Research Journal Paper (Q1) For Structural Engineering Journals - Machine Learning Methods

Machine Learning Method for Analytical Investigation on Rocking Walls by Artificial Neural Network Abstract In this Study, behaviour of self-centered post-tensioned concrete rocking wall was studied by soft computing and machine learning techniques, developing artificial neural network (ANN) model. To generate required dataset for modelling, experimental results were employed and effective input parameters including length, height and thickness of wall, number, area and yield stress of tendons, distance of each tendon from edge of wall, prestressing ratio of tendons, compressive strength of concrete and axial stress ratio of wall were selected for training network model. Sensitivity analysis of input parameters and network performance validation was conducted. MONN model was verified by experimental results and showed proposed method was able to accurately predict initial stiffness, secant stiffness, lateral behaviour, cyclic response, and neutral axis. Prediction of these parameters required simultaneous control and design of self-centered wall by MONN model that was a new approach in field of designing these systems. The accuracy of MONN was compared to building codes and existing procedures. This approach by machine learning method was a new effort for investigating behaviour of self-centered concrete rocking walls that showed the proposed network can be used to as an