Introduction

Structural optimization is an important area of modern civil engineering. The objective is to design structures that satisfy all performance requirements while minimizing or maximizing a selected engineering criterion. Typical goals include reducing structural weight, increasing stiffness, improving load-carrying capacity, reducing cost, or enhancing reliability.

In traditional design procedures, engineers often rely on experience and iterative trial-and-error modifications. Structural optimization provides a systematic mathematical framework that allows the computer to search for the best design among many feasible alternatives.

The development of the Finite Element Method (FEM) has made optimization particularly attractive. FEM enables accurate evaluation of structural responses such as displacements, stresses, strains, and buckling loads. Optimization algorithms then use these responses to improve the design automatically.

Formulation of an Optimization Problem

A structural optimization problem can be written in the general form:

subject to

where $\mathbf{x}$ - vector of design variables, $f(\mathbf{x})$ - objective function, $g_i(\mathbf{x})$ - inequality constraints, $h_j(\mathbf{x})$ - equality constraints.

The design variables may represent:
• Cross-sectional dimensions,
• Plate thicknesses,
• Material properties,
• Nodal coordinates,
• Topological parameters.

Types of Structural Optimization

Size Optimization

Size optimization modifies dimensions of structural members while keeping the geometry unchanged.

Examples:
• Beam height,
• Cross-sectional area of truss bars,
• Plate thickness.
This is the most common optimization problem in civil engineering because it is relatively easy to formulate and solve.

Shape Optimization

Shape optimization changes the geometry of the structure while preserving its connectivity.

Examples:
• Arch profile,
• Bridge girder shape,
• Position of structural boundaries.

Topology Optimization

Topology optimization determines the optimal material distribution inside a given design domain. The optimization process identifies where material should be retained and where it can be removed.
Applications include:
• Lightweight structures,
• Conceptual structural design.

Weight Minimization

One of the most common objectives is minimizing structural weight:

where $\rho$ -- material density, $A_i$ -- cross-sectional area, $L_i$ -- member length.

Typical constraints include stress limits and displacement limits.

Compliance is a measure of structural flexibility:

where $\mathbf{F}$ - external load vector, $\mathbf{u}$ - displacement vector.

Minimizing compliance is equivalent to maximizing stiffness.

The design must satisfy:

for all structural members.

This ensures that the structure remains within allowable strength limits.

Serviceability requirements often limit structural displacements:

Examples include deflection limits for beams and floors.

Optimization Methods

Gradient-Based Methods

These methods use derivatives of the objective function and constraints.

Examples:
• Steepest Descent,
• Newton Method,
• Sequential Quadratic Programming (SQP).

Advantages:
• Fast convergence,
• Efficient for large FEM models.

Disadvantages:
• May converge to local minima,
• Require sensitivity analysis.

Evolutionary Algorithms

These methods are inspired by biological evolution.

Examples:
• Genetic Algorithms,
• Differential Evolution,
• Particle Swarm Optimization.

Advantages:
• Global search capability,
• No derivative information required.

Disadvantages:
• High computational cost,
• Slower convergence.

In practical engineering applications, optimization is usually coupled with FEM.

The computational workflow is:
• Define design variables.
• Generate FEM model.
• Compute structural response.
• Evaluate objective function and constraints.
• Update design variables.
• Repeat until convergence.

The FEM equilibrium equation is:

where the stiffness matrix $\mathbf{K}$ depends on the design variables.

Consequently, every optimization iteration requires a structural analysis.

Numerical Example: Weight Optimization of a Steel Bar

Consider a steel bar subjected to a tensile force $P$.

The objective is to minimize weight by selecting the optimal cross-sectional area $A$ while satisfying a stress constraint.

Problem Formulation

Objective:

\begin{equation}
\min W = \rho A L
\end{equation}

Constraint:

\begin{equation}
\sigma = \frac{P}{A}
\leq \sigma_{allow}
\end{equation}

Parameters:

• $P = 200$ kN,
• $L = 3$ m,
• $\rho = 7850$ kg/m$^3$,
• $\sigma_{allow} = 250$ MPa.

The analytical solution is:

\begin{equation}
A_{opt}
=
\frac{P}{\sigma_{allow}}
=
\frac{200 \times 10^3}
{250 \times 10^6}
=
8 \times 10^{-4}\ \text{m}^2
\end{equation}

Python Implementation

import numpy as np
from scipy.optimize import minimize
P = 200e3          # N
L = 3.0            # m
rho = 7850         # kg/m^3
sigma_allow = 250e6
def objective(A):
    return rho * A[0] * L
def stress_constraint(A):
    return sigma_allow - P / A[0]
x0 = [0.002]
constraints = [
    {'type': 'ineq',
     'fun': stress_constraint}
]
bounds = [(1e-5, 0.01)]
result = minimize(
    objective,
    x0,
    bounds=bounds,
    constraints=constraints
)
A_opt = result.x[0]
print("Optimal area =", A_opt, "m^2")
print("Weight =", objective([A_opt]), "kg")

The numerical result converges to approximately:

\begin{equation}
A_{opt} \approx 8 \times 10^{-4}\ \text{m}^2
\end{equation}

which agrees with the analytical solution.

Summary

Structural optimization is a fundamental tool in modern engineering design. By combining optimization algorithms with FEM analysis, engineers can systematically improve structural performance while reducing material consumption and cost.

The most common objectives are:
• Weight minimization,
• Compliance minimization (stiffness maximization),
• Stress reduction,
• Displacement control.
For civil engineering applications, size optimization is often the first and most practical step, while topology optimization represents the current state-of-the-art approach for achieving highly efficient structural layouts.