کاربرد مدل‌های فیزیکی برای تعیین ویژگی‌های شکافت و جریان در شکست سدهای خاکریزه‌ای

نویسندگان

چکیده

در پدیدة شکست سد خاکی، تعیین دقیق ویژگی‌های جریان، زمان و شکافت همراه با تحلیل اجزای هیدروگراف در کاهش آسیب‌پذیری دشت‌های سیلابی پایین‌دست و خسارت‌های مالی و جانی آن، نقش بسزایی دارد. در این پژوهش از مدل‌های فیزیکی با خصوصیت‌های مختلف هندسی و مکانیکی برای بررسی سازوکار فرسایش و تکامل شکافت استفاده شد. در تحلیل اجزای هیدروگراف حاصل از شکافت نیز نتایج آزمایشگاهی با داده‌های تاریخی حاصل از مدل ریاضی BREACH مقایسه شد. برای توسعة روابط جدید در تعیین دبی اوج (Qp)، پارامترهایی معرفی شده است که تاکنون در تعیین آن به کار نرفته‌ است. در این زمینه، از داده‌های تاریخی، آزمایشگاهی و شکست فرضی سدهای واقعی، به‌عنوان مطالعة موردی استفاده شد. پژوهش‌ها نشان می‌دهد زمان توسعة شکافت (tf)، وابستگی زیادی به ویژگی‌های هیدرولیکی و شکافت داشته و بر این اساس، روابط جدید بر مبنای آن پیشنهاد شده است. همچنین، روابط جدید معرفی‌شده مربوط به ارتفاع شکافت (Hb) به ارزیابی دقیق‌تر فرایند شکافت در حالت‌های روگذری جریان و رگاب منتهی می‌شود. محدودة تغییرات عرض متوسط شکافت (Bave) برحسب ارتفاع جریان گذرنده از روی شکافت، که در این مطالعه تعریف شده است، دامنة وسیعی از سدهای خاکریزه‌ای را فرامی‌گیرد. نتایج تحلیل‌های آماری نشان داد که ضریب تبیین (R2) برای رابطه‌های پیشنهادی برای تعیین Qp برحسب پارامترهای شکافت و با استفاده از برنامه‌ریزی بیان ژن (GEP) و رگرسیون غیرخطی به‌ترتیب، برابر 0/84 و 0/84 است. این ضریب برای تعیین روابط tf با رگرسیون و برحسب پارامترهای هیدرولیکی و شکافت به‌ترتیب، معادل 0/87 و 0/88 محاسبه شد. مقدار R2 در تعیین رابطه‌های جدید Hb با GEP و برحسب پارامترهای بدنه و مخزن در حالت‌های روگذری جریان و رگاب نیز به‌ترتیب، برابر 0/99 و 0/99 بود.

کلیدواژه‌ها


عنوان مقاله [English]

Application of physical models for determination of breach and flow characteristics in earthfill dams failure

نویسندگان [English]

  • Yousef Hasanzadeh
  • Kamran Kouzehgar
  • Saeid Eslamian
  • Mikaeil Yousefzadeh Fard
  • Alireza Babaeian Amini
چکیده [English]

Climate variations are one of the most important factors affecting water resources. Severe floods and runoff increase the risk of the vulnerability of hydraulic structures. The failure of an embankment dam causes extensive financial, human, and environmental damages. Overtopping and internal erosion are the main causes of the embankment failure. More than 46% of the embankment failures around the world are attributed to overtopping. The breach process during this event is generally divided into two parts: the breach initiation stage; and the breach development stage. In the initial stage, the outflow discharge from the dam is not considerable and includes a small stream along the channel, while in the breach development stage, the outflow and erosion processes are significant. On the other hand, about 48% of failures and accidents affecting embankment dams are related to piping. In an embankment dam failure, accurate determination of flow, time, and breach characteristics along with the analysis of hydrograph components can play an important role in reducing the financial losses and fatality.
Several equations have been established using literature data and field observations to calculate the peak outflow discharge (Qp) as a function of the height of water above the breach (Hw) and stored volume above the breach (Vw) (Costa 1985; Gupta and Singh 2012; Hooshyaripor et al. 2014; MacDonald and Langridge?Monopolis 1984; Pierce et al. 2010). The failure time (tf) has been investigated by various researchers. The tf equations were obtained as a function of the average breach width (Bave) (USBR, 1988), Hw, or their combination (Von Thun & Gillette, 1990). Therefore, the height of breach (Hb) and Vw are also considered as important parameters (Froehlich, 2008). On the other hand, Hb is found to be an important factor affecting the hydraulic and breach characteristics (Dhiman & Patra, 2019; Wang et al., 2020).
In this research, physical models with different geometrical and mechanical properties have been used to investigate the mechanisms of erosion and breach evolution. Therefore, several experimental models were constructed and tested in the laboratory flume at three heights of 0.3, 0.4, and 0.5 m. Therefore, five different soil combinations were used and studied. The breach hydrograph components were analyzed, and the breach characteristics were examined as well as the hydraulic characteristics. The BREACH mathematical model is one of the most common models used to evaluate the breach parameters and output hydrographs in embankment dam failures. This model is developed by the National Weather Service (NWS) and is based on the principles of hydraulics, hydrology, and geotechnics. In overtopping failures, the flow over the crest is calculated by the broad-crest weir formula. While the flow into the pipe is simulated by the orifice equation. The results of the model are compared with the observations of some historical failures, which indicates the higher accuracy among the existing mathematical models. In the present study, the output hydrograph components of the experimental models were also compared with the historical data obtained from the BREACH mathematical model. Gene Expression Programming (GEP) was employed as one of the artificial intelligence methods along with the nonlinear regression to obtain a suitable relationship between the input parameters of the models. In order to evaluate the efficiency of the models, three statistical indices, including the root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), and coefficient of determination (R2) are used for performance assessment of the proposed equations in the present study. To develop new equations for the determination of Qp, there are several input variables, that were not introduced before. Therefore, multiple combinations of datasets including historical, experimental, and the hypothetical failure of real dams have been used in this study. Based on observations, the values of tf are highly correlated with hydraulic and breach characteristics. Therefore, new relations have been proposed based on those parameters. On the other hand, the newly introduced equations related to the Hb can lead to more accurate assessment of the breach process in the overtopping and piping failures.
According to statistical analyses, the values of R2 for the proposed equations of Qp, obtained from GEP, as well as nonlinear regression, were 0.84 and 0.84, respectively, based on the breach parameters. The values of the recent coefficient for the development of tf relations obtained from the regressions were 0.87 and 0.88, respectively, according to the hydraulic and breach characteristics. By application of GEP for the determination of Hb in both overtopping and piping failure cases, the values of R2 were 0.99 and 0.99, respectively, as a function of the embankment and hydraulic characteristics. Soil gradations play an important role in increasing the erosion rate and reducing the tf. Therefore, breach formation in coarse-grained particles will take less time to evolve due to its lower shear stresses compared to fine-grained soils. Similarly, the Bave has a significant effect on the breach output hydrograph and its components. The Bave variations based on Hw defined in this study, could cover a wide range of embankment dam failures.

کلیدواژه‌ها [English]

  • Hydrograph components
  • Erosion mechanism
  • Empirical relationships
  • Artificial Intelligence
  • Regression models
  • Water resources