1
ESD.77 MSDO
–
Spring 2010
Space Shuttle External Tank Optimization
Anonymous MIT Students
System Design and Management
Massachusetts Institute of Technology
Abstract
A
simplified
model
of
the
Space
Shuttle
External
Tank
was
used
to
set
up
a
Multidisciplinary System Design Optimization
problem. First, Return of Investment (ROI) was
established as single objective. Both gradient
based and heuristic methods were used to solve
this
problem.
Sequential
Quadratic
Programming (SQP) was the gradient based
method chosen. In the other hand, a Genetic
Algorithm was used as the heuristic optimization
tool.
In
addition,
sensitivity
analysis
was
performed to the optimal solution found. Finally,
a multi-objective problem was set up adding the
total tank weight (TW) as second objective.
Adaptive Weighted Sum (AWS) was the method
selected to solve the problem.
Introduction
After the Cold war, and especially in times of
economic crisis, manned space programs have
been questioned per its high costs and the
associated safety risks. Reusable launch vehicles
(RLV’s)
have emerged as an alternative to
reduce
expenses
by
reusing
equipment
and
commercializing
space
flights.
Recent
discussions in Obama’s administration about the
financials of future human space exploration
have
motivated
us
to
explore
the
use
of
Multidisciplinary System Design Optimization
(MSDO) as a tool to improve the business case
of current space systems.
The most successful RLV has been, without
question, NASA’s Space Shuttle.
At a high
level, the elements of the Space Shuttle (at
launch) are: the external tank, two solid rocket
boosters and the orbiter vehicle. The external
tank has several functions: to provide the fuel
and the oxidizer for the main engines (liquid
hydrogen and liquid oxygen) and to serve as
structure to the system (the solid rocket boosters
and the obiter vehicle are attached to the tank at
launch). The tank is the only element that is not
reused and is also the heaviest.
Framing the optimization problem
In this analysis we will use a simplified model of
the external tank as described in Figure 1. This
model assumes the tank is divided in three main
sections: the hemisphere, the cylinder and the
nose
cone.
Table
1
shows
the
six
design
variables considered in this problem and Table 2
shows the parameters assumed.
Figure 1.
Graphic representation of the External
Tank simplified model
Table 1.
Design Variables
Symbol
Variable
Variable name
x
1
HR2
Height /radius ratio
x
2
L
Length of cylindrical body
x
3
R
Radius of the hemisphere
H
L
R
TCO
TCY
TH
*HR2=H/R

This
** preview**
has intentionally

**sections.**

*blurred***to view the full version.**

*Sign up*