|
Abstract
|
To estimate the cost of a building prior to the detail design phase, engineers and project managers need suitable tools and
guidelines. Steel is an important construction material that is used in high volumes in buildings and has a significant role in
the total cost of projects. In this paper, the application of the artificial neural network (ANN) method to predict the quantity
of steel used in the steel moment-resisting frame (MRF) structures is presented. First, more than 1100 steel MRF structures
were designed applying the changes in the influenced parameters, then these models were transferred to the ANN, and finally,
the results of the performed parametric study were analyzed. The obtained results demonstrate that by using the proposed
ANN method, the weights of the structures can be estimated with an acceptable accuracy prior to the starting of the design
process. Based on the performed parametric study, several sets of required inputs in terms of the parameters of the story
height, the span length, the number of stories, the seismicity rate of the construction site, ductility, the class of soil site and
column cross section type influenced on the weight per unit area of the structure are submitted.
|