THE USE OF ARTIFICIAL NEURAL NETWORKS (ANN) FOR THE PRELIMINARY DESIGN OF HIGH-RISE BUILDINGS

Abstract eng:
The need for high-rise buildings is increasing in order to meet the challenges posed by rapid urbanization and the exponential growth in population. With recent advancements in computing technologies, innovative materials and new structural systems, the need of prior knowledge about proportioning the geometry and configuration of structural members is also increasing. A relatively quicker and reliable estimation of approximate sizes of members can greatly facilitate the preliminary design and feasibility of the project. This study uses an Artificial Neural Network (ANN) based approach to directly determine some key design parameters based on experience gained from previously designed buildings. It uses a heuristic tool using ANN that can provide fast and reliable results based on two algorithms (i.e. Multilayer Perceptron with Back Propagation using Levenberg-Marquardt algorithm MLP-BP, and PCA-Sparse-Extreme Learning Machine with online Sequential learning). The ANN models are trained to determine structural design indicators from architectural parameters. The proposed approach can not only provide means for quick estimation of design output, but can also for crosschecking of code based & performance based design very quickly. The objective is to provide means of assisting the design team and clients to make key design decisions based on cumulative experience rather than relying on judgment of individual designers. The approach is demonstrated through the sample networks trained on various case study high-rise buildings for which the required architectural and structural design results have been generated through detailed design.

• Export as: BibTeX | MARC | MARCXML | DC | EndNote | NLM | RefWorks
• Add to your basket