Yazar "Kacar, Ilyas" seçeneğine göre listele
Listeleniyor 1 - 10 / 10
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Calibrating the combined hardening rule parameters for burr-free forging simulation of the torque rod joint(Sage Publications Ltd, 2024) Kacar, Ilyas; Yildirim, SefaJoints used in the automotive industry are widely manufactured by forging. A cold upsetting process can provide burr-free forging which reduces economic loss by preventing material waste. In this study, finite element simulations for the upsetting of a torque rod joint made of 41Cr4 steel are performed. The novelty of the present study lies in the fact that the upsetting performance is investigated through simulation having hardening model in order to replace the existing forging process. The performance of hardening models is studied for an accurate simulation and optimum parameters are determined. A combination of the bilinear isotropic hardening rule and Chaboche's nonlinear kinematic hardening rule is employed with the associated flow rule and Hill48 yield criterion to set up a plasticity model of the upsetting process for the first time. The parameters of the bilinear isotropic hardening rule are determined from monotonic tensile tests. The Chaboche's parameters are determined by using hysteresis loops obtained from strain-controlled low-cycle fatigue tests. The parameters of both rules are combined. Furthermore, they are calibrated using inverse analysis based on the optimization method. Genetic algorithm is used for optimization. The experimental diameter and height measurements of the joint are compared with those obtained from the optimized model. The results show that the application of the combined hardening rule provides better prediction performance of the upset dimensions with minimum dimensional tolerance. The calibrated parameters are presented for the upsetting process. The calibrated parameters of the combined hardening model for the upsetting are YS = 446.64 MPa, TM = 3363.05 MPa, C-1 = 452.31 MPa, gamma(1) = 55.165, C-2 = 212.13 MPa, gamma(2) = 12.24, C-3 = 194.191 MPa, gamma(3) = 10.00 where YS, TM, C-1,C- gamma 1,C- C-2,C- gamma 2,C- C-3,C- gamma 3 are hardening models' parameters. Absolute percent true error (APE) is 0.19%. The parameters are YS = 1.93 MPa, TM = 6.98 MPa, C-1 = 580.79 MPa, gamma(1) = 1.08, C-2 = 597.23 MPa, gamma(2) = 0.98, C-3 = 565.05 MPa, gamma(3) = 2.87 in the case of cyclic load. APE is 1.66%. Also upsetting force requirement and material flow path are presented. The forging process can be replaced by the burr-free upsetting process with necessary changes in the die and press bench design. This replacement will save the 128-gr material per each one of the torque rod joint part.Öğe Explaining Data Preprocessing Methods for Modeling and Forecasting with the Example of Product Drying(Univ Namik Kemal, 2024) Korkmaz, Cem; Kacar, IlyasAlthough regression is a traditional data processing method, machine and deep learning methods have been widely used in the literature in recent years for both modelling and prediction. However, in order to use these methods efficiently, it is important to perform a preliminary evaluation to understand the data type. Therefore, preevaluation procedures are described in this study. Experimental uncertainty analysis was performed to determine the measurement uncertainties in the measurement devices and sensors used in the drying experimental setup. Significant and insignificant relationships between variables in the data set were determined by Pearson correlation matrix. Autocorrelation and partial autocorrelation functions were used to determine the time series lag in the drying data and an AR(5) series with 5 lags was determined. The data were found to have variable variance due to peaks and troughs in the raw data resulting from the natural behaviour of the drying process. Modelling success was achieved with the normalisation pre -evaluation process performed without distorting the raw data. Thus, it has been shown that better models can be obtained compared to traditional models. In order to avoid unnecessary time and computational costs in the trial and error method used to determine the number of hidden layers and neurons in the machine learning method, various formulas proposed in the literature were compared. It is shown that the correlation coefficient alone is not sufficient to determine the goodness of the model. In modelling the data in this study, the NARX model was found to converge to the desired value faster and with less error than ANFIS and LSTM models. In the simulation of a rotary drum dryer, the optimum number of mesh elements was determined as 1137 by mesh independence analysis. In this way, unnecessary over -calculations were also prevented. Of course, all these methods are already available in statistical science. However, in this study, the methods to be used for modelling and prediction purposes are carefully selected and explained with examples, especially for young researchers who are outside this field to gain speed and easy comprehension.Öğe Free vibration/buckling analyses of noncylindrical initially compressed helical composite springs(TAYLOR & FRANCIS INC, 2016) Kacar, Ilyas; Yildirim, VebilThis article presents the use of the stiffness matrix method based on the first-order shear deformation theory to predict the fundamental natural frequencies and buckling loads of noncylindrical unidirectional composite helical springs subjected to initial static axial force and moment. This theoretical study about such springs with circular/rectangular cross-sections and large pitch angles is performed for the first time in the literature. The validity of the present results is verified by the benchmark studies related with initially compressed isotropic cylindrical springs.Öğe Investigation of forging performance for AA6082(Springer London Ltd, 2021) Tunc, Ozkan; Kacar, Ilyas; Ozturk, Fahrettin6XXX series aluminum alloys are generally excellent alternatives to steels for many forged parts in aerospace and automotive industries. In this study, the forging performance of the 6082 aluminum alloy is investigated in order to replace the existing material for forged steel parts. The effect of artificial aging of the alloy on the microstructure and mechanical properties is studied. Optimum aging conditions are determined. Results reveal that AA6082 could be a good replacement for applications where shock and vibrational loads exist. The rod end automotive part currently manufactured from AISI1045 can be replaced by AA6082 without any design changes. The major drawback is that the cold forging of the aged alloy is poor due to its brittle nature and crack initiations. Therefore, warm or hot forging is recommended to overcome the poor forgeability.Öğe Parameter Calibration of Chaboche Kinematic Hardening Model by Inverse Analysis Using Different Optimization Methods in the Case of Pipe Bending(Society of Automotive Engineers Turkey, 2024) Akkoyun, Ozan; Kacar, IlyasDrag link is one of the important parts in steering system used in automotive. The ball joint, ball joint housing, and pipe compose the drag link. In this study, finite element analysis is used to simulate the deformation process. St 52 steel material is used. A yield criterion, an associated flow rule, and Chaboche's kinematic hardening rule were used in the finite element simulations of processes involving high plastic deformation. A series of low-cycle tensile/compression tests is performed to determine the parameters of Chaboche's kinematic hardening rule. The success of the simulation results depends on the more accuracy of the finding parameters. Some optimization methods are used in the calibration progression of these parameters and the results are compared. For the purpose of optimization, the angle of the pipe after bending is set as 16.6° as soon as possible. As design variables, the Chaboche kinematic hardening rule parameters were adjusted. Consequently, calibrated parameters were obtained for St52 pipe bending. By analysing and verifying the candidate points, optimization methods are compared. The optimum parameters are determined as JSH350 MPa, C=2984.3 MPa, and ^=100 while their initial values are Y5K373.806 MPa, C=4016 MPa, and y=94. It is concluded that the optimization process gives more consistency in the bending process. © 2024 Society of Automotive Engineers Turkey. All rights reserved.Öğe Parameter determination and calibration of the combined plasticity model through inverse analysis for cold forging simulation of 7075-T6 aluminum alloy(Gazi Univ, Fac Engineering Architecture, 2022) Kacar, Ilyas; Kilic, SuleymanIn this study, plasticity models have been developed to be used in cold forging simulations of AA7075-T6 alloy which is widely used in aviation industry. In addition, the coefficients of the obtained models have been calibrated using genetic algorithm optimization method. As the hardening rule in models; bilinear isotropic is combined with Chaboche's nonlinear kinematic hardening rule (three-termed). Plasticity models have been obtained by using the associated flow rule and Hill48 yield criterion in addition to the hardening rules. Experimental stress values have been compared with those obtained from the models. As a result, the most suitable hardening model for monotonic/cyclic loading deformation conditions is presented and the effects of the model parameters on the simulation results are shown.Öğe Parameters Calibration of the Combined Hardening Rule through Inverse Analysis for Nylock Nut Folding Simulation(Tech Science Press, 2021) Kacar, IlyasLocking nuts are widely used in industry and any defects from their manufacturing may cause loosening of the connection during their service life. In this study, simulations of the folding process of a nut's flangemade from AISI 1040 steel are performed. Besides the bilinear isotropic hardening rule, Chaboche's nonlinear kinematic hardening rule is employed with associated flow rule and Hill48 yield criterion to set a plasticity model. The bilinear isotropic hardening rule's parameters are determined by means of a monotonic tensile test. The Chaboche's parameters are determined by using a low cycle tension/compression test by applying curve fitting methods on the low cycle fatigue loop. Furthermore, the parameter calibrations are performed in the finite element simulations by using an optimization approach based on the inverse analysis. Dimensional accuracy for the nut is of primary concern due to the tolerance constraints of the nut manufacturers. Experimental diameter and height measurements of the folded locking nut are compared with those obtained fromthe optimized model. The results reveal that the folding dimensions can be predicted more accurately when the model parameters are determined by using the combined hardening rule. The calibrated parameters are presented for the folding and cycling deformation processes.Öğe Prediction of Strain Limits via the Marciniak-Kuczynski Model and a Novel Semi-Empirical Forming Limit Diagram Model for Dual-Phase DP600 Advanced High Strength Steel(Assoc Mechanical Engineers Technicians Slovenia, 2020) Kacar, Ilyas; Ozturk, Fahrettin; Toros, Serkan; Kilic, SuleymanThe prediction capability of a forming limiting diagram (FLD) depends on how the yield strength and anisotropy coefficients evolve during the plastic deformation of sheet metals The FLD predictions are carried out via the Marciniak-Kuczynski (M-K) criterion with anisotropic yield functions for DP600 steel of various thicknesses. Then, a novel semi-empirical FLD criterion is proposed, and prediction capabilities of the criterion are tested with different yield criteria. The results show that the yield functions are very sensitive to anisotropic evolution. Thus, while the FLD curves from the M-K model and the proposed model are not the same for each thickness, the proposed model has better prediction than the M-K model.Öğe Review of warm forming of aluminum-magnesium alloys(ELSEVIER SCIENCE SA, 2008) Toros, Serkan; Ozturk, Fahrettin; Kacar, IlyasAluminum-magnesium (Al-Mg) alloys (5000 series) are desirable for the automotive industry due to their excellent high-strength to weight ratio, corrosion resistance, and weldability. However, the formability and the surface quality of the final product of these alloys are not good if processing is performed at room temperature. Numerous studies have been conducted on these alloys to make their use possible as automotive body materials. Recent results show that the formability of these alloys is increased at temperature range from 200 to 300 degrees C and better surface quality of the final product has been achieved. The purpose of this paper is to review and discuss recent developments on warm forming of Al-Mg alloys. (C) 2008 Elsevier B.V. All rights reserved.Öğe Time-series prediction of organomineral fertilizer moisture using machine learning(Elsevier, 2024) Korkmaz, Cem; Kacar, IlyasThis study aims to model and forecast the drying process of three new types of commercial organomineral fertilizers: gold sulfur, 25.5.5, and 5x10 at elevated temperatures. They absorb and release moisture, depending on the conditions. Accurate prediction of drying behaviour is essential. Drying was carried out at temperatures of 70, 75, and 80 degrees C through natural convection. The data are unimodal time series of the moisture rate (MR). MR ). Supervised machine/deep learning techniques such as nonlinear autoregressive (NAR) network, adaptive neural fuzzy inference systems (ANFIS), long short-term memory (LSTM) network, gated recurrent unit (GRU), and hybrid of convolutional neural network (CNN) and recurrent neural network (RNN) are used in addition to wellknown regression-based formulas. The models can predict 30 minutes ahead. An error analysis was performed for performance comparison using the metrics, root mean square error (RMSE), RMSE ), and coefficient of determinant, R 2 . Two types of validation were performed by monitoring error convergence and using prediction curves. The most effective forecasting was obtained at an air temperature of 80 degrees C for all materials with machine learning. While RMSE =0.0030 having R 2 = 0.98843 by the LSTM network for gold sulfur (80 degrees C), ANFIS was the best for 25.5.5 and 5x10 with RMSE =0.0056, R 2 = 0.75585 and RMSE =0.0060, R 2 = 0.81938, respectively, in the MR prediction. GRU was remarkable for both its speed of 13.77 sec and RMSE =0.009. CNN-RNN has a more complex structure but lower performance. The results demonstrate that machine learning techniques are better than regressions. Among them, ANFIS provides the most reliable results. Regressions, including exponential terms, were good at providing a general curve shape but not peaks and drops. In addition, regressions are not good at forecasting.
