نوع مقاله : مقاله پژوهشی
نویسندگان
1 گروه مهندسی عمران، دانشگاه بیرجند، بیرجند، ایران
2 گروه مهندسی عمران، دانشکده مهندسی، دانشگاه بیرجند، بیرجند، ایران
3 گروه مهندسی عمران، دانشکده مهندسی، گروه پژوهشی فناوریهای نوین در مهندسی عمران، دانشگاه بیرجند، بیرجند، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Inorganic matrix composites are considered an advanced method for structural retrofitting due to their low weight, high strength, rapid instal-lation, and environmental compatibility. This study evaluates the confine-ment behavior of circular concrete columns wrapped with inorganic ma-trix composites. A database of 92 experimental specimens was compiled, considering key input parameters: column diameter, number of confine-ment layers, fiber elastic modulus, fiber thickness, ultimate strain, and unconfined compressive strength. Existing empirical models were found inadequate for accurately predicting strength enhancement. Consequently, an artificial neural network (ANN) with nine hidden neurons was devel-oped, achieving high prediction accuracy (R = 0.9952 and 0.9990; MSE = 0.000187 and 0.0000069; MAPE = 1.6871% and 0.6795% for training and testing, respectively). Among previous models, Triantafillou et al.’s formula showed the best performance. Sensitivity analysis quantitatively indicated that unconfined compressive strength (fco) has the greatest in-fluence on strength enhancement (92.20%), while the number of confine-ment layers has the least influence (5.81%). The proposed ANN model provides a reliable tool for accurate prediction and practical design of inorganic matrix composites -confined concrete columns.
کلیدواژهها [English]
ارسال نظر در مورد این مقاله