شناسايي آثار عدم قطعيت پارامترهاي موثر بر تحليل پايداري سواحل چندلايه رودخانه ها در مقابل ساز و كار گسيختگي صفحه اي

Identifying the effects of parameters uncertainty on the stability analysis of composite riverbanks against planar failure mechanism

تحليل پايداري سواحل رودخانه در بسياري از مدل ها، عمدتاً با استفاده از يك ضريب اطمينان به صورت نسبت نيروهاي محرك به مقاوم موثر بر سطح گسيختگي، صورت مي گيرد. اين نيروها توسط مجموعه اي از عوامل كنترل كننده شامل مشخصات هندسي، ژئوتكنيكي و هيدرولوژيكي قرار دارند، تحت تاثير قرار مي گيرند. در اين پژوهش، آثار عدم قطعيت موجود در تعيين پارامترهاي موثر بر تحليل پايداري سواحل چندلايه رودخانه ها از نوع چسبنده بر اعتبار نتايج تحليل پايداري، ارزيابي شده است. بدين منظور از مدل تحليل پايداري سواحل غيرهمگن رودخانه ها BSTEM استفاده شده و آثار تغييرات هر پارامتر مجزا بر ضريب اطمينان پايداري ساحل در مقابل ساز و كار گسيختگي صفحه اي در شرايطي كه مقدار ساير پارامترها ثابت باشد، بررسي شده است. ساحل منتخب مورد بررسي از سواحل در معرض عقب نشيني رودخانه آرنو ايتاليا برگزيده شد. عدم قطعيت هاي موجود در تخمين پارامترها نيز با توجه به دو منبع اصلي خطا شامل خطاهاي اندازه گيري و خطاهاي ناشي از تغييرپذيري طبيعت ذاتي هر پارامتر مدنظر قرار گرفت. نتايج اين تحقيق نشان داد كه به دليل بزرگ بودن عدم قطعيت هاي ناشي از تغييرپذيري طبيعت ذاتي پارامترها، احتمال بروز خطا در تخمين ها بسيار زياد است. در اين ميان عدم قطعيت پارامترهاي ارتفاع، زاويه ساحل، چسبندگي مصالح و رقوم سطح ايستابي به ترتيب داراي درجه اهميت بيشتري هستند، اما ساير پارامترهاي موثر در رابطه، از درجه اهميت كمتر تا بي اهميت برخوردار بودند.

Rivers have not always flowed in the same directions as they do today. They constantly change their route، for various reasons. Therefore، understanding of river pattern changes، especially near cities، roads and coastal infrastructures is the most important issue. Hence، it is essential to assess unstable banks condition by applying riverbank stability analysis models which will lead to better understanding of the movement pattern of river’s course، so its results can be used to accurately model the flow and sediment transport along the river's route. Nevertheless، access to all required data for the riverbank stability analysis along the entire course of river is very difficult and practically impossible due to the complexity of the different mechanisms of bank erosion. In previous studies، evaluation of uncertainty in the analysis of bank stability has been done with the assumption of homogeneity of materials forming the riverbank، while such stream banks are rarely found in nature. The aim of this study is to evaluate the effect of parameter uncertainty arising from the variability of input parameters influencing the stability analysis of composite banks on the reliability of the simulated factor of safety for planar failure. The area selected for this study is located in Arno River، Italy (referred to as the reference bank) and it has been subjected to bank retreat. Among various types of riverbank failure، planar failures are the most common type، being associated with steep، relatively low banks composed of cohesive sediments. Moreover، this type of failure represents the most frequent mechanism observed on fine-grained banks along the Arno River. In this paper، the bank stability and toe erosion model (BSTEM)، developed by Simon et al. (2006) has been used. BSTEM is a steady-state bank stability model capable of analyzing the stability of non-homogenous bank materials that calculates a bank factor of safety for multi-layer stream banks. The model uses two different limit equilibrium algorithms to calculate factor of safety (FS) for multi-layer stream banks using data from the bank geometry، geotechnical properties and hydrological conditions. At the highest level، two sources of uncertainty exist: uncertainty due to variability and uncertainty due to limited knowledge. Uncertainty due to variability first as these uncertainties ultimately combine to contribute to uncertainty due to limited knowledge. Five distinct subclasses of uncertainty due to variability are proposed: inherent natural randomness; value diversity (socio-political); behavioral diversity; societal randomness and technological surprise. Inherent natural randomness is attributed to ‘the nonlinear، chaotic and unpredictable nature of natural processes’. The natural variability of river systems should be a fundamental consideration in integrated river basin management and restoration In these sensitivity analyses، the parameter value ranges (PR) that induce less than a ±15% change in the simulated factor of safety (relative to the reference bank) are highlighted by the grey-shaded areas. The ±15% threshold is arbitrary but is selected to enable direct comparison with the study performed by Samadi et al. (2009) for a homogenous bank. The shaded areas thus delineate the parameter variations (relative to the reference bank) that have a relatively minor influence on the simulated factor of safety. In fact Samadi et al. (2009) used the relationship between Parameter Range (PR) and parameter uncertainty (PU) to determine the overall influence of parameter uncertainty on the reliability of homogenous bank stability modeling. To compare their results to non-homogeneous riverbanks، just these contractual relationships were used. This was achieved by defining four categories to describe each parameter's influence on reliability: (i) insignificant (PU>PR). The sensitivity analyses results and variations in the factor of safety values due to the variation of the effective parameters on bank stability in the aforementioned threshold were considered in this study. As we reported، bank height and bank angle are both very important while the impact of parameters uncertainty of the tension crack depth، river flow depth and matric suction angle are insignificant on the reliability of the bank stability results in both homogeneous and non-homogeneous riverbanks. However، other factors show some differences between the studied composite riverbank (Arno River) results and the results obtained from analysis of homogeneous bank in the study of Goodwin Creek by Samadi et al. (2009). For example، unlike the previous case studies (homogenous bank)، the bank material cohesion in both layers of Arno River is of less importance، seems to be due to the less cohesiveness of bank materials in the current study rather than the previously studied homogeneous bank. Similarly، it has been indicated that soil unit weight is not important and shows this parameter’s insignificant impact on predictive reliability of the calculated factor of safety values. Another important issue is the higher importance and sensitivity of the upper than lower bank to simulate bank stability factor of safety which is completely obvious in the role of the internal friction angle in the upper bank rather than the lower bank and also homogeneous riverbank materials. In addition، groundwater levels are another important parameter under the specific conditions of the present study.