Sharif University of TechnologyScientia Iranica1026-309830620231101Field and laboratory investigations of the effective contact surface between sleeper and ballast particles in railway tracks191119242303810.24200/sci.2022.59574.6315ENH.Refahiat-NikooSchool of Railway Engineering, Iran University of Science and Technology, Tehran, IranJ.-A.ZakeriCenter of Excellence in Railway Transportation, School of Railway Engineering, Iran University of Science and Technology,
Tehran, IranS.MohammadzadehSchool of Railway Engineering, Iran University of Science and Technology, Tehran, IranJournal Article20211215Ballast is a porous layer that reduces the effective contact surface between sleeper and ballast. Also, the sleeper does not have a smooth bottom surface, which causes a reduction of the contact surface between the sleeper and ballast particles. At field experiments, by installing sensitive papers under the sleepers of a track, the cyclic load was imposed by passing trains, and after passing a certain amount of loads, the sensitive pressure paper was removed, and the contact surface was measured while in the laboratory investigation, the ballast box test was used to achieve a similar goal. The values obtained for the contact surface showed that the 200 kN axle load contact area is about 4.9% for concrete sleepers and about 8.1% for wooden sleepers, and this value increases with increasing axle load and load cycle. The average contact area for ballast type 1 was 4.6%, for ballast type 4 was 5.5%, and for the sample of ballast from the field test site was 5.3%. Two types of soft and stiff under sleeper pads (USP) were used. The results showed that the effective contact area between sleeper and ballast was 21.6% and 15.9% for soft and stiff USP, respectively.https://scientiairanica.sharif.edu/article_23038_ace308dd71ad032513fa9674bfe26f7b.pdfSharif University of TechnologyScientia Iranica1026-309830620231101Multi-gene GP and GA-FIS models to deal with scaling problem in the ANFIS model for estimating roughness coefficient in erodible channels192519412308710.24200/sci.2023.60241.6679ENM.ZanganehDepartment of Civil Engineering, Faculty of Engineering, Environmental Hazard Institute, Golestan University, Golestan, IranJournal Article20220407Estimation of the roughness coefficient is important for reliable hydraulic design in erodible channels. In this paper, the capability of multi-gene Genetic Programming (GP), a combined Genetic Algorithm and Fuzzy Inference System (GA-FIS) model, and Multi Regression (MR) methods are employed to estimate the roughness coefficient. These methods try to extract either an explicit or an implicit relationship between the roughness coefficient and input variables. In addition, traditional GP, widely used by researchers, and conventional empirical formulas are implemented to evaluate the models. Results show that the employed methods are more accurate than empirical methods. In addition, the effects of some other parameters, such as non-dimensional water depth and shear Reynolds number, are highlighted over the roughness coefficient while previously ignored in the empirical methods. Also, findings prove that the GA is a helpful tool to optimize a FIS compared with gradient-based models like ANFIS, while the scale of input variables is not in the same order. The R2 for multi-gene GP and GA-FIS are 0.8504 and 0.8842, respectively, while this value for the most accurate empirical method (Yalin, 1992) is 0.6286.https://scientiairanica.sharif.edu/article_23087_a86c3a52836c0ae7189127f9e2b8f676.pdfSharif University of TechnologyScientia Iranica1026-309830620231101Global thin plate spline differential quadrature as a meshless numerical solution for two-dimensional viscous Burgers' equation194219542300010.24200/sci.2022.60247.6685ENA. M.BehrooziDepartment of Civil Engineering, Persian Gulf University, Shahid Mahini St., Bushehr, P.O. Box 75169, Iran0000-0002-7663-8727M.VaghefiDepartment of Civil Engineering, Persian Gulf University, Shahid Mahini St., Bushehr, P.O. Box 75169, IranJournal Article20220410This paper is aimed to present the Global Thin Plate Spline Differential Quadrature method for the numerical solution of viscous Burgers’ equation. This mesh-less and high-order model is introduced with the motive of diminishing computational effort and dealing with irregular geometries. Thin Plate Spline Radial basis function is used as a test function to determine coefficients of derivatives in differential quadrature. The present algorithm is applied to discretize and solve two-dimensional Burgers’ equation in both rectangular and irregular non-rectangular computational domains with randomly distributed computation nodes. To evaluate the capability of the present model, several problems with different boundary and initial conditions and Reynolds Numbers are solved and the obtained results are compared with the analytical solutions and other previous numerical models. The obtained results show the higher accuracy of the present model for solving Berger's equation with fewer computational nodes compared to the previous models even in irregular domains.https://scientiairanica.sharif.edu/article_23000_a705657b70511a8af8430fd36224698e.pdf