پیش‌بینی سختی خاک در پایۀ پل‌ها با در نظرگیری اندرکنش خاک-سازه با استفاده از مدل شبکه عصبی مصنوعی

بحث دربارۀ مدل سازی اندرکنش خاک-گروه شمع به دلیل تعداد زیاد پارامترهای دخیل در آن یکی از مباحث پیچیده و البته مورد توجه محققان در سال های اخیر بوده است که به روش های مختلف به آن پرداخته اند. در سال های اخیر روش شبکۀ عصبی مصنوعی در بسیاری از مسائل مرتبط با مهندسی ژئوتکنیک از جمله مسائل مرتبط با شمع ها استفاده شده است. در این پژوهش پاسخ دینامیکی گروه شمع خاک زیر یک پل تک پایه با درنظرگیریاثر اندرکنش خاک سازه به روش تحلیلی محاسبه شد . به علاوه یک مدل شبکه عصبی با استفاده از مدل چند لایه پیشخور برای پیش بینی سختی اندرکنشی گروه شمع-خاک واقع در خاک دانه ای در نظر گرفته شد. هم چنین تاثیر پارامترهای مختلف بر محاسبۀ مقدار سختی اندرکنشی گروه شمع-خاک بحث و بررسی شد. در نهایت با توجه به مقدار ضریب تبیین 0/99 و کم ترین میزان مربعات 1 مدل شبکه عصبی ارائه شده می تواند / خطای 0/00000088 یک ابزار قدرتمند برای پیش بینی میزان سختی اندرکنشی گروه شمع-خاک و طراحی بهینۀ گروه شمع برای دست یابی به طول و قطر بهینۀ شمع ها و کاهش هزینۀ مربوط به تجهیزات حفاری باشد.

Estimation of Soil-Pile Stiffness under the Bridge Piers Considering Soil-Structure Interaction using Artificial Neural Network Model

IntroductionLandslide is known as one of major natural hazards. Landslide susceptibility mapping is known as efficient approach to mitigate the future hazard and reduce the impact of landslide hazards. The main objective of this research is to apply GIS spatial decision making systems for landslide hazard mapping in the 5th segment of ArdebilMianeh railroad. Evaluation of the landslide criteria mapping and their relevancy for landslide hazard can be also considered. To achieve the research objectives, an integrated approach of FuzzyAnalytic Hierarchy Process (AHP), Fooler Hierarchical Triangle and Fuzzy logic methods were employed in GIS Environment.Material and methodsWithin this research, we also aimed to apply GIS spatial decision making systems and in particular GIS multi criteria decision analysis which are available in Arc GIS and Idrisi softwares. We have identified 8 casual factors (including: density of vegetation, land use, faults desistance, distance from rivers, distance from roads, slope, aspect, geology) based on literature review. Accordingly, these layers were prepared in GIS dataset by means of applying all GIS ready, editing and topology steps. The criterion weighting was established based FAHP approach. The criteria weights was derived and rank of each criterion was obtained. Accordingly, the landslide susceptible zones were identified using GISMCDA approaches.Results and discussionFinally the functionality of each method was validated against known landslide locations. This step was applied to identify most efficient method for landslide mapping. According to the results and based on the values derived from Qs, P, and AUC, the accuracy of fuzzy method was accordingly about 0.33, 0.74 and 0.76, respectively. In context of FuzzAHP the accuracy of 1.08, 0.88 and 0.94 were obtained. While, the accuracy of Fooler Hierarchical Triangle were obtained 0.78, 0.84 and 0.91, accordingly.ConclusionAs results indicated integration of FuzzyAHP represented more accurate results. Results of this research are great of important for future research in context of methodological issues for GIScience by means of identifying most efficient methods and techniques for variety of applications such landslide mapping, suitability assessment, site selection and in all for any GISMCDA application.