چکیده:
مورفولوژی رودخانه، علم شناخت سامانه رودخانه از نظر شکل و شکل مسطحِ، مشخصههای هیدرولیکی، راستا و نیمرخ طولی بستر و روند و سازوکار تغییرات این مشخصات است. از طریق بررسی ریختشناسی رودخانه، میتوان شرایط کنونی و توان تغییرات احتمالی آن را در آینده بهتر درک کرد. خصوصیات مورفولوژی با گذشت زمان تغییر میکند و تحت تأثیر عواملی همچون بده و سرعت جریان، میزان انتقال و خصوصیات رسوب، جنس مواد تشکیلدهنده بستر و کنارهها، شرایط زمینشناسی و عوامل دیگر است. از طرفی، سازههای هیدرولیکی انواع مختلفی دارد و استفاده از آنها به نوع منبع آب و هدف مورد استفاده وابسته است. در هر صورت، به طور کلی از مهمترین سازههای هیدرولیکی میتوان به پلها اشاره کرد. احداث پلها در مسیر رودخانهها به بروز تغییرات در رفتار مورفولوژیکی آنها منجر میشود؛ بنابراین، در این پژوهش با توجه به اهمیت بررسی ساخت سازههای هیدرولیکی بر بستر رودخانهها، مدلسازی تأثیر این سازهها بر رفتار مورفولوژی رودخانه شلمانرود در استان گیلان بررسی شد. بهمنظور ایجاد مدل، در ابتدا الگوریتم به آنالیز دادههای ارائهشده میپردازد تا انواع خاصی از الگوها یا روندها را جستجو کند. سپس از نتایج این آنالیز بهدفعات استفاده میشود تا پارامترهای مطلوب برای ایجاد مدل دادهکاوی حاصل شود. همچنین برای مقایسه روش مورد استفاده در این پژوهش، از روش دادهکاوی استفاده شد. به طور کلی، نتایج نشان میدهد که استفاده از الگوریتم دادهکاوی اپریوری در مدلسازی تأثیر سازههای هیدرولیکی احداث شده بر بستر دو رودخانه شلمانرود و پلرود روشی مناسب میباشد؛ زیرا هر سه شاخص دقت، حساسیت و ویژگی این الگوریتم از الگوریتم وکا بالاتر است.
1- Introduction
Changes in river pattern is one of the most important issues of river engineering that affects the activities and structures of the river bank. It is important to study the morphological changes of river channels in order to find appropriate control solutions to solve the dynamic problems of these areas. The morphology of the riverbed pattern over time is a function of various factors such as geological formations, flood discharge, changes caused by human factors, vegetation, topography and tectonic movements. Statistical studies of river patterns have presented the morphology of a number of geostatistical rules. In this regard, the sinusoidal pattern of the riverbed depends on the dominant role of processes and a set of factors that are applied over time. A hydraulic structure refers to a structure in which all or part of the body of water is in contact with water in a way that changes the natural flow of water. These structures are used for purposes such as transmission, energy dissipation and flow cessation. There are different types of hydraulic structures and their use depends on the type of water source and the purpose used. In any case, in general, stair is one of the most important hydraulic structures. The construction of bridges in the course of rivers causes changes in the morphological behavior of rivers.
2- Methodology
The two most important rivers in Gilan province are Plorud and Shalmanrud located in the east of Gilan. Data mining algorithm was used to model the effect of hydraulic structures. The algorithm determines the method used to search for the pattern in the data and is, in fact, like a mathematical procedure for solving a particular problem. Data mining algorithms refer to a set of inferences and calculations that provide a model of the data. In order to create a model, the algorithm first analyzes the presented data to search for specific types of patterns or trends. It then uses the results of this analysis several times to achieve the desired parameters to create a data mining model. In the next step, these parameters are used to extract accurate operational patterns and statistical processes in the entire data set. One of the most efficient algorithms used in the data mining process is the use of a priori algorithm. For modeling through a priori algorithm, six parameters of distance from bottom to bottom, cross-sectional area, full cross-section width, maximum full cross-sectional depth, average full cross-sectional depth and width-to-depth ratio were used. Another algorithm used to investigate the effect of constructed structures on the morphological behavior of measured rivers is Veca algorithm. The Weka algorithm consists of a set of methods such as categorization, clustering, and association rules and feature selection. In categorization, each data is assigned to a predefined class, but in clustering there is no information about the classes in the data. Therefore, considering the importance of investigating construction of hydraulic structures on riverbeds, modeling the effect of hydraulic structures on the morphological behavior of Shalmanrud River in Guilan province will be evaluated in this study.
3- Result
Of the 53 rules obtained, 12 were more attractive and accurate based on the data. Comparison of the rules obtained by using the a priori algorithm showed that among the six indicators measured in data mining operations, the indicators of maximum depth of section with average confidence of 92% and 25% of the participation rates in the extractive laws are in the first place; width to depth ratio with an average confidence of 88% and 24% participation rate in extractive laws is in the second rank; cross-sectional area index with an average confidence of 85% and 21% participation rate in extraction laws is in the third place; full cross-sectional index with an average 79% confidence and 15% participation rate in extractive laws is in the fourth rank; average depth of full section with an average confidence of 77% and 10% participation rate in extraction laws is in the fifth rank and downstream distance index with an average confidence of 75% 55 Percentage of participation in extraction laws is in the sixth place. Based on this algorithm, ten rules were extracted from the data set used. The first rule with an accuracy of 0.883 includes river morphology, width-to-depth ratio, cross-sectional area and maximum full cross-sectional depth. The second law with an accuracy of 0.867 includes river morphology, cross-sectional area, full cross-sectional width and distance from the downstream. The third rule with 0.769 accuracy includes river morphology, cross-sectional area, full cross-sectional width and maximum full cross-sectional depth. Based on the results obtained from the accuracy and usefulness of the indicators, the width-to-depth ratio index has the most beneficial mode in the extraction rules.
4- Discussion and Conclusion
In order to determine which of the two tested algorithms is more effective, the accuracy, sensitivity and specificity of the two algorithms used were measured along with the accuracy, sensitivity and specificity that can be used for each of the available algorithms. These indices were calculated to determine the accuracy of the classification for each of the categories. In fact, this criterion indicates the success rate of the classifier method in identifying samples related to each category. The call rate with the attribute, which is calculated as the a priori criterion for each of the available categories, shows the percentage of reliability of the output of the classifier method. In general, the results indicated the appropriateness of using the a priori data mining algorithm in modeling the impact of hydraulic structures on the bed of the two rivers Shalmanerood and Plerood because all three indicators of accuracy, sensitivity and specificity of this algorithm were higher than the Weka algorithm.