2571-1075ISSN (Online)
Purpose - Reinforced concrete (RC) shear walls are critical lateral-load resisting elements; however, reliable prediction of their axial–flexural interaction behavior remains difficult, particularly for irregular geometries and nonuniform reinforcement layouts. This study aims to develop an accurate and versatile analytical framework to evaluate the global axial–flexural response of RC shear walls. Design/methodology/approach - A fully nonlinear, code-independent numerical framework is formulated based on strain compatibility, equilibrium enforcement, and curvature-controlled sectional analysis. The model incorporates advanced stress–strain relationships for confined and unconfined concrete, a bilinear steel constitutive law, and a high-resolution fiber discretization scheme capable of representing arbitrary cross-sectional shapes. The framework generates complete moment–curvature responses and axial–moment (P–M) interaction diagrams under uniaxial bending. Findings - The results exhibit strong agreement with established analytical models and reported experimental trends. The framework accurately captures nonlinear degradation, neutral-axis migration, confinement effects, and the influence of reinforcement distribution on axial–flexural capacity. Practical implications - The proposed model provides a reliable tool for performance-based assessment, design, and optimization of RC shear walls beyond simplified code provisions. Originality/value - The study introduces a geometry-independent, fully nonlinear modeling approach that enables detailed evaluation of irregular RC shear walls with enhanced accuracy and practical applicability.
Read MoreDoi: https://doi.org/10.54216/IJBES.110201
Vol. 11 Issue. 2 PP. 01-19, (2025)
The study put forward an integrated artificial intelligence-based approach to the analysis and prediction of contracting disputes in Engineering Projects, especially through Machine Learning methods and Deep Learning methods. Current ways of managing contracts cannot effectively deal with the complicated nature of Legal Texts and do not provide for early identification of potential disputes. This developed System was built using the Python Programming Language, using key libraries for Natural Language Processing (NLP) and Machine Learning (ML). The cache of Contract Documents in all formats was transformed into numerical vectors using TF-IDF once all Document Processing and Clean-up Procedures were completed. Multiple Models were built, with trained versions of each, including Logistic Regression, SVM, Voting Classifiers and an MLP (Multi-Layer Perceptron) based Neural Network model. Since each Contracting Dispute was modelled separately to improve overall prediction accuracy, initial recommendations for resolution are generated. Results show that the MLP performed in a SUPERIOR fashion, with an Overall Model Accuracy of 88%, and F1 Score of 0.874, effectively classifying Contracting Disputes relating to Delays, Payments and Scope Variations. The application of this framework to an actual example taken from the construction industry in Syria reaffirmed the capability of automating contract text review and improving risk management. This reinforces the importance of artificial intelligence as a tool for increasing proactive decision-making and minimizing conflict in engineering projects.
Read MoreDoi: https://doi.org/10.54216/IJBES.110202
Vol. 11 Issue. 2 PP. 13-24, (2025)