ASIAN JOURNAL OF CIVIL ENGINEERING (BUILDING AND HOUSING) VOL. 7, NO. 6 (2006) PAGES 603-619
GENETICALLY OPTIMIZED ARTIFICIAL NEURAL NETWORK BASED OPTIMUM DESIGN OF SINGLY AND DOUBLY REINFORCED CONCRETE BEAMS
B. Saini∗a, V.K. Sehgala and M.L. Gambhirb Department of Civil Engineering, National Institute of Technology, Kurukshetra, Haryana-136119, India b Thapar Institute of Engineering and Technology, Patiala, Punjab, India
a
ABSTRACT
Optimum design of singly and doubly reinforced beams with uniformly distributed and concentrated load has been done by incorporating actual self weight of beam, parabolic stress block, moment-equilibrium and serviceability constraints besides other constraints. Also, this design expertise has been incorporated into a genetically optimized artificial neural network based on steepest descent, adaptive and resilient back-propagation learning techniques. The initial solution for the optimization procedure has been obtained using limit state design as per IS: 456-2000.
Keywords: reinforced concrete, cost optimization, beam design, artificial neural network, genetic algorithms, hybrid systems
Notation: The following symbols are used in this paper: ast = area of tension reinforcement; asc = area of compression reinforcement; b = width of beam in mm; Cc= cost of concrete per unit volume in Rs./m3; Cf = cost of form work per unit area in Rs./m2; Cs= cost of steel per unit volume in Rs./m3; d = effective depth of beam in mm; d ' = effective cover to reinforcement in mm; Err_g = Error goal of ANN; F(x) = squared error function of ANN; fck = characteristic compressive cube strength of concrete in MPa; fy = characteristic yield strength of steel in MPa; δall = allowable deflection in mm;
∗
Email-address of the corresponding author: babitasaini6@rediffmail.com
604
B. Saini, V. K. Sehgal and M. L. Gambhir
δtot = sum of short and long term deflection in mm; l = span in m; lr = Learning rate of ANN; Mc = moment capacity of...
No comments