Civil Engineering / İnşaat Mühendisliği
Permanent URI for this collectionhttps://hdl.handle.net/11147/13
Browse
12 results
Search Results
Article Citation - WoS: 2Citation - Scopus: 2Trait-based heterogeneous populations plus (TbHP+) genetic algorithm(Elsevier Ltd., 2009) Tayfur, Gökmen; Sevil, Hakkı Erhan; Gezgin, Erkin; Özdemir, SerhanThis study developed a variant of genetic algorithm (GA) model called the trait-based heterogeneous populations plus (TbHP+). The developed TbHP+ model employs a memory concept in the form of immunity and instinct to provide the populations with a more efficient guidance. Also, it has an ability to vary the number of individuals during the search process, thus allowing an automatic determination of the size of the population based on the individual qualities such as character fitness and credit for immunity. The algorithm was tested against the classical GA model in convergence and minimum error performance. For this purpose, 5 different mathematical functions from the literature were employed. The selected functions have different topological characteristics, ranging from simple convex curves with 2 variables to complex trigonometric ones having several hilly shapes with more than 2 variables. The developed model and the classical GA model were applied to finding the global minima of the functions. The comparison of the results revealed that the developed TbHP+ model outperformed the classical GA in faster convergence and minimum errors, which may be explained by the adaptive nature of the new paradigm.Article Citation - WoS: 21Citation - Scopus: 25Predicting Hourly-Based Flow Discharge Hydrographs From Level Data Using Genetic Algorithms(Elsevier Ltd., 2008) Tayfur, Gökmen; Moramarco, TommasoThis study developed a genetic algorithm model to predict flow rates at sites receiving significant lateral inflow. It predicts flow rate at a downstream station from flow stage measured at upstream and downstream stations. For this purpose, it constructed two different models: First is analogous to the rating curve model (RCM) of Moramarco et al. [Moramarco, M., Barbetta, S., Melone, F., Singh, V.P., 2005. Relating local stage and remote discharge with significant lateral inflow. J. Hydrologic Eng., ASCE, 10(1)] and the second is based on summation of contributions from upstream station and lateral inflows using kinematic wave approximation. The model was applied to predict flow rates at three different gauging stations located on Tiber River, Upper Tiber River Basin, Italy. The model used average wave travel time for each river reach and obtained average set of parameter values for all the events observed in the same river reach. The GA model was calibrated, for each river reach and for each formulation, by three events and tested against three other events. The results showed that the GA model produced satisfactory results and it was superior over the most recently developed rating curve method. This study further analyzed the case where only water surface elevation data were used in the input vector to predict flow rates. The results showed that using elevation data produces satisfactory results. This has an implication for predicting flow rates at ungauged river sites since the surface elevation data can be obtained without needing the detailed geometry of river section which could change significantly during a flood.Article Citation - WoS: 81Citation - Scopus: 97Evaluation of Natural Zeolite as a Viscosity-Modifying Agent for Cement-Based Grouts(Elsevier Ltd., 2008) Şahmaran, Mustafa; Özkan, Necati; Keskin, Süleyman Bahadır; Uzal, Burak; Yaman, İsmail Özgür; Erdem, Tahir KemalThe effects of natural zeolite on the rheological and workability properties of the grout mixtures were studied. Setting times of grouts were also determined as part of the experimental study. For comparison, grout mixtures were also prepared with a commercially available viscosity modifying admixture (VMA). The experimental results show that addition of natural zeolite modifies both the rheological and workability properties of grouts. For a constant superplasticizer (SP) content, an increase in the zeolite amount significantly increases the yield stress, the apparent and plastic viscosity, and reduces the fluidity and deformability. Moreover, an increase in the amount of SP causes a significant reduction in both the yield stress and plastic viscosity of the grouts. It was also observed that, grouts prepared with natural zeolite addition have a pseudo-plastic behavior, and shear-thinning behavior increases with an increase in the zeolite amount. Therefore, it has been shown that using natural zeolite as a VMA it is possible to obtain grouts that have satisfactory rheological properties, especially if natural zeolite is used in combination with a superplasticizer.Article Citation - Scopus: 31Spatial Interaction Modeling of Interregional Commodity Flows(Elsevier Ltd., 2007) Çelik, Hüseyin Murat; Guldmann, Jean-MichelDrawing from both the spatial price equilibrium theoretical framework and the empirical literature on spatial interaction modeling, this paper expands models of interregional commodity flows (CFs) by incorporating new variables and using a flexible Box-Cox functional form. The 1993 US commodity flows survey provides the empirical basis for estimating state-to-state flow models for 16 commodity groups over the 48 continental US states. The optimized Box-Cox specification proves to be superior to the multiplicative one in all cases, while selected variables provide new insights into the determinants of state-to-state CFs.Article Citation - WoS: 23Citation - Scopus: 24Modelling Sediment Transport From Bare Rilled Hillslopes by Areally Averaged Transport Equations(Elsevier Ltd., 2007) Tayfur, GökmenTreating the dynamics of sediment transport as two-dimensional on interrill-areas and as one-dimensional in rill sections, areally averaged sheet sediment transport equations are developed. The two-dimensional sheet sediment transport equation is averaged over an individual interrill-area width and then along the interrill-area length to obtain local-scale areally averaged interrill-area sheet sediment transport equation (local-scale areal averaging). Similarly, the cross-sectionally-averaged rill sediment transport equation is averaged along an individual rill length to obtain local-scale areally averaged rill sediment transport equation (local-scale areal averaging). In order to minimize computational effort and economize on the number of model parameters, the local-scale areally averaged equations are then averaged over a whole hillslope section (large-scale areal averaging). These equations constitute the areally averaged model. The expectations of the terms containing more than one variable are obtained by the method of regular perturbation. In the large-scale areal averaging it is assumed that all the randomness in the state variable is due to the randomness in the parameters of the process. Comparison of the results obtained from the areally averaged model with those of the point-scale model indicates that the areally averaged model uses far less data and yet it performs as well as the point-scale model. The results of the developed model indicate that on a rilled-surface most of the sediment loads comes from rill sections. The developed model is successfully tested against experimental data obtained from a bare rilled hillslope. It predicted measured runoff and sediment rates with mean absolute errors of 11.07 l/min and 0.382 kg/s, respectively.Article Citation - WoS: 53Citation - Scopus: 63Artificial Neural Network (ann) Prediction of Compressive Strength of Vartm Processed Polymer Composites(Elsevier Ltd., 2005) Seyhan, Abdullah Tuğrul; Tayfur, Gökmen; Karakurt, Murat; Tanoğlu, MetinA three layer feed forward artificial neural network (ANN) model having three input neurons, one output neuron and two hidden neurons was developed to predict the ply-lay up compressive strength of VARTM processed E-glass/ polyester composites. The composites were manufactured using fabric preforms consolidated with 0, 3 and 6 wt.% of thermoplastic binder. The learning of ANN was accomplished by a backpropagation algorithm. A good agreement between the measured and the predicted values was obtained. Testing of the model was done within low average error levels of 3.28%. Furthermore, the predictions of ANN model were compared with those obtained from a multi-linear regression (MLR) model. It was found that ANN model has better predictions than MLR model for the experimental data. Also, the ANN model was subjected to a sensitivity analysis to obtain its response. As a result, the ANN model was found to have an ability to yield a desired level of ply-lay up compressive strength values for the composites processed with the addition of the thermoplastic binder.Article Citation - Scopus: 33Modeling Freight Distribution Using Artificial Neural Networks(Elsevier Ltd., 2004) Çelik, Hüseyin MuratStudies about freight distribution modeling are limited due to the limitations in data availability. Existing studies in this subject, generally either use the conventional gravity models or the regression based models as modeling techniques. The present study, using the 1993 US Commodity Flow Survey Data, models inter-regional commodity flows for 48 continental states of the US with three different artificial neural networks (ANN). The results are compared with those of Celik and Guldmann's (2002) Box-Cox Regression Model. The ANN using conventional gravity model variables provides a slight improvement with respect to this Box-Cox model. However, the ANNs using theoretically relevant variables provide surprising improvements in comparison to the Box-Cox model. It is concluded that ANN architecture is a very promising technique for predicting short-term inter-regional commodity flows.Article Citation - WoS: 175Citation - Scopus: 203Fuzzy Logic Model for the Prediction of Cement Compressive Strength(Elsevier Ltd., 2004) Akkurt, Sedat; Tayfur, Gökmen; Can, SeverA fuzzy logic prediction model for the 28-day compressive strength of cement mortar under standard curing conditions was created. Data collected from a cement plant were used in the model construction and testing. The input variables of alkali, Blaine, SO3, and C3S and the output variable of 28-day cement strength were fuzzified by the use of artificial neural networks (ANNs), and triangular membership functions were employed for the fuzzy subsets. The Mamdani fuzzy rules relating the input variables to the output variable were created by the ANN model and were laid out in the If-Then format. Product (prod) inference operator and the centre of gravity (COG; centroid) defuzzification methods were employed. The prediction of 50 sets of the 28-day cement strength data by the developed fuzzy model was quite satisfactory. The average percentage error levels in the fuzzy model were successfully low (2.69%). The model was compared with the ANN model for its error levels and ease of application. The results indicated that through the application of fuzzy logic algorithm, a more user friendly and more explicit model than the ANNs could be produced within successfully low error margins.Article Citation - WoS: 135Citation - Scopus: 157The Use of Ga-Anns in the Modelling of Compressive Strength of Cement Mortar(Elsevier Ltd., 2003) Akkurt, Sedat; Özdemir, Serhan; Tayfur, Gökmen; Akyol, BurakIn this paper, results of a project aimed at modelling the compressive strength of cement mortar under standard curing conditions are reported. Plant data were collected for 6 months for the chemical and physical properties of the cement that were used in model construction and testing. The training and testing data were separated from the complete original data set by the use of genetic algorithms (GAs). A GA-artificial neural network (ANN) model based on the training data of the cement strength was created. Testing of the model was also done within low average error levels (2.24%). The model was subjected to sensitivity analysis to predict the response of the system to different values of the factors affecting the strength. The plots obtained after sensitivity analysis indicated that increasing the amount of C3S, SO3 and surface area led to increased strength within the limits of the model. C2S decreased the strength whereas C3A decreased or increased the strength depending on the SO3 level. Because of the limited data range used for training, the prediction results were good only within the same range. The utility of the model is in the potential ability to control processing parameters to yield the desired strength levels and in providing information regarding the most favourable experimental conditions to obtain maximum compressive strength.Article Citation - WoS: 11Citation - Scopus: 13Experimental and Artificial Neural Network Modeling Study on Soot Formation in Premixed Hydrocarbon Flames(Elsevier Ltd., 2003) İnal, Fikret; Tayfur, Gökmen; Melton, Tyler R.; Senkan, Selim M.The formation of soot in premixed flames of methane, ethane, propane, and butane was studied at three different equivalence ratios. Soot particle sizes, number densities, and volume fractions were determined using classical light scattering measurement techniques. The experimental data revealed that the soot properties were sensitive to the fuel type and combustion parameter equivalence ratio. Increase in equivalence ratio increased the amount of soot formed for each fuel. In addition, methane flames showed larger particle diameters at higher distances above the burner surface and propane, ethane, and butane flames came after the methane flames, respectively. Three-layer, feed-forward type artificial neural networks having seven input neurons, one output neuron, and five hidden neurons for soot particle diameter predictions and seven hidden neurons for volume fraction predictions were used to model the soot properties. The network could not be trained and tested with sufficient accuracy to predict the number density due to a large data range and greater uncertainty in determination of this parameter. The number of complete data set used in the model was 156. There was a good agreement between the experimental and predicted values, and neural networks performed better when predicting output parameters (i.e. soot particle diameters and volume fractions) within the limits of the training data.