Estimation of Reinforcement Steel Weight and Concrete Consumption in Reinforced Concrete Residential Buildings Using Artificial Neural Networks

نوع: Type: Thesis

مقطع: Segment: masters

عنوان: Title: Estimation of Reinforcement Steel Weight and Concrete Consumption in Reinforced Concrete Residential Buildings Using Artificial Neural Networks

ارائه دهنده: Provider: Sayed Ali Mahdavi

اساتید راهنما: Supervisors: Dr. Jalal Akbari

اساتید مشاور: Advisory Professors:

اساتید ممتحن یا داور: Examining professors or referees: Dr. Mohammad Sadegh Ketabi Dr. Mahram Mansouri Zadeh

زمان و تاریخ ارائه: Time and date of presentation: 2026

مکان ارائه: Place of presentation: آمفی تاتر

چکیده: Abstract: Abstract Accurate estimation of material quantities—particularly reinforcement steel weight and concrete volume—is one of the major challenges in the early stages of design and cost management of construction projects. Significant fluctuations in material prices in recent years have further emphasized the need for fast and reliable methods for preliminary material estimation. Traditional quantity take-off methods require completion of structural design and preparation of detailed construction drawings and therefore do not provide sufficient efficiency and accuracy during the initial phases of a project. The present study aims to develop an intelligent model based on artificial neural networks for estimating reinforcement steel weight and concrete volume in reinforced concrete residential buildings with intermediate moment-resisting frames. In this research, the required dataset was compiled through the design and analysis of 115 reinforced concrete buildings in accordance with the Iranian Standard No. 2800 and the Ninth Chapter of the Iranian National Building Regulations. One building was independently modeled, analyzed, and designed by the author using ETABS and SAFE software, and code-based checks—including torsional irregularity, redundancy factor (ρ), fundamental period, stability index, interstory drift, beam–column joint shear, and foundation controls—were performed to ensure data reliability. The model input parameters included the number of bays in two directions, plan area, number of stories, structural height, number of columns per floor, and seismic base shear. The total reinforcement weight and total concrete volume were considered as the model outputs. After data normalization, the proposed model was developed using a multilayer perceptron neural network with the optimal architecture of 7–10–8–2 (7 input neurons, two hidden layers with 10 and 8 neurons, and 2 output neurons) and trained using stochastic gradient descent with momentum. The dataset was divided into training (70%), validation (15%), and testing (15%) subsets. Model performance was evaluated using the correlation coefficient (R) and mean squared error (MSE). The results demonstrated high prediction accuracy; the MSE values for training, validation, and testing datasets were 0.00155, 0.00177, and 0.00133, respectively, while the corresponding correlation coefficients were 0.9762, 0.9737, and 0.9817. The relative error in most samples was less than 10%, and correlation analysis indicated that base shear, number of stories, and structural height had the greatest influence on the outputs. Additionally, nonlinear regression relationships with coefficients of determination around 0.92 were derived for rapid material estimation. Based on the findings, artificial neural networks can serve as an efficient tool in the preliminary stages of structural design to provide fast and reasonably accurate estimates of reinforcement and concrete quantities without requiring detailed design. The strong agreement between actual and predicted values and the stable performance of the model across the full range of data indicate good generalization capability for new projects with different specifications. This approach, in addition to reducing time and cost, can effectively support economic decision-making, project feasibility assessment, optimal design selection, and material supply planning.