International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 04 Issue: 07 | July -2017
p-ISSN: 2395-0072
www.irjet.net
Design of reinforced concrete structures using neural networks Chandan.M.K1, Raghu Prasad.B.K2, Amarnath.K3 1Student,
The Oxford College of Engineering, Bangalore 2Retd.IISc Professor, Bangalore 3 Professor and Head, Dept. of Civil Engineering, The Oxford college Of Engineering, Karnataka, India ---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract - Optimization techniques play an important role
in structural design. The purpose is to find the best ways so that a designer or a decision maker can derive maximum benefit from the available resources. In the present study a column, a beam and a G+4 storey model are modelled using STAAD PRO v8i software. Static analysis of the structure is carried out and the results like axial forces (P) , bending moments (M) , support reactions(R) are recorded. The results are tabulated along with other parameters like Area of steel(Ast), Breadth of beam (B), Depth of beam or slab (D) , Characteristic compressive strength of concrete (fck) , Characteristic strength of steel (fy) , Design bending moment (Mu) and percentage of steel (pt). The design of Reinforced concrete members under uni axial bending is done manually as per IS-456:2000 and SP 16 and percentage of steel is calculated and noted down. The results from the static analysis of the structure which are in tabulated form are tested and trained in MATLAB neural network toolbox. The predicted values for the percentage of steel by neural network toolbox are noted down. The percentages of error for the predicted values are almost negligible when compared to those obtained by conventional method for most of the cases and are in good agreement with one another. Key Words: Optimization, Artificial Neural Network , STAAD PRO , MATLAB , IS-456 : 2000 etc...
from his past experiences he knows what to look out for and what not to worry about. Similarly it would be beneficial if machines too, could utilize past events as part of the criteria on which their decisions are based, and this is the role that neural networks seek to fill.
1.2 Artificial neural networks ANNs consist of a number of processing units analogous to neurons in the brain called nodes. Each node has a function called node function which are associated with a set of local parameters . The local parameters determine the output of the node when input is given. If the local parameters are modified , the node functions may get altered. Hence, the artificial neural networks can be defined as the information-processing system in which the elements called neurons, process the information.
1.3. Structure of neural network The neural networks can be single layered or multi-layered. A single layered neural network is composed of two input neurons and one output neuron. A multi-layered artificial neural network(MNN) consists of input layer, output layer and a hidden layer of neurons. The hidden layer of neurons is also called as intermediate layer of neurons. A three layered neural network is shown in the figure below.
1. INTRODUCTION The artificial neural network (ANN) was developed 50 years ago. ANNs are the simplifications of biological neural networks. The neural networks are very important tool for studying the structure of human brain. Due to the complexity and for complete understanding of biological neurons , many architectures of ANN have been reported in the present study.
1.1 Aim of neural networks The aim of neural networks is to replicate the human ability to adapt to changes taking place in the current environment. This depends on the capability to learn from the events that have happened in the past and to be able to apply that to future situations. For example : The decisions made by trainee doctors are rarely based on a single symptom due to complexity of human body. But an experienced doctor is far more likely to make a good and effective decision than a trainee , because Š 2017, IRJET
|
Impact Factor value: 5.181
|
Fig -1: The above figure shows densely interconnected three layered static neural network in which each circle represents an artificial neuron.
ISO 9001:2008 Certified Journal
|
Page 2012