Effect of number of neurons and layers in artificial neural network and membership and defuzzification functions on fuzzy logic for prediction of lateral confinement coefficient of carbon fiber reinforced polymer (cfrp) strengthened columns

Turkey is located in an active earthquake zone and has experienced many earthquakes to date. It has paid the consequences of this both financially and morally. With the increasing population in Turkey, the existing building stock is increasing day by day. Many unplanned constructions and uncontrolled houses have emerged. For the last 30 years or so, the concept of reinforcement of structures that emerged with this unplanned construction and uncontrolled housing has been carried to engineering practices around the world. In recent years, our country has been shaken severely by the 1998 Adana-Ceyhan earthquake, 1999 Düzce earthquake, 2011 Van earthquake and most recently 2023 Kahramanmaraş earthquake. There have been many earthquakes until today and then the regulations have been updated and new ones have been issued. Since 1940, 7 earthquake regulations have been published. In recent years, in line with the developing technology and needs, the regulations have been revised in accordance with the age, the variety of materials used and building models have increased. Different methods or materials are preferred for repair-strengthening or improvement of the structure in reinforced concrete structural elements. Among these materials, Fiber Reinforced Polymers (LP) are used in the reinforcement of reinforced concrete structural elements. LP is a good option because it is a lightweight material that can be applied faster and easier than traditional strengthening methods and has high strength. In this study, the lateral envelopment coefficients of circular, rectangular and square reinforced concrete columns of different cross-sections reinforced with carbon based fibrous polymer (CFRP) were modeled with artificial intelligence methods using experimental data in the literature. The experimental data obtained from the literature were modeled with artificial neural networks and fuzzy logic methods using MATLAB software program and the column lateral envelopment coefficient predictions were performed by optimizing the artificial intelligence models of the columns with different cross-sections.