9047.182
8765
210.4
Vapp 155.0316
150.4
118.09
FON 107.3951
103.9756
120.43
SLN 110.1838
109.5424
0.17643
$/RPM 0.113477
0.10590

VT
Area

Pareto Analysis of Significant Technology Metrics

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Parametric Dynamic Requirements Exploration

Response
contours may
be set here

Horiz Vert Factor

Radius
ULF
CmbMach
DPayld
Thrust
Area
DStealth
Auxtnk
SFC
Response
O&S
TOGW
LDWOD
TOWOD
Vapp
Ps
AltRng

Current X
Grid Density Update Mode
-0.8888
20 x 20
Immediate
0
0
-1
-0.8888
-0.857
-1
-1
1
Contour Current Y
Lo Limit
Hi Limit
5500 5031.4066
?
5500
40000
37137.19
?
40000
15 4.8129793
?
15
0 -20.93512
?
0
153 152.42056
?
153
695 700.21755
695
?

21420

Slide bars control variable values

Constraints are set here

Vapp

TOWOD

Thrust
(lbs.)

LDWOD

Ps
TOGW

14535
360

Area (ft^2)

520

White area indicates available design space. Filled regions indicate areas which violate set constraints
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Parametric Dynamic Requirements Exploration

As the combat
radius slide bar
is moved to the
right, the design
space shrinks to
reflect the new
constraint.

Horiz Vert Factor

Radius
ULF
CmbMach
DPayld
Thrust
Area
DStealth
Auxtnk
SFC
Response
O&S
TOGW
LDWOD
TOWOD
Vapp
Ps
AltRng

Current X
Grid Density Update Mode
-0.4232
20 x 20
Immediate
0
0
-1
-0.8888
-0.857
-1
-1
1
Contour Current Y
Lo Limit
Hi Limit
5500 5031.4066
?
5500
40000
37137.19
?
40000
15 4.8129793
?
15
0 -20.93512
?
0
153 152.42056
?
153
695 700.21755
695
?

With a parametric design,

the slide bars to the left
are now enabled to give
the designer freedom to
change the space, moving
both the design point
AND the contraints

21420
Vapp

TOWOD

Thrust
(lbs.)

LDWOD

Ps

TOGW

14535
360

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Area (ft^2)

520

Parametric Dynamic Requirements Exploration

Horiz Vert Factor
Radius
ULF
CmbMach
DPayld
Thrust
Area
DStealth
Auxtnk
SFC
Response
O&S
TOGW
LDWOD
TOWOD
Vapp
Ps
AltRng

Current X
Grid Density Update Mode
0.3500
20 x 20
Immediate
0
0
-1
-0.8888
-0.857
-1
-1
1
Contour Current Y
Lo Limit
Hi Limit
5500 5031.4066
?
5500
40000
37137.19
?
40000
15 4.8129793
?
15
0 -20.93512
?
0
153 152.42056
?
153
695 700.21755
695
?

21420
Vapp

Thrust
(lbs.)

TOWOD
LDWOD

Ps
TOGW

14535
360

Area (ft^2)

White area now represents the smaller design space

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

520

Horiz

Vert

Factor

Current X

Radius

0.964

ULF

Grid Density

Update Mode

20 x 20

Immediate

Exploring the Space

0.71

CmbMach

DPayld

-1

Thrust

0.88888

Area

0.857

DStealth

-1

Auxtnk

-1

SFC

0.3333

Response
O&S
TOGW
LDWOD
TOWOD
Vapp
t_Radius
t_Rate

Ps
AltRng

Contour

Current Y

Lo Limit

Hi Limit
?

5130

5475.7833

45000

47224.344

30

26.563468

30

15

13.613377

15

151

150.2058

151

12656.5

10828.741

3.8115

4.0987387

780

807.60615

780

1540

1545.1224

1540

With a parametric environment, the

contours can be moved and re-evaluated
A designer can explore hundreds if not
thousands of potential design points for
multiple criteria

21420

TOWOD

Vapp
Ps

Alternate
Range

Turn Radius
Turn Rate
TOGW

Thrust
(lbs.)

O&S
14535
380
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Area (ft^2)

520

Contour Profiler Allows Visualization of Constraints in 3D

X

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

EngWt
TWR
WSR
FCDSUB
LPT_SD
LPT_BD
LPT_metal
LPT_BT
LPT_Vane_T
LPT_Load
LPT_eff
Duct_bleed
HPT_SD
HPT_BD
HPT_metal
HPT_BT
HPT_2VT
HPT_1VT
HPT_Load
HPT_eff
xfactor1
T4max
C_Dens
HPC_FSPR
HPCPR
HPC_eff
HPC_TS
LPCPR
LPC_eff
Fantech
FPR
Fan_eff
% CO2/ASM

EngWt
TWR
WSR
FCDSUB
LPT_SD
LPT_BD
LPT_metal
LPT_BT
LPT_Vane_T
LPT_Load
LPT_eff
Duct_bleed
HPT_SD
HPT_BD
HPT_metal
HPT_BT
HPT_2VT
HPT_1VT
HPT_Load
HPT_eff
xfactor1
T4max
UTEP
delphi
C_Dens
HPC_FSPR
HPCPR
HPC_eff
HPC_TS
LPCPR
LPC_eff
Fantech
FPR
Fan_eff
% Below Rule

The Contour Profiler allows the Response

Surface Equations to be viewed in three
dimensions
Constraints can be applied as shown with
the red surface

The radio buttons allow the axes to be

changed instantly
Tools allow the surface to be rotated and
viewed on multiple axes
The surface can be re-shaped by changing
the input data with slide bars which allows
trade studies to be performed and visualized
in real-time

Potential Solution Surface Plots of % CO2 and % Below LTO

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Dynamic Interactive Design Space TradeTrade-off Environment for an SST

Metric
Responses

Upper/Lower bounds of
the design space

Influence of parameter on response

(either or or no influence)
The larger the slope, the greater the influence

Hairlines move and update

responses in real-time
Optimization can be performed

1450671

TOGW 837264.1

T/W

TIT

X2

X4

Lower bound values of SLN and $/RPM

indicates NO feasible space
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

X5

X6

Y2

t/c
root

t/c
tip

HT
Area

-1 min value of Y2
0 current value of Y2
1 max value of Y2
All are in a non-dimensional space

-1

1
-1

-1

-1

-1

-1

1
-1

-1

-1

-1

-1

X3

-1

-1

CL
design

-1

OPR

-1

FPR

-1

-1

0.24

-1

-1

Wing
Area

-0.33

Design
Variables

-0.3301

-1

765601.0
18545
TOFL 10327.29
8980
14787
LDGFL 9047.182
8765
210.4
Vapp 155.0316
150.4
118.09
FON 107.3951
103.9756
120.43
SLN 110.1838
109.5424
0.17643
$/RPM 0.113477
0.10590

VT
Area

Probabilistic vs. Deterministic

Moving from deterministic design to robust/probabilistic design methods is
rooted in the recognition that uncertainty exists and has a significant impact
on system affordability
Want to analytically answer the questions:
How much design margin is really necessary?
What is the sensitivity of objectives to sources of uncertainty?
What can be done to reduce this impact?

Probabilistic Design
Robust Design

Common obstacles to implementation:

Organizational inertia
Lack of probabilistic analysis tool to bridge the gap between deterministic and
probabilistic methods
Computational Costs, if not approached intelligently

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Options for Probabilistic Design

I
II
III

Sophisticated
Analysis Code

Metamodel
(e.g., Response Surface)

Sophisticated
Analysis Code

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

100%

Monte
Carlo

(most exact)

P
0%

Time-Consuming
Computationally Intense

Monte
Carlo

Fast Probability
Integration
(approximates Monte Carlo)

100%
P
0%

100%
P
0%

Objective

Probability Distributions Can Be Input Into the RSEs

Select
Distribution

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Output
Distributions
100%

Constraint = 11,000 ft

90%
80%

Probability

With probability distributions, thousands

of designs across a user-specified
distribution can be analyzed
This allows a designer to assess technical
feasibility and economic viability
Without RSEs or metamodels, this
analysis would be impossible, due to the
execution time of large parametric spaces
RSEs are an enabler for this method of
design

70%
60%
50%
40%
30%
20%
10%
0%

7,000

P(success) = 4.6%
8,000

9,000

10,000

11,000

12,000

13,000

Takeoff Field Length (ft)

14,000

15,000

Investigate
Design
Space

Evaluate
System
Feasibility

Identify
Technology
Alternatives

System Feasibility:

Select Best
Family of
Alternatives

Evaluate
Technology
Alternatives

HSCT Design Space Representation

100%

90%

90%

80%

80%

70%
60%
50%
40%
30%
20%

70%
60%
50%

40%
30%
20%

P(success) = 3.3%

10%

Constraint = 106 EPNLdB

100%

Probability

Modeling
and
Simulation

Constraint = 155 kts

Define
Concept
Space

Probability

Define
the
Problem

10%

P(success) = 3.1%

0%

0%

120

130

140

150

160

170

180

190

90

200

95

100%

90%

90%

80%

80%

50%
40%
30%
20%
10%

Probability

60%

70%
60%
50%
40%
30%
10%
0%

0%
95

100

105

110

Sideline Noise (EPNL dB)

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

115

120

20%

P(success) = 0%
90

110

Constraint = $0.10/RPM

100%

70%

105

Flyover Noise (EPNL dB)

Constraint = 103 EPNLdB

Probability

Approach Speed (kts)

100

115

120

P(success) = 0%
0.09

0.10

0.11

0.12

0.13

0.14

0.15

0.16

Average Required Yield/Passenger Mile ($ FY96)

Joint Probability Decision Making

Correlation between
CO2/ASM and LTO NOx:

Monte Carlo Simulation

% Reduction of
CO2/ASM

% Below LTO
NOx Rule

-15.5%
-5.3%
-17.8%

-71.9%
-32.7%
-58.3%

Probability Density

Monte
Carlo Data

0.002

Histogram
of Data

0
40
5

-100 -30

-10

cov(x, y )

-20

Assuming -90 -70 -50 -30 -10

Normal
NOx = -61%
Distributions
NOx = 19.8%

CO2 =

-16%
CO2 = 4.3%

f ( x, y )

1
2

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

2
xy

exp

2(1

1
2
xy

= 0.1464

x
x

xy
x

y
y

y
y

Visualizing Potential Solutions to Meet Future Goals

Rim of lowest
frequency of
combinations

Assuming the Goal distributions are

normal, plot Joint distribution and
overlay future Goal target values to
determine if any combination of
technology metrics will achieve the
target
Interpret the joint probability as:
Highest frequency implies that of the
combinations considered, the majority
of the solutions will fall in this region
The outer edges or rim represent the
limits of what can be obtained with the
technology metric ranges that were
specified

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-90%

-80%

-70%

-15%

CO2/ASM

-20%

-25%

LTO NOx
Region of highest
frequency of
combinations

Selecting Potential Solutions to Meet Future Goals

Path of least resistance (based on the

degree of difficulty)
Physically realizable solution
Balance with the impact to other metrics
(noise, cost, vehicle performance)
Balance with the cost to achieve and the
time to develop with the performance
capability needed
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-90%

-80%

-70%

-15%

CO2/ASM

To acquire the technology metric values

that meet the new Goal values, simply
overlay the Monte Carlo Simulation
data
Pick the data points that meet the goals
and extract the associated technology
metric values
If multiple combinations exist that will
satisfy the new goals, then selection of
the appropriate path may be determined
by:

-20%

-25%

LTO NOx

Data points exist which

will satisfy both goals
concurrently.
Extract and investigate
needed capability

300 pax Dynamic Technology Gap Analysis

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

300 pax Dynamic Technology Gap Analysis

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

300 pax Dynamic Technology Gap Analysis

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

FON Goal

SLN Goal

Flyover Noise Reduction

% NOX Goal

Skew in joint
distribution indicates
correlation between
the variables

Sideline Noise Reduction

Points that satisfy emission reduction

goals for CO2 and NOx can be queried
in the other dimensions

% DOC+I

% NOX Reduction

HC Goal

% DOC+I Goal

% CO2/ASM Goal

No Configurations Meet HC Goal

% CO2/ASM Reduction

Joint Probability Distributions in Multiple Dimensions

Hydrocarbon Reduction
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

ASDL Probabilistic Methods Process

FPI

CDF
Customer Driven Requirements
(Concept & Technology Set Specific)

100%

Aero

Architecture B

Aspiration Space
Technology
Insertion
Impact

x
x

x
x

0%

Structures

Architecture A

Thrust

Probability

DISCIPLINARY RSEs

Objective

Design Point

x
Physics
Driven
Growth

x
Growth Spurs

Engine Weight

Competitive Assessment
Strategic Decision Making

Weights
Requirements
Space

RSEs

Dynamic
Contour
Plots

Constraints
%$/RPM

TOFLmod
SLNmod

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Technology
Space
Metrics/Objectives

Responses

Metrics/Objectives

Responses

Concept Space

Constraints

Metrics/Objectives
Constraints

SYNTHESIS & SIZING

Responses

Etc.

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

7788.6

6.13%

9797

0%

9230.9

-24.8%

6941.6

3.84%

162.4

0%

156.4

-16.5%

130.6

1.77%

109.3

0%

107.4

-27.75%

77.6

0.72%

111.3

0%

110.5

-23.08%

85.0

4.24%

0.1130

0%

0.1084

-21.59%

0.0850
0%

Wing
Weight

0%

Fuselage
Weight

0%

Engine
Weight

0%

0%

0%

-27.26%

0%

10706.9

-24 %

0%

-21 %

$/RPM

11456

46 %

SLN

7.00%

0%

TIF environment provides

transparency. The behavior of
these trends was invisible to
us before the parametric
environment was in place.

FON

596469

-10 %

Forecasting environment if
no specific technologies were
in mind

Vapp

-30.27%

7%

Impact of degradation of a
technology over the life of the
system

LdgFL

855352.7

-40 %

Identify code fidelity needed

to model a technology

TOFL

0%

-35%

Features of the TIF

Environment

921547.2

Dimensional
Values

TOGW

7.74%

% Change
from Baseline

Technology Impact Forecast Environment

Noise
Supersonic
Suppression
Drag

Technology Gap Analysis Tradeoffs

Horiz Vert

lbs

Minimum required
improvement in wing weight

=0.

.09

= $0

D3
R&

RPM
R&
0.4

LdB

EPN

=
D3

102

=
SLN

Wing Weight
(Structures)

.1
=$0
B
PM
NLd
$/R
3 EP
=10
SLN

Baseline

k
750

.095
= $0

W=

RPM

Feasible
Space

G
TO

7%

R&

D3

=0

.5

Minimum
Feasible
Minimum required Solution

improvement in drag

-35%
-24%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Supersonic Drag
(Aerodynamics)

0%

Factor
Wing Weight
Fuse Weight
Engine Weight
Noise Suppression
Supersonic Drag

Response
TOGW
SLN
$/RPM
R&D3

Current X
-26.6
-12.3
15.6
-5.778
-8.28
Contour Current Y
750000 749665.46
103 102.90273
0.1 0.0989133
0.5 0.3786822

Lo Limit
.

Hi Limit
750000

.
.
.

103
0.1
0.5

Technology Evaluations

Technologies:

T1: Composite Wing

T2: Composite Fuselage
T3: Circulation Control
T4: HLFC
T5: Environmental Engines T6: Flight Deck
T7: Propulsion Materials
T8: ISSA Structures T9: Smart Wing
T10: Active Flow Control T11: Active Noise Control

Full Factorial Investigation

107.4

0%

107.4

-27.9%

77.4

0%

110.5

0%

110.5

-23.4%

84.7

+9.1%

0.1183

0%

0.1084

-15.1%

0.0920

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

T1

T2

T3

T4

T5

T6

T7

T8

T9

T10

On

0%

Off

120.4

On

-22.9%

Off

156.3

On

0%

Off

159.4

On

+1.9%

Off

5984

On

9231

Off

0%
-35.2%

On

9517

Off

+3.1%

On

7629

Off

-28.8%

On

10707

Off

0%

% Change
from Baseline

$/RPM

11100

On

SLN

+3.7%

Off

FON

583504

On

Vapp

-31.8%

Off

LdgFL

855352

On

TOFL

0%

Dimensional
Values

TOGW

Off

+4.4% 892528

T11

300 PAX Dynamic Technology Environment

+3.49%

% CO2/ASM
% Below
LTO NOx
% DOC+I

0%
-23.40%
+26.70%
-31.50%
-88.18%
+5.85%
0%

on

off

on

on

off

off

on

off

on

on

off

off

on

off

on

off

on

off

on

off

on

off

on

on

off

off

on

off

% Change
From Baseline

-17.35%

off

off

off

off

off

off

off

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

% Reduction in CO2

0%
-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

0.00%
-31.50%

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

This Baseline
is the
baseline configuration.
+T3 +T5 +T7 +T8 +T9 +T10 +T11a/b+T13 +T14 +T15a +T16

300 PAX Dynamic Technology Environment

+3.49%

+26.70%
-82.76%
-88.18%
+5.85%
0%

on

off

on

off

on

on

off

on

off

off

on

off

on

off

on

on

off

off

on

off

on

on

off

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

-23.40%

on

% Below
LTO NOx

0%

off

% CO2/ASM

on

off

off

off

off

off

off

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

Addition of the TAPS advanced combustor

(T3) does not change the flowpath, but
results in a significant reduction in the LTO
NOx characteristics.
% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

+T5

0.00%

-82.76%

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

+26.70%
-83.54%
-88.18%
+5.85%
-2.99%

on

off

on

off

on

on

off

on

off

off

on

off

on

on

off

off

on

off

on

on

off

on

off

off

on

on

-17.35%
% Change
From Baseline

% DOC+I

-23.40%

off

% Below
LTO NOx

-3.91%

off

% CO2/ASM

on

off

on

off

off

off

off

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The highly loaded compressor system (T5)

improves the efficiency of the HPC, thus
reducing vehicle fuel burn and CO2 production.
+T5

+T7

% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-3.91%

-83.54%

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

+26.70%
-84.08%
-88.18%
+5.85%
-7.79%

on

off

on

off

on

on

off

on

off

off

on

on

off

off

on

off

on

off

on

on

off

off

on

off

on

on

-17.35%
% Change
From Baseline

% DOC+I

-23.40%

off

% Below
LTO NOx

-9.50%

off

% CO2/ASM

on

off

on

on

off

off

off

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The highly loaded turbine system (T7) improves

the work extraction of the turbine and the adiabatic
efficiency, which has a positive influence on
vehicle fuel burn, hence reducing CO2/ASM.
+T5

+T7

+T8

% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-9.50%

-84.08%

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

+26.70%
-78.57%
-88.18%
+5.85%
-8.87%

on

off

on

off

on

on

off

on

off

off

on

off

on

off

on

on

off

off

on

off

on

on

off

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

-23.40%

on

% Below
LTO NOx

-9.76%

off

% CO2/ASM

on

off

on

on

on

off

off

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The advanced compressor disk alloy (T8) allows for

an increase in the compressor discharge temperature,
allowing higher overall pressure ratios and, hence, a
higher cycle efficiency to lower CO2. A slight NOx
penalty is incurred due to a higher OPR.
+T5

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-9.76%

-78.57%

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

-88.18%
+5.85%
-13.19%

on

off

on

on

off

off

on

off

on

on

off

off

on

off

on

on

off

off

on

off

on

off

on

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

+26.70%
-80.66%

on

% Below
LTO NOx

-23.40%

off

% CO2/ASM

-17.62%

on

off

on

on

on

on

on

off

off

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The low conductivity ceramic TBC (T10) further

enables higher gas temperatures in the turbine, and
begins to mitigate the increase in NOx through a
large reduction in cooling flow requirements.
+T5

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

% Reduction in CO2

-6%
-12%
-18%

-17.62%

-24%

-80.66%

-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

26.70%
-77.92%
-88.18%

+5.85%
-13.72%

on

off

on

on

off

off

on

off

on

on

off

off

on

off

on

on

off

off

on

off

on

off

on

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

-23.40%

on

% Below
LTO NOx

-18.23%

off

% CO2/ASM

on

off

on

on

on

on

on

on

on

off

off

off

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The 2700 F CMC vanes and liner (T11a/b) have a

minimal impact to CO2 since T9 and T10 already
raise the turbine vane temperature, and benefits are
not seen from incremental effects.
+T5

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

% Reduction in CO2

-6%
-12%
-18%

-18.23%

-24%

-77.92%

-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

-88.18%
+5.85%
-15.35%

on

off

on

off

on

on

off

on

off

off

on

off

on

off

on

on

off

off

on

off

on

on

off

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

+26.70%
-78.73%

on

% Below
LTO NOx

-21.56%
-23.40%

off

% CO2/ASM

on

off

on

on

on

on

on

on

on

off

on

on

off

off

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The turbine tip clearance control

(T14) influences the adiabatic
efficiency, which improves fuel burn
and reduces the CO2/ASM.
+T5

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

-21.55%
-78.73%

0%
-10%
-20%
-30%
-40%
-50%
-60%
-70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

300 PAX Dynamic Technology Environment

+3.49%

-88.18%
+5.85%
-16.45%

on

on

off

off

on

on

off

on

off

off

on

off

on

on

off

off

on

off

on

off

on

on

off

off

off

on

-17.35%
% Change
From Baseline

% DOC+I

+26.70%
-78.91%

on

% Below
LTO NOx

-22.27%
-23.40%

off

% CO2/ASM

on

off

on

on

on

on

on

on

on

off

on

on

on

on

T3

T4

T5

T7

T8

T9

T10

T11a

T11b

T12

T13

T14

T15a

T16

The aspirated seal for the turbine (T16)

influences the adiabatic efficiency, which
reduces the fuel burn, decreasing the
amount of CO2/ASM produced.
% Reduction in CO2

-6%
-12%
-18%
-24%
-30%

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

+T5

+T7

+T8

+T9 +T10 +T11a/b+T13 +T14 +T15a +T16

0%
-10%
-20%
-30%
-40%
-50%
-60%
-22.27%
-78.91% -70%
-80%
-90%
-100%

% Reduction in NOX

Baseline +T3
0%

Additive Creation of the Overall Environment

Desirements
Constraints

Responses

Baseline +

Constraints

Desirements

Desirements
Constraints

Top Level Requirements

Technology kk-factors

Responses

Concepts (Design Variables)

Responses

Desirements
Constraints

Responses

Top Level Requirements

Design/Economic Variables

Technology kk-factors

Assumption: Interactions among the input variables exist only within each group
(Or regroup the inputs to eliminate interaction across subspaces)
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Project PROMETHEUS

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Varying Fidelity M&S Initiative

Safety
Safety

Aerodynamics
Aerodynamics

Ge ometry

Economics
Economics

Synthesis & Sizing

Structures

Mission

Structures

Manufacturing

Integrated Routines
Table Lookup
S&C

Approximating Functions
Direct Coupling of Analyses

Conceptual Design Tools

Increasing
Sophistication and
Complexity

(First-Order Methods)
Propulsion

S&C

Performance

Preliminary Design Tools

(Higher-Order Methods)

Propulsion

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Performance

Manufacturing

Our Motivation
High-fidelity, physics-based analyses need inclusion
earlier in the Design Process
Advanced Concepts
Multidisciplinary Design
Complex Tradeoffs
Shortened Design Cycle
Et cetera
Low-order results not trustworthy to guide vehicle
definition outside results of historical database
Utilization of CFD, FEM, during the conceptual design
phase is a figurative Holy Grail
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Varying Fidelity M&S Initiative

Access to Space (TBCC)

Parametric
Robust
Optimization
Multidisciplinary
Environment for
Technology and

Hyperspace
Exploration of
Unconventional
Systems
Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Hypersonic Missiles
Strategic Missiles
Project
PROMETHEUS

Supersonic Vehicles
Morphing Vehicles
Unmanned Air Vehicles
Electric Propulsion
(Fuel Cells, Solar, Hybrids)

Collaborative Visualization Environment (CoVE)

Design Process

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Collaborative Visualization Environment - CoVE

The CoVE will be a large projection screen permitting multiple
linked design applications to be simultaneously displayed.
Will synchronize early conceptual design tools with highfidelity analysis programs.
Key decision-makers will make design choices on-the-fly
and will immediately see the impact of their decisions.
High fidelity tools will be run via a Beowolf cluster:
Will provide high computational power.
Parallel computing will run design applications
simultaneously.
Backup Storage
Several terabytes of memory dedicated to each project.
Will permit CoVE users to access previous design iterations.

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

The CoVE Vision

Collaborative Design Environment with Advanced Visualization
High-fidelity Conceptual Design Tools at near real-time results
Real time analysis of problems
Ability to design collaboratively with team even between geographically
dispersed locations
Integrated design tools
State of the art visualization
Multi-disciplinary optimization in a collaborative environment
Physics based computing
Integrates decision makers to the design process

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Collaborative Visualization Environment (CoVE)

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Knowledge Management System Demo

RFP
&
TRD

Multi-Attribute
Possible
Ranked
Selected
Ranked
Matrix of Document
First
Order
Possible
Downselect
Decision
High
Fidelity
Detailed
Well
Defined
Concepts
Concepts
Conduct
Lit
Establish
Concepts
Selected
Final
Concept
Concepts
Alternatives
M&S& Brainstorm
Concepts
Making
Sizing
Final
Concept
Analysis
Final
Concept
Search
Baseline
Concepts
Step 2
Step 3
Brainstorming
Synthesis
Step 1
Step 4
Step 5
QFDIdentify
RFP & TRD
Define
Establishing
Evaluate
Rank
Evaluate
VS
Recieved
Customer
Concepts
the Need
Concepts
Concepts
Requirements
Measures of
executes& Alternatives
executes
executes
executes
Merit
executes
contributes
to
Baseline
Matrix of
Customer
Collaboration
Alternatives
Document
Team
Baseline
Step 8
Step 10
executes
Team
Step 7
Step 9
Step 6
Perform House of
Document
Final Paper &
Team
Tree
Affinity
Create High
Evaluate
Team Downselect
Detailed
Final
Presentation
Fidelity M&SDiagramTeam
Concepts
Quality
Diagram
Analysis

Design

Team

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Collaborative Visualization Environment (CoVE)

DURIP Award
10 ft high x 18 ft wide Display Wall
Allows complete design space and
analysis visualization

Future Plans
Integrated High Fidelity Distributed Computing
Facility for real time display of computationally
intensive analyses and simulations

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

PROMETHEUS - CoVE Environment

Dr. Dimitri N. Mavris, Director ASDL

School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]

Collaborative Visualization Environment (CoVE)

Geographically distributed collaboration

Dr. Dimitri N. Mavris, Director ASDL
School of Aerospace Engineering
Georgia Institute of Technology
Atlanta, GA 3033230332-0150
[email protected]