Civil Engineering / İnşaat Mühendisliği
Permanent URI for this collectionhttps://hdl.handle.net/11147/13
Browse
40 results
Search Results
Now showing 1 - 10 of 40
Article Citation - WoS: 3Citation - Scopus: 3Ensemble and Optimized Hybrid Algorithms Through Runge Kutta Optimizer for Sewer Sediment Transport Modeling Using a Data Pre-Processing Approach(Elsevier, 2023) Safari, Mir Jafar Sadegh; Gül, Enes; Dursun, Ömer Faruk; Tayfur, GökmenUncontrolled sediment deposition in drainage and sewer systems raises unexpected maintenance expenditures. To this end, implementation of an accurate model relying on effective parameters involved is a reliable benchmark. In this study, three machine learning techniques, namely extreme learning machine (ELM), multilayer perceptron neural network (MLPNN), and M5P model tree (M5PMT); and three optimization approaches of Runge Kutta (RUN), genetic algorithm (GA), and particle swarm optimization (PSO) are applied for modeling. The optimization and ensemble hybridization approaches are applied in the modeling procedure. For the case of hybrid optimized models, the ELM and MLPNN models are hybridized with RUN, GA, and PSO algorithms to develop six hybrid models of ELM-RUN, ELM-GA, ELM-PSO, MLPNN-RUN, MLPNN-GA, and MLPNN-PSO. Ensemble hybrid models are developed through coupling the ELM and MLPNN models with the M5PMT algorithm. The data pre-processing approach is applied to find the best randomness characteristic of the utilized data. Results illustrate that the RUN-based hybrid models outperform the GA- and PSO-based counterparts. Although the MLPNN-RUN and MLPNN-M5PMT hybrid models generate better results than their alternatives, MLPNN-M5PMT slightly outperforms MLPNN-RUN model with a coefficient of determination of 0.84 and a root mean square error of 0.88. The current study shows the superiority of the ensemble-based approach to the optimization techniques. Further investigation is needed by considering alternative optimization techniques to enhance sediment transport modeling. © 2023 International Research and Training Centre on Erosion and Sedimentation/the World Association for Sedimentation and Erosion ResearchArticle Citation - WoS: 16Citation - Scopus: 19Comparative Analysis of Estimation of Slope-Length Gradient (ls) Factor for Entire Afghanistan(Taylor & Francis, 2023) Ansari, Ahmad; Tayfur, GökmenSlope length gradient (LS) is one of the crucial factors in the Universal Soil Loss Equations (USLE, RUSLE). This study aimed at estimating the slope-length and slope-steepness (LS) factor for the entire watersheds of Afghanistan by using three different methods, namely; (1) LS-TOOLMFD (Method 1); (2) The Method of Equations (Method 2); and (3) The approach of Moore and Burch (Method 3). The first method uses the digital elevation model (DEM) in the ASCII format, and the other two methods use the DEM in the spatial domain. The results show that the LS-factor of the study area ranges from 0.01 to 44.31, with a mean of 5.24 and standard deviation of 6.95, according to Method 1; 0.03 to 163.49, with a mean of 9.6 and standard deviation of 13.58, according to Method 2; and 0 to 3985, with a mean of 7.16 and standard deviation of 29.7, according to Method 3. The study reveals that Methods 1 and 2 are more appropriate than Method 3 because Method 3 yields high LS-factor values close to or at streamlines located near mountainous regions. The highest LS values are found to be in the northeast, north, and central regions of Afghanistan, which is consistent with the high mountains and deep valley geomorphology, indicating that these regions are particularly vulnerable to soil erosion by rainfall-runoff processes. The sediment delivery ratio (SDR) for the Upper-Helmand River Basin (Upper-HRB) is also estimated by the RUSLE, employing the LS factors produced by the three methods. The results revealed that the average annual soil loss is found to be, respectively, 9.3, 18.2, and 11.1 (ton/ha/year) by using the three methods, corresponding to SDR of 23.5%, 12.1%, and 19.9%.Article Citation - WoS: 10Citation - Scopus: 11Experimental Investigation of Sediment Movement as a Result of Homogeneous Earth-Fill Dam Overtopping Break Over a Simplified Urban Area(Elsevier, 2023) Taşkaya, Ebru; Bombar, Gökçen; Tayfur, GökmenWhen an earth-fill dam breaks, dam body sediment and water flow simultaneously move to downstream area causing devastation. Dam break studies in the literature have concentrated mostly on the water flow part while ignoring the sediment movement by designing the dam body as a movable metal gate. This study, however, is the first one experimentally investigating flow and sediment transport due to an earth-fill dam break by constructing the dam body from sediment. Sediment propagation as a result of homogenous earth-fill dam overtopping break was experimentally studied in a laboratory flume of 18.4 m long and 2.0 m wide, and 0.88 m in height in the Hydraulics Laboratory of Izmir Katip Celebi University, Izmir, Turkey. Downstream section right after the dam body was designed as a smooth bed and rough bed. The rough bed, resembling a simplified urban area, was created by using thirteen 10 × 10 × 10 cm sized concrete blocks. The dam body was constructed as homogenous with uniform material having D50 = 0.441 mm. The earth-fill dam body was built using the standard compression methods; each layer of sediment with a thickness of 10 cm was laid in layers, and the body was prepared with a crest width of 10 cm, a transverse base width of 200 cm, a longitudinal base width of 202 cm and height of 60 cm with upstream and downstream slopes of 1:1.6. The water level behind the dam was gradually raised until it overtopped the crest level. A pre-breach was formed at the top of the dam to trigger the break. During each dam break event, water depths were measured by three ULS-40D level meter sensors at different locations, and the final sediment bathymetry map was generated using the ULS-40D Probes at 10 × 10 cm grids. The results showed that, in both smooth and rough downstream bed cases, the dam body eventually collapsed while a great portion of it was carried away by the flood flow. The sediment spreading occurred all over the downstream area, showing significant non-uniform variation in thickness both longitudinally and transversely, especially in the simulated urban area. All the residential areas, while breaking in motion, were submerged under the muddy flow. Some blocks were almost submerged while sediment heights reached half level of some blocks at the end of the experiment. Sediment heights were higher in the urban area.Article Citation - WoS: 13Citation - Scopus: 12Identification of Groundwater Potential Zones in Kabul River Basin, Afghanistan(Elsevier, 2021) Tani, Hamidullah; Tayfur, GökmenGroundwater (GW) plays a vital role in the socio-economic growth of Kabul River Basin (KRB) in Afghanistan. Since the GW resources in the basin have not been properly managed, there is a need for sound strategies by first identifying the potential GW zones. This study assesses the potential groundwater zones for the KRB using the Geographic Information Systems (GIS) and the Analytic Hierarchy Process (AHP). In this direction, seven different thematic maps of rainfall, lithology, land use/land cover, slope, soil, drainage density, and lineament density are first prepared using the GIS. The AHP is then employed to assess the weights of different themes. Finally, the weighted overlay option in the GIS is used to generate the map of the groundwater potential zones (GWPZ). The Very Good zones are mostly located in the downstream and central parts of the KRB, covering around 1543 km(2) area. The Good and the Poor zones are found to be randomly distributed, covering about 39 444 km(2) and 27 658 km(2), respectively. The Very Poor zones are located in the west, southwest, and in some central parts of the basin, covering about 2272 km(2). It is found that only 18% of the total average annual precipitated water of 6.88 x 10(9) m(3)/year infiltrates into the subsurface and ultimately contributes to recharging of the groundwater.Article Citation - WoS: 6Citation - Scopus: 6Kinematic Reverse Flood Routing in Natural Rivers Using Stage Data(Springer, 2022) Tayfur, Gökmen; Moramarco, TommasoIn many developing countries, due to economic constraints, a single station on a river reach is often equipped to record flow variables. On the other hand, hydrographs at the upstream sections may also be needed for especially assessing flooded areas. The upstream flow hydrograph prediction is called the reverse flood routing. There are some reverse flood routing pocedures requiring sophisticated methods together with substantial data requirements. This study proposes a new reverse flood routing procedure, based upon the simple kinematic wave (KW) equation, requiring only easily measurable downstream stage data. The KW equation is first averaged along a channel length at a fixed time, t, assuming that channel width is spatially constant, and then the spatially averaged equation is averaged in time, Δt. The temporally averaged terms are approximated as the arithmetical mean of the corresponding terms evaluated at time t and t + Δt. The Chezy roughness equation is employed for flow velocity, and the upstream flow stage hydrograph is assumed be described by a two parameter gamma distribution (Pearson Type III). The spatially averaged mean flow depth and lateral flow are related to the downstream flow stage. The resulting routing equation is thus obtained as a function of only downstream flow stage, meaning that the method mainly requires measurements of downstream flow stage data besides the mean values of channel length, channel width, roughness coefficient and bed slope. The optimal values of the parameters of reverse flood routing are obtained using the genetic algorithm. The calibration of the model is accomplished by using the measured downstream hydrographs. The validation is performed by comparing the model-generated upstream hydrographs against the measured upstream hydrographs. The proposed model is applied to generate upstream hydrographs at four different river reaches of Tiber River, located in central Italy. The length of river reaches varied from 20 to 65 km. Several upstream hydrographs at different stations on this river are generated using the developed method and compared with the observed hydrographs. The method predicts the time to peak with less than 5% error and peak rates with less than 10% error in the short river reaches of 20 km and 31 km. It also predicts the time to peak and peak rate in other two brances of 45 km and 65 km with less than 15% error. The method satisfactorily generates upstream hydrographs, with an overall mean absolute error (MAE) of 42 m3/s.Article Citation - WoS: 8Citation - Scopus: 9Empirical, Numerical, and Soft Modelling Approaches for Non-Cohesive Sediment Transport(Springer, 2021) Tayfur, GökmenThis paper reviews the modelling approaches and outstanding issues with regard to non-cohesive sediment transport which has been experimentally and numerically studied for many decades owing to its importance to hydraulic structures, morphology and related areas. About 311 papers are reviewed that included laboratory experiments, field observations, and analytical and numerical modelling studies. The reviewed papers cover the period 1938-2020. Of 311, 95 papers are included in this paper. The modeling approaches include empirical, physics-based, spatially averaged, and soft methods. The empirical models have oversimplified the process while the physics-based models are indispensable when the detailed analysis is required. On the other hand, when the objective is to obtain cumulative sediment loads, it would be advantageous to employ the spatial averaging modelling and/or the soft computing methods due to less computational burden and data requirements. The outstanding issues are related to the particle fall velocity, particle velocity, incipient motion, and transport function that require further experimental investigations especially for unsteady non-uniform transport processes.Article Citation - WoS: 27Citation - Scopus: 27Discrepancy Precipitation Index for Monitoring Meteorological Drought(Elsevier, 2021) Tayfur, GökmenWidely employed precipitation drought indices, one way or another, impose probability distribution functions to the data when performing the drought analysis. This may be a plausible approach when the data do not have strong discrepancy which can impede the distribution. The precipitation data in semi-arid and especially in arid regions do have a strong discrepancy due to the sporadic rainfall occurring in such regions. Therefore, in the analysis of the drought for such regions, imposing any probability distribution function to the data could be futile. This study hence developed a new drought index called the Discrepancy Precipitation Index (DPI) for assessing and monitoring the meteorological drought. The method does not impose any probability distribution on the precipitation data. The method is based on the discrepancy of the data with respect to the mean value. The drought classifications are proposed based on the D-score values. Its drought classification ranges are straightforward as those of the Standard Precipitation Index (SPI). The method is applied to assess the meteorological drought at several stations located at different climatic regions such as the arid climate (Mauritania), semi-arid climate (Afghanistan) and the Mediterranean climate (Turkey). The results reveal that the DPI is more representative drought assessment tool for the arid climate regions. At semi-arid climate regions, the DPI can be an alternative drought index to the widely employed (the log-SPI and/or the gamma-SPI) indices. For the Mediterranean climate regions, the DPI can be used together with the other indices. The Discrepancy Measure (DM) is introduced to assess the strength of the discrepancy of the precipitation data series. As the DM of a precipitation series increases, the DPI captures more historical droughts.Article Citation - WoS: 13Citation - Scopus: 14Soil Erosion Model Tested on Experimental Data of a Laboratory Flume With a Pre-Existing Rill(Elsevier Ltd., 2020) Aksoy, Hafzullah; Gedikli, Abdullah; Yılmaz, Murat; Eriş, Ebru; Ünal, N. Erdem; Yoon, Jaeyoung; Tayfur, GökmenPrediction of sediment discharge transported within flow is strongly needed in order to provide measures for a well-established erosion control and water quality management practice. Initiated by runoff generation and erosion processes sediment transport is influenced by microtopography over hillslopes of hydrological watersheds. Consideration of microtopography provides more accurate results. In this study, a process-based two-dimensional rainfall-runoff mathematical model is coupled with erosion and sediment transport component. Both the rainfall-runoff and sediment transport components make simulations in rills and over interrill areas of a bare hillslope. Models at such fine resolution are rarely verified due to the complexity of rills and interrill areas. The model was applied on a data set compiled from laboratory experiments. Erosion flume was filled with granular sand to replace a bare soil. A longitudinal rill and an interrill area were pre-formed over the soil in the flume before the simulated rainfall exerted on. The flume was given both longitudinal and lateral slopes. The simulated rainfall was changed between 45 mm/h and 105 mm/h and exerted on granular uniform fine and medium sand in the erosion flume with longitudinal and lateral slopes both changing from 5% to 20%. Calibration of the model shows that it is able to produce good results in terms of sedigraphs, which suggest also that the model might be considered an important step to verify and improve watershed scale erosion and sediment transport models.Article Citation - WoS: 1Citation - Scopus: 1Generalized Regression Neural Network and Empirical Models To Predict the Strength of Gypsum Pastes Containing Fly Ash and Blast Furnace Slag(Springer Verlag, 2020) Erdem, Tahir Kemal; Cengiz, Okan; Tayfur, GökmenGypsum is widely used in constructions owing to its easy application, zero shrinkage, and excellent fire resistance. Several parameters can affect the properties of gypsum pastes. To study the strength of the gypsum pastes experimentally by trying all these parameters is time-consuming and costly. Therefore, artificial intelligence methods can be very useful to predict the paste strength, which, in turn, can reduce the number of trial batches. Based on experimental data, the generalized regression neural network (GRNN) and empirical models were developed to predict strength of gypsum pastes containing fly ash (FA) and blast furnace slag (BFS). Gypsum content, pozzolan content, curing temperature, curing duration, and testing age constituted the input variables of the models while the paste strength was the target output. The trained and tested GRNN model was found to be successful in predicting strength. Sensitivity analysis by the GRNN model revealed that the curing duration and temperature were important sensitive parameters. In addition to the GRNN model, empirical models were proposed for the strength prediction. The same input variables formed the input vectors of the empirical models. The same dataset used for the calibration of the GRNN model was employed to establish the empirical models by employing genetic algorithm (GA) method. The empirical models were successfully validated. The GRNN and GA_based empirical models were also tested against the multi-linear regression (MLR) and multi-nonlinear regression (MNLR) models. The results showed the outperformance of the GRNN and the GA_based empirical models over the others.Article Citation - WoS: 31Citation - Scopus: 34Data Pre-Post Processing Methods in Ai-Based Modeling of Seepage Through Earthen Dams(Elsevier Ltd., 2019) Sharghi, Elnaz; Nourani, Vahid; Behfar, Nazanin; Tayfur, GökmenIn this paper, seepage of Sattarkhan earthen dam in northwest Iran was simulated using various artificial intelligence (AI) models (e.g., Feed forward neural network, Adaptive neural fuzzy inference system and Support vector regression) and linear ARIMA model based on different input combinations. Both jittering pre-processing and ensembling post-processing methods were also used in order to enhance the performance of the used AI-based data driven methods. For this purpose, various jittered datasets were produced by imposing noises (at different levels) to the original time series to enlarge the training data sample space. Further, three techniques of simple linear, weighted linear and nonlinear neural averaging were considered for pre-post processing purpose. The obtained results indicated that using both jittering and ensembling (especially neural ensemble) enhanced the modeling performance by almost 30% in the testing phase. (C) 2019 Elsevier Ltd. All rights reserved.