Fuzzy inference system to formulate compressive strength and ultimate strain of square concrete columns wrapped with fiber-reinforced polymer

In this paper, fuzzy inference system (FIS) is employed to develop a more accurate approach to evaluate the strength and strain capacity of axially loaded concrete columns with the square section confined by fiber-reinforced polymer (FRP) wraps. To do so, an experimental database containing 261 test data on compressive strength and 112 test data on ultimate strain is collated from the literature. By using subtractive clustering algorithm to extract cluster centers from the experimental database, the structure of FIS model is identified. To select the best FIS model, several constant and linear (i.e., zeroth- and first-order) Takagi–Sugeno FIS models with different numbers of rules are developed and their performances in terms of the model output errors with respect to training data set as well as validation data set are compared. The finally proposed FIS models for calculation of strength and strain contain as few as three rules. Besides, the proposed FIS models are expressed as closed-form formulations, which can be conveniently used in practice. The outputs of the proposed FIS models agree favorably with the test data and outperform the existing models by providing more accurate prediction of both strength and strain capacity. In view of the FIS models, a parametric study is carried out to examine the influence of various variables including the section corner radius as well as the elastic modulus and tensile strength of FRP on the capacity of FRP-confined square columns.