Form-Finding of Long Span Bridges with

Continuum Topology Optimization

and a Buckling Criterion

Salam Rahmatalla (Graduate Student)

Prof. Colby C. Swan (Adviser)

Department of Civil and Environmental Engineering

Center for Computer-Aided Design
The University of Iowa
 Objective Statement

To develop a continuum topology formulation

capable of finding structural forms of
maximum buckling stability.
 Introduction Structural Optimization

. Size Optimization

. Shape Optimization

. Topology Optimization
Alternative Topology Optimization Formulations

Discrete topology optimization Continuum topology optimization

(Ground structures)
 Elements of Continuum topology optimization
Applied loads

numnp
( X) = N ( X)
i =1
i i Starting Design domain Supports

i
Intermediate design

( X) = 0 ( X) = 1.0

0 ( X) 1.0 E ( X) = E ( (X )) [mixing rule]

 Sparsity of Long-Span Bridges

Sunshine Skyway bridge cable-stayed bridge in Tampa, Akashi Bridge suspension bridge In Japan
Florida

Most long-span bridges occupy < 1% of their envelope volume.

 SPARSITY in Topology Optimization
Fixed-mesh model of full envelope volume;
must capture the form of the structure with realistic sparsity
must capture mechanical performance of the structure

Fine meshes are required;

Implies large computational expense;
20 elements 50 elements 224 elements

40 elements 100 elements 224 elements

50% Material constraint usage 10% Material constrain usage 3% Material constraint usage
 Structure modeled as linearly elastic
system
Stability analysis performed via
linearized buckling analysis

Linear elastic problem: K L u = f ext

Eigenvalue problem: K L (b) + G (u, b) = 0 T K L

= T
G
1
Objective function: L(u, b) =
min( )

Design sensitivity analysis:

dL T G 1 K L a T K L f ext
= ( + ) + (u ) ( .u )
db b b b b
G
K L ua = T
u
Design Constraints

Bounds on individual design

variables
0 J 1 J = 1,2, K ,NUMNP

Material usage constraint

(X) d

C
d

 Initialization Starting guess

Structural analysis
Find u(b) and (b)

Gradient-based Sensitivity analysis (Linearization)

Optimization
Algorithm Optimization step
(Linear programming sub-problem)

Optimal?
No
Yes
Stop
 Fix-end Beam Problem

F d F
Design domain
with 10% material
constraint.

Resulting
topology
= 1.496*106 = 5.09*106

First buckling
mode.
= 10.0 = 19.51

Minimizing the Maximizing the min. Maximizing the min.

general critical buckling load, critical buckling load,
compliance. nonlinear formulation. linearized buckling.
 Circle Problem

Undeformed
configuration

= 4.09*105 = 2.66*105
Design domain
With material
2.5% Constraint
Deformed
configuration

= 1.57 *109 = 3.28*109

Minimizing structural Maximizing the min.
Compliance critical buckling load,
linearized buckling.
 Canyon Bridge Problem
Design domain
With 12.5% material 100m
Constraint and 10 kPa
300m
applied load.

1000m
Minimizing the
general compliance

= 1.7 *103 = 4.84*103

Maximizing the min.
critical buckling load,
linearized buckling.

= 2.49*103 = 3.88*103

Maximizing the min. 80 elements 80 elements

critical buckling load,
linearized buckling.
(Non-designable deck)
200 elements 200 elements
= 1*10 4
= 6.98*103
 Long-Span Bridge Problem 2-Supports
3000m

Design domain with 200m

12.5% material constraint
and 10 kPa applied load. 375m 375m
(a)
50 elements

50 elements = 2.92*105
Compliance Minimization = 58.3
200 elements
(b)
Compliance Minimization 60 elements
(finer mesh)

900 elements
(c)
Optimimum linearized buckling = 3.16*106
stability = 412

200 elements
(d)
 Long-Span Bridge Problem 3-Supports

3000m
Design domain with 250m
12.5% material constraint
and 10 kPa applied load.
375m 150m 375m

50 elements = 2.06*104
Compliance Minimization = 5.29*102

200 elements

50 elements = 1.52*105
Optimimum linearized buckling = 5.15*103
stability

200 elements
 Summary & Conclusion

A formulation has been developed and tested for form-finding of

large-scale sparse structures;

The formulation is based on linearized buckling analysis;

The structural form and topology are optimized to achieve

maximum buckling stability;

The formulation yields concept designs that resemble existing

large-scale bridge structures;