چكيده به لاتين
Strengthening concrete columns with Fiber Reinforced Polymer (FRP) has gained significant attention in recent decades due to advantages such as lightweight, corrosion resistance, and high strength. The method presented in ACI 440-2R for strengthening columns uses FRP wrapping, which enhances the column's performance by providing confinement to the concrete. This method has been evaluated as suitable for columns with compression-controlled behavior, but it lacks efficiency for columns with predominantly flexural behavior. Some researchers have suggested the use of the Near Surface Mounted (NSM) method for FRP bars with or without FRP wrapping in such cases. Given the novelty of the method, Experimental studies under static loading have less frequently considered the impact of various factors on the performance of columns retrofited by these methods. Therefore, in the present research, 33 experimental samples were constructed, strengthened, and subjected to pure compression, eccentric loading, and pure bending. The variables considered included the type and number of NSM FRPs, the type and number of FRP wrapping layers, groove dimensions, and eccentricity of the load. Subsequently, an analytical model for flexural samples, a finite element model, and code-based calculations for samples under pure and eccentric compression, as well as a fuzzy model for predicting the capacity of rectangular section samples under various eccentricities, were presented. The Takagi-Sugeno-Kang (TSK) fuzzy model algorithm was modified to handle incomplete data in modeling. Experimental results indicated better performance of samples strengthened with the hybrid method compared to those strengthened with the NSM method in all loading conditions. The maximum capacity increase due to the hybrid and NSM methods for samples under bending, pure compression, and eccentric loading was 91% and 50%, 34% and 9%, and 36% and 17%, respectively. Additionally, the positive effect of increasing the number of FRP bars, the number of FRP wrapping layers, and groove dimensions on the capacity was observed. Increasing these parameters in bending, pure compression, and eccentric loading led to respective increases compared to their corresponding control samples. Considering the price of glass fibers to be approximately one-third that of carbon fibers, the threefold price increase of CFRP compared to GFRP did not result in a threefold capacity increase, making GFRP a suitable economic alternative to CFRP. Finally, the maximum error of the presented analytical and finite element models was 6% and 8%, respectively, indicating the models' accuracy. The analytical model accounted for the limitations related to the bonding of CFRP bars to concrete. Furthermore, a parametric study on the eccentricity of the load in the finite element model showed that the corbel in the present test conditions could not transfer the desired eccentricity, thereby reducing it. Also, the mean errors of the predicted model compared to the experimental values, the coefficient of variation, and the mean absolute percentage error for the modified fuzzy model compared to the conventional TSK model across all data for the six-law models were 1.004, 13.6%, 6.1% and 0.957, 23.5%, 12.8%, respectively, indicating the superior performance of the modified fuzzy model algorithm.
