Sürdürülebilir Kalkınma Amaçları
Yazarlar
- Gul, Muhammet(3)
- Akpinar, Muhammed Halil(1)
- Celik, Erkan(1)
- Dogan, Gurkan(1)
- Drobne, S.(1)
- Durmuş, Aydın(1)
- Guneri, Ali Fuat(1)
- Gülçimen, Fevzi(1)
- Imak, Andaç(1)
- Karakaya, Hakan(1)
- Kljajic Borstar, M.(1)
- Kurum, Hasan(1)
- Köküm, Mehmet(1)
- Oksuztepe, Eyyup(1)
- Onat, Onur(1)
- Pekel, Engin(1)
- Polat, Mehmet(1)
- Uyanik, Hakan(1)
- Yogurtcu, Hakan(1)
- Zadnik Stirn, L.(1)
- Zerovnik, J.(1)
- Özçelik, Salih Taha Alperen(1)
- Şengür, Abdulkadir(1)
- Şentürk, Erman(1)
Yayınlar
- MakaleYalnız künye2026Scopus SJRQ1Modeling stack voltage and remaining useful life of proton exchange membrane fuel cells using deep neural networks
Dogan, Gurkan; Imak, Andaç
Currently, various energy conversion systems are being developed to reduce environmental impacts. Proton exchange membrane fuel cells (PEMFCs), which stand out with their high efficiency characteristics, have emerged as an important alternative. However, these advantages, along with high costs and uncertainties regarding remaining useful life (RUL) estimation, restrict their further commercialization, especially in the automotive industry. This study presents a forecasting framework using a long short-term memory (LSTM) architecture enhanced with the Bahdanau attention mechanism for predicting RUL and long-term degradation trends. The model thus exhibits a performance that can more effectively capture dependencies in time series and adaptively weight significant time steps. Furthermore, the tree-structured parzen estimator (TPE) was chosen for the appropriate selection of hyperparameters in the developed forecasting framework. The validity, stability, and performance of the proposed model under different operating conditions were examined on the FC1 and FC2 datasets. Experimental studies have shown that the proposed method is the most successful among the best methods, achieving RE metric values of 1.628 × 10−3, 0.003 × 10−1, 99.16 × 10−2, and 0.044 × 10−1 respectively in estimating the best RMSE, MAPE, R2, and RUL.
- MakaleAçık Erişim2026Scopus SJRQ3An Artificial Neural Network Modeling of Solar Drying of Mint: Energy, Exergy, and Drying Kinetics
Gülçimen, Fevzi; Karakaya, Hakan; Durmuş, Aydın
Abstract: Solar drying is an environmentally friendly method for preserving agricultural products; however, limited studies have jointly examined energy–exergy performance and data-driven prediction methods under varying airflow conditions, which restricts performance optimization. This study aims to investigate the effects of inlet airflow rate on energy utilization ratio, energy generation, exergy loss, exergy efficiency, and drying kinetics of mint leaves, and to evaluate the predictive capability of artificial neural network modeling. Thin-layer drying experiments were conducted in a forced convective solar dryer equipped with a newly designed solar air collector at different air mass flow rates. Energy and exergy analyses were conducted using experimental thermodynamic data, and energy utilization and exergy loss were modeled with a feed-forward back-propagation artificial neural network. The energy utilization ratio ranged from 7.45 to 87.1% and increased with decreasing airflow rate. The average exergy losses for air mass flow rates of 0.012, 0.026, and 0.033 kg s–1 were 16.2, 8.2, and 6.88 W, respectively. Both experimental and artificial neural network results showed that increasing airflow rate reduced energy utilization ratio and exergy loss. The artificial neural network predictions exhibited very close agreement with experimental data. The results show that artificial neural networks can accurately predict energy utilization ratio and exergy loss in a forced convective solar dryer. Airflow rate is a key parameter for thermodynamic performance, and the integrated experimental–ANN approach offers a robust tool for optimizing and designing solar drying systems.
AtıfOpenAlex1
İncele - BildiriAçık Erişim2025Toplam Elektron Içeriǧi ve Uzay Iklim Verileri Kullanilarak Transfer Öǧrenme ile Deprem Öncüllerinin Analizi
Uyanik, Hakan; Şentürk, Erman; Köküm, Mehmet; Özçelik, Salih Taha Alperen; Akpinar, Muhammed Halil; Şengür, Abdulkadir
The stress changes occurring in the Earth's crust before earthquakes may cause some anomalies in the ionosphere and cause significant changes in the Total Electron Content (TEC) and Space Climate (SCC) parameters. The analysis of these changes offers a significant potential in the detection of earthquake precursors. In this study, the analysis of earthquake precursors was performed using TEC and UC data and the effectiveness of transfer learning and machine learning techniques was investigated. Features were extracted from five-day TEC and UC data using a pre-trained Multi-Input One-Dimensional Convolutional Mixer Network (ÇG-TB-EMA) and these features were classified with different machine learning algorithms. Within the scope of the study, earthquakes with a magnitude of Mw≥6.0 that occurred during the low and high solar activity periods of the 24th solar cycle were analyzed. The earthquakes were selected from the Far East region between 10°S-45°N latitudes and 120°E-160°E longitudes, where intense seismic activity is observed. Experimental studies have shown that during the low solar activity period, support vector machines (SVM) and artificial neural networks (ANN) models exhibited the highest performance with 96.04% accuracy and 96.08% F1 score, respectively. During the high solar activity period, the classification success decreased slightly, and SVM and ANN models achieved the best results with 90.28% accuracy and 89.86% F1 score. These findings reveal that the effect of solar activity on ionospheric variables should be taken into account in earthquake precursor detection. For future studies, long-term analyses are recommended to better understand the effects of solar activity.
- MakaleYalnız künye2018Scopus SJRQ3TIME-INDEPENDENT RHEOLOGICAL BEHAVIOUR OF PEANUT BUTTER: MODELLING AND ARTIFICIAL NEURAL NETWORK APPLICATION
Yogurtcu, Hakan
In this study, the rheological behaviour of semisolid peanut butter samples with a fat ratio of 42.7 % were examined in a temperature interval of 10-50 degrees C via rotational rheometer at different shear rates. Peanut butter samples display non-Newtonian yield-pseudoplastic behaviour. Seven model equalities were used for representing the time-independent rheological behaviour of peanut butter and the representation ability of the models were determined by way of determination coefficient (r(2)), mean absolute percentage error (MA%E), chi-square (chi(2)) and root mean square error (RMSE) statistical parameters. It was observed that the O'foli model was the most successful among the models applied to experimental data. Temperature dependency of viscosity at shear rates of 0.1, 0.4 and 0.75 s(-1) was determined via Arrhenius equality and activation energies varied between the interval of 46818.63-51952.52 J/mol. Three-layer model suggested for estimating viscosity using artificial neural network method was observed to be quite successful.
- MakaleYalnız künye2018Scopus SJRQ1Application of Artificial Neural Networks to the prediction of out-of-plane response of infill walls subjected to shake table
Onat, Onur; Gul, Muhammet
The main purpose of this paper is to predict missing absolute out-of-plane displacements and failure limits of infill walls by artificial neural network (ANN) models. For this purpose, two shake table experiments are performed. These experiments are conducted on a 1:1 scale one-bay one-story reinforced concrete frame (RCF) with an infill wall. One of the experimental models is composed of unreinforced brick model (URB) enclosures with an RCF and other is composed of an infill wall with bed joint reinforcement (BJR) enclosures with an RCF. An artificial earthquake load is applied with four acceleration levels to the URB model and with five acceleration levels to the BJR model. After a certain acceleration level, the accelerometers are detached from the wall to prevent damage to them. The removal of these instruments results in missing data. The missing absolute maximum out-of-plane displacements are predicted with ANN models. Failure of the infill wall in the out-of-plane direction is also predicted at the 0.79g acceleration level. An accuracy of 99% is obtained for the available data. In addition, a benchmark analysis with multiple regressionis performed. This study validates that the ANN-based procedure estimates missing experimental data more accurately than multiple regression models.
AtıfScopus33WoS29OpenAlex14
İncele - BildiriYalnız künye2017Forecasting daily patient visits in an emergency department by ga-ann hybrid approach
Pekel, Engin; Gul, Muhammet; Celik, Erkan; Drobne, S.; Zerovnik, J.; Zadnik Stirn, L.; Kljajic Borstar, M.
An Emergency Department (ED) plays a crucial role in the health system by providing acute care for patients who attend hospital without prior appointment. An accurate forecasting of patient visits in EDs contributes to health care decision makers to better allocate ED human resources and medical equipment. In this paper, a hybrid genetic algorithm–artificial neural network (GA–ANN) approach is developed. The forecasting performance of the hybrid approach is obtained using the real-world data set collected from a public hospital in Istanbul, Turkey. The hybrid GA–ANN model is shown to perform well in terms of forecasting accuracy. In order to contribute to the current knowledge, this paper is a novel attempt of applying GA-ANN to model ED patient arrivals, and the results can be used to aid in strategic decision-making on ED resource planning in response to predictable arrival variations.
AtıfScopus4İncele - MakaleAçık Erişim2016Scopus SJRQ3Switched reluctance motor control without position sensor by using data obtained from finite element method in artificial neural network
Polat, Mehmet; Oksuztepe, Eyyup; Kurum, Hasan
Switched reluctance motors (SRMs) require sensitive rotor information in order to work sensitively and properly and especially for the commutation of phase currents. Position information is generally acquired from a sensor which is generally placed into the stator or connected to the rotor. However, sensors have disadvantages in terms of cost, safety, complexity and volume. In order to remove these disadvantages, recent studies concentrate on SRMs working without a position sensor. In this study control methods working without a position sensor are examined, the applied method is presented comprehensively. The education and test data of the artificial neural network is not obtained from simulation or experimental results; it is obtained from magnetic analyses carried out in ANSYS 10 and the results are used in position speed and current sensorless control of SRM. The validity of proposed method is demonstrated by using DS 1103.
AtıfWoS8OpenAlex11İncele - MakaleYalnız künye2016Scopus SJRQ1An artificial neural network-based earthquake casualty estimation model for Istanbul city
Gul, Muhammet; Guneri, Ali Fuat
Earthquakes over the world cause serious damages for people and the economy. The establishment of an appropriate model to estimate potential losses or injuries in earthquake disasters is fundamental to decrease their impacts and losses and effectively respond and mitigate. Turkey as a country that experienced many earthquakes in the last century and had serious human and financial losses needs a comprehensive knowledge of consequences of devastating earthquakes to be able to plan for the future. Artificial neural networks (ANNs) have abilities to solve and analyse complex relations as an appropriate method to estimate number of injuries. In this study, an ANN model is built up for the earthquake casualty prediction, which takes earthquake occurrence time, earthquake magnitude, and population density as the predictors and employs 21 Mw > 5 earthquake disasters occurred in Turkey from 1975 as samples for the training of the network. The model was then tested on a study region consisting of four districts in Istanbul which is estimated to have the highest injury rate according to the earlier reports and generated estimations of the expected number of injured people. Results show that 99.9 % of the variability in the number of injured people is predictable with using this ANN model. Comparison of actual values and estimated output values in the ANN model was also found apparently very close to each other. According to the test case study results, when the value of earthquake magnitude is 6.5 Mw, number of injured people has increased in a sharp trend compared to the previous magnitude value (6 Mw) for all districts. Estimated number of injured people in daytime is obtained higher than at night when the earthquake magnitude is 5 and 5.5 Mw. The highest value in estimated number of injured people has emerged in Fatih district as 7241 for a 7.5 Mw daytime earthquake. On conclusion, it was deduced that the model can reveal accurate estimation of casualties and it can provide information in order to develop mitigation policies, especially in earthquake emergency service management.
AtıfWoS32OpenAlex46İncele