بررسی پارامتری عوامل موثر در مدلسازی بتن در مقیاس مزو

مدل‌های عددی راهکاری مقرون به صرفه در مقایسه با تست های میدانی و آزمایشگاهی می‌باشند که در صورت لحاظ نمودن خصوصیات دقیق ماده توانایی پیش بینی رفتار مصالح را با دقت قابل قبولی دارند. مزیت دیگر مدل‌های عددی در مقیاس مزو، قابلیت ارائه رفتار اندرکنش اجزاء مختلف مصالح غیرهمگن مانند بتن می‌باشد که این ویژگی در پیش بینی رفتار بتن در مقابل بارهای دینامیکی، کارآیی بالایی دارد. به منظور انجام مدل‌سازی بتن در مقیاس مزو پلتفرمی برای ایجاد نمونه‌های بتنی با درنظرگیری سنگدانه، ملات و الیاف ایجاد گردید. پلتفرم ایجاد شده در این پژوهش قادر است سنگدانه‌های بیضوی مطابق با منحنی دانه‌بندی دلخواه به صورت کاملا تصادفی تولید و با کنترل عدم تداخل سنگدانه‌ها در کنار یکدیگر به صورت از پیش تعیین نشده درون مرزهای قالب جایگذاری کند و پس از معین شدن هندسه نمونه و مش‌بندی آن، مدل ساخته‌شده تبدیل به متن ورودی برنامه‌های تحلیل عددی می‌شود و بارگذاری فشاری و کششی بر روی آن انجام می‌گیرد. در تحلیل‌های انجام شده حداکثر مقاومت فشاری و میزان شکل‌پذیری بتن و در مواردی مسیر گسترش ترک مورد بررسی قرار گرفته است.

parametric study of effective factors in meso-scale concrete modeling

numerical models are cost-effective solutions compared to field and laboratory tests that have the ability to predict the behavior of materials with acceptable accuracy if the exact properties of the material are taken into account. another advantage of numerical models on the meso-scale is the ability to provide the interaction behavior of different components of inhomogeneous materials such as concrete, which is highly effective in predicting the behavior of concrete against dynamic loads. in order to model concrete at the meso-scale, a platform was created to create concrete samples considering aggregates, mortars, and fibers. the platform created in this research can produce elliptical aggregates according to the desired granulation curve in a completely random manner. by controlling the non-interference of aggregates next to each other stochastically placed within the mold boundaries and after determining the sample geometry and grid, the constructed model is converted into input text of numerical analysis programs. compression and tensile loading are performed on it. using numerical models, the effect of changes in these parameters was investigated: the volume ratio of aggregates in the sample, the allowable elongation of elliptical aggregates, whether or not to consider the itz element, and the use of fibers. in the performed analyzes, the maximum compressive strength and ductility of concrete and, in some cases, the crack propagation path have been investigated. results show that considering both the maximum strength and the area under the stress-strain diagram; it can be concluded that among the percentages of aggregate volume ratios to the sample equal to 65, 70, 75, and 80%, the volume ratio of 75% to the optimal ratio to achieve the best compressive performance concrete is closer.comparison of the shape of the elliptical aggregates in the modeling showed that the closer the aggregate form is to the sphere, the maximum resistance increases, but the strength of the concrete decreases with a steeper slope after the peak.although the analysis of concrete samples at the meso-scale requires much higher computational costs than the macro analysis, this analysis can provide diagrams of tensile and compressive strength of concrete more accurately and also to test new samples with changes in the main parameters acceptable estimation of behavior. provide concrete without the need to build a physical sample. studies have shown that unstructured tetrahedron meshing provides graphs closer to empirical results. in analyses with unstructured mesh, mesh size has minimal effect on the results; in the case of models with a structured cubic mesh, the mesh size severely overshadows the analysis results. suppose the appropriate size for uniform meshing is calibrated using the results of non-structured mesh samples. in that case, results close to the meshes of non-structured samples can be obtained with much less computational time and cost. although considering the itz helps the stress-strain curve at the front of the peak to be closer to the laboratory graphs and also to predict the location of the crack more accurately than without modeling without itz, it causes a decrease in resistance at the rear of the peak and is predicted. this can be partially controlled by calibrating the itz resistance reduction value.