Scribd
Upload
31 views1 page

Binder1 50 PDF

Neural networks can be used to approximate solutions to complex civil engineering problems and accelerate convergence in numerical simulations. They have been combined with iterative solving techniques like preconditioned conjugate gradient algorithms to provide initial solutions and act as a preconditioner. This hybrid approach was applied to finite element analysis of bridges. Neural networks have also been used for design automation and optimization of large civil engineering systems. Specifically, research has integrated penalty functions, stability theorems, and neural dynamics to develop robust optimization models for problems with thousands of variables and non-linear constraints. These hybrid computational models allow for automated optimal design of complex structures subject to real-world design code constraints.

Uploaded by

Abdul Rahman
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
31 views1 page

Binder1 50 PDF

Neural networks can be used to approximate solutions to complex civil engineering problems and accelerate convergence in numerical simulations. They have been combined with iterative solving techniques like preconditioned conjugate gradient algorithms to provide initial solutions and act as a preconditioner. This hybrid approach was applied to finite element analysis of bridges. Neural networks have also been used for design automation and optimization of large civil engineering systems. Specifically, research has integrated penalty functions, stability theorems, and neural dynamics to develop robust optimization models for problems with thousands of variables and non-linear constraints. These hybrid computational models allow for automated optimal design of complex structures subject to real-world design code constraints.

Uploaded by

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

8 OVERVIEW OF NEURAL NETWORKS IN CIVIL ENGINEERING

analysis, neural networks can be used to create domain specific equation solvers utilizing the knowledge of
a particular domain such as highway bridges. But, neural networks can provide only an approximate
solution where an exact solution is usually required. Consolazio (2000) proposes combining neural
networks with iterative equation-solving techniques such as a preconditioned conjugate gradient (PCG)
algorithm (Adeli and Kumar, 1999). In particular, he uses the BP neural network algorithm to compute
approximate displacements at each iteration, while the overall PCG algorithm steers convergence to the
exact solution. The neural network part of the algorithm improves the efficiency of the algorithm by (a)
providing a good initial solution and (b) playing the role of the preconditioner in the PCG algorithm. The
author applies the method to finite element analysis of flat-slab highway bridges and concludes the neural
network to be an effective method for accelerating the convergence of iterative methods. Use of neural network
in finite element analysis is also discussed by Li (2000).

2.3.2
Design Automation and Optimization
Automation of design of large one-of-a-kind civil engineering systems is a challenging problem due partly
to the open-ended nature of the problem and partly to the highly nonlinear constraints that can baffle
optimization algorithms (Adeli, 1994). Optimization of large and complex engineering systems is
particularly challenging in terms of convergence, stability, and efficiency. Most of the neural networks
research has been done in the area of pattern recognition and machine learning (Adeli and Hung, 1995).
Neural network computing can also be used for optimization (Berke et al., 1993).
Adeli and Park (1995a) present a neural dynamics model for optimal design of structures by integrating
penalty function method, Lyapunov stability theorem, Kuhn-Tucker conditions, and the neural dynamics
concept. A pseudo-objective function in the form of a Lyapunov energy functional is defined using the
exterior penalty function method. The Lyapunov stability theorem guarantees that solutions of the
corresponding dynamic system (trajectories) for arbitrarily given starting points approach an equilibrium
point without increasing the value of the objective function. The robustness of the model was first verified
by application to a linear structural optimization problem, the minimum weight plastic design of lowrise
planar steel frames (Park and Adeli, 1995). Optimization algorithms are known to deteriorate with the
increase in size and complexity of the problem. The significance of the new optimization model is that it
provides the optimum design of large structures with thousands of members subjected to complicated and
discontinuous constraints with excellent convergence results.
In order to achieve automated optimum design of realistic structures subjected to actual constraints of
commonly-used design codes such as the American Institute of Steel Construction (AISC) Allowable Stress
Design (ASD) and Load and Resistance Factor Design (LRFD) specifications (AISC, 1995, 1998), Adeli
and Park (1995b, 1996a) developed a hybrid CPN-neural dynamics for discrete optimization of structures
consisting of commercially available sections such as the wide-flange (W) shapes used in steel structures.
The computational models are shown to be highly stable and robust and particularly suitable for design
automation and optimization of large structures no matter how large the size of the problem is, how
irregular the structure is, or how complicated the constraints are. For their innovative landmark work the
authors were awarded a patent by the U.S. Patent and Trademark Office on September 29, 1998 (United States
Patent Number 5,815,394).
An important advantage of cold-formed steel is the greater flexibility of cross-sectional shapes and sizes
available to the structural steel designer. The lack of standard optimized shapes, however, makes the
selection of the most economical shape very difficult if not impossible. This task is further complicated by

You might also like