نوع مقاله : مقاله پژوهشی
نویسندگان
1 دانشیار، گروه مهندسی عمران، واحد ارومیه، دانشگاه آزاد اسلامی، ارومیه، ایران
2 مربی، گروه مهندسی عمران، دانشگاه فنی و حرفهای، تهران، ایران
3 مربی، گروه معماری و شهرسازی، دانشگاه فنی و حرفهای، تهران، ایران
چکیده
کلیدواژهها
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Designing a mix for high strength SCC concrete requires heightened precision and strict control over the quantities of constituent materials. Moreover, conducting compressive strength tests for concrete at various ages involves significant time and cost, generates waste, and causes environmental harm. With technological advancements, machine learning is now capable of solving complex problems; specifically, it can estimate the compressive strength of concrete based on the provided mix design. The learning process, however, is divided into various methods and categories depending on the expected output, many of which have become obsolete over time, with more accurate approaches being introduced. In this study, silica fume was incorporated in the production of SCC concrete, and various algorithms were employed to examine the influence of different material ratios on compressive strength. In this regard, Artificial Neural Networks (ANN) and Adaptive Neuro Fuzzy Inference Systems (ANFIS) with 1 to 10 hidden layers were developed, and their results were compared. Findings from the first part of the study revealed that the ANFIS and ANN models with 10 and 6 hidden layers, respectively, achieved higher accuracy than the other models. In the second part, to validate the AI model results, six high strength concrete mix designs with various material proportions were prepared and tested experimentally, and the results were compared with the predictions of the ANFIS and ANN networks. The results indicated that the ANFIS network with ten hidden layers achieved, on average, an accuracy of 96% in estimating the compressive strength of high strength SCC concrete.
کلیدواژهها [English]
)