1. تخمین آبشستگی در مجاورت پایه های پل جفت و سه تایی با استفاده از دسته بندی c-میانیگن فازی شبکه انفیس
- Author
-
افشین کیانی, سعید شعبانلو, and فریبرز یوسفوند
- Abstract
Background and Objectives: scouring is a phenomenon that occurs due to the passage of water flow at the boundaries of contact with other objects in hydraulic structures. The basis of this phenomenon is the creation of a vacuum at the contact boundaries of two environments, due to the water velocity gradient. In general, scouring occurs in the vicinity of various structures such as bridge foundations, submerged spillways, downstream of spillways. With the increase in the dimensions of the scour hole, there is a possibility of overturning and destroying these structures. So that every year significant costs are spent on the repair and reconstruction of various structures, including bridge foundations that have been destroyed due to erosion. Estimation and prediction of scouring around the piers play a significant role to design these structures since with increasing dimensions of scour hole, stability of the pier is threatened; as a result, the structure may be destructed. In this study, scour hole in the vicinity of twin and three piers was estimated by using fuzzy c-means clustering of ANFIS (ANFIS-FCM) network technique. Methodology: Firstly, the parameters affecting scour hole around twin and three piers including Froude number (Fr), the ratio of the pier diameter to the flow depth (D/h), and the ratio of the distance between the piers to the flow depth (d/h) were detected. Subsequently, seven ANFIS-FCM models were defined by means of these dimensional input parameters. It should be stated that 70% of the experimental data were utilized to training the models and 30% of the rest were applied to testing. Then, the superior ANFIS-FCM model and the most important input parameter were introduced by implementing a sensitivity analysis. The premium model as a function of all input parameters simulated the scour values with a reasonable accuracy. Findings: For instance, the correlation coefficient (R), the scatter index (SI), and the Nash-Sutcliff efficiency coefficient (NSC) are respectively computed to be 0.988, 0.106, and 0.976. Furthermore, the Froude number was considered as the most important input parameter. Finally, a computer code was introduced to simulate the scour hole around the twin and three piers. Conclusion: In this study, a neuro-fuzzy technique of the new artificial intelligence method called ANFIS Network Fuzzy C-Mean Classification (ANFIS-FCM) was used to simulate the scour depth in the vicinity of paired and triple bridge foundations. To validate the results of the simulations, 70% of the observed values were used to train the artificial intelligence model and the remaining 30% were used to test it. Then, using the input parameters, seven ANFIS-FCM models were defined, and by analyzing the results Modeling, the best model and the most effective input parameters were introduced. The superior model (ANFIS-FCM 1) simulated scour values with acceptable accuracy. This model estimated scour values according to all input parameters and Fr and D/h parameters were identified as the most effective input parameters. For example, the values of RMSE, MAE and VAF were calculated as 0.025, 0.019 and 97.507 respectively for the test conditions of ANFIS-FCM 1 model (superior model). It should be mentioned that in order to estimate scour depth in the vicinity of double and triple bridge foundations, a computer code was proposed for use in practical work for engineers. One of scour hole, stability of the pier is threatened; as a result, the structure may be destructed. In this study, scour hole in the vicinity of twin and three piers was estimated by using fuzzy c-means clustering of ANFIS (ANFIS-FCM) network technique. Firstly, the parameters affecting scour hole around twin and three piers including Froude number (Fr), the ratio of the pier diameter to the flow depth (D/h), and the ratio of the distance between the piers to the flow depth (d/h) were detected. Subsequently, seven ANFIS-FCM models were defined by means of these dimensional input parameters. It should be stated that 70% of the experimental data were utilized to training the models and 30% of the rest were applied to testing. Then, the superior ANFISFCM model and the most important input parameter were introduced by implementing a sensitivity analysis. The premium model as a function of all input parameters simulated the scour values with a reasonable accuracy. For instance, the correlation coefficient (R), the scatter index (SI), and the Nash-Sutcliff efficiency coefficient (NSC) are respectively computed to be 0.988, 0.106, and 0.976. Furthermore, the Froude number was considered as the most important input parameter. Finally, a computer code was introduced to simulate the scour hole around the twin and three piers. [ABSTRACT FROM AUTHOR]
- Published
- 2024
- Full Text
- View/download PDF