Katipoğlu, Okan Mert; Bazrafshan, Ommolbanin; Mohammadi, Mojtaba; Alsumaiei, Abdullah A.

This study aims to develop a robust multivariate regression framework for accurately predicting monthly Sea Surface Temperature (SST) using a core set of ten primary atmospheric and oceanographic variables, including relative humidity, specific humidity, temperature, mean square slope of waves, mean wave direction, 2-meter air temperature, top net solar radiation, total precipitation, mean evaporation rate, and potential evaporation. The originality of this research lies in the simultaneous integration of meteorological and oceanographic parameters into multiple machine learning architectures, coupled with a detailed interpretability analysis using SHAP values, which provide a comprehensive understanding of the influence of variables on SST prediction. The dataset was split into 60 % for training, 20 % for testing, and 20 % for validation. Four machine learning models: Random Forest, CatBoost, one-dimensional Convolutional Neural Network, and Support Vector Machine were optimized using RandomizedSearchCV, with MinMaxScaler applied to ensure consistent feature scaling across models. Model performance was evaluated using Mean Squared Error, Root Mean Squared Error, Mean Absolute Error, Coefficient of Determination, and Explained Variance metrics. Meanwhile, SHAP analysis identified the most influential features as 2-meter air temperature, mean evaporation rate, mean square slope of waves, air temperature, and specific humidity. Results revealed that the Support Vector Machine achieved the highest accuracy overall, followed by CatBoost, confirming that combining meteorological and oceanographic predictors significantly enhances SST forecasting performance. The physical consistency analyses, supported by the bulk air–sea heat flux formulations for sensible and latent heat fluxes, confirm that the machine-learning model’s most influential predictors: near-surface air temperature, specific humidity, and evaporation-related variables are fully consistent with the fundamental thermodynamic controls governing SST. This research not only advances SST prediction accuracy but also provides a transparent interpretability framework, making it a valuable reference for climate modeling, marine resource management, and environmental monitoring applications.

Akiner, Muhammed Ernur; Tınğ, Naim Süleyman; Alsumaiei, Abdullah A.; Katipoğlu, Okan Mert

This study investigates the prediction of daily total global solar radiation using advanced machine learning and hybrid models. The input variables include various meteorological and environmental parameters such as total precipitation, average snow height, total solar intensity, existing snow height, soil temperature, actual pressure, average temperature, cloudiness, minimum cloud height, and relative humidity. Based on SHAP value analysis, 5 cm Soil Temperature (+67.08) and Total Solar Intensity (+34.76) emerged as the most influential features for the models, while parameters like Current Snow Depth (+0.33) and Average Snow Depth (+0.36) contributed minimally. The analysis spans a period from 2006 to 2022 using daily data. The dataset is divided into 70 % for training and 20 % for validation, with 10 % of the training data allocated for testing. Several models were employed, including AdaBoost, CatBoost, Deep Belief Network (DBN), Light Gradient Boosting Machine (LGBM), Adaptive Neuro-Fuzzy Inference System (ANFIS), Echo State Neural Networks (ESN), and Autoencoder-based regression. According to the analysis results, the Catboost model produces the most accurate predictions with RMSE: 97.25 KW.hr/m2, MAE: 73.64 KW.hr/m2, AIC: 17075, NSE: 0.65 and, KGE: 0.72 values, while the DBN model shows the second best model with RMSE: 98 KW.hr/m2, MAE: 74.48 KW.hr/m2, AIC: 17103, NSE: 0.64 and, KGE: 0.74. The analysis results have the potential to provide information to decision makers and planners, especially in the evaluation of the performance of solar panels, the optimization of agricultural processes, and the development of climate change adaptation strategies.

Akgül, İsmail; Kaya, Volkan; Ünver, Edhem; Karavaş, Erdal; Baran, Ahmet; Tuncer, Servet

Coronavirus disease (Covid-19) has recently emerged as a serious public health threat, spreading rapidly worldwide and threatening millions of lives. With an increasing number of cases and mutations, medical resources are being drained daily owing to the rapid transmission of the disease, and the health systems of many countries are negatively affected. Therefore, it is important to use the available resources appropriately and in a timely manner to detect and treat the disease. In this study, VGG16 and ResNet50 deep learning models were used to quickly evaluate x-ray images and perform a prediagnosis of Covid-19, and an alternative model (IsVoNet) was proposed. Following model training, success accuracies of 99.92%, 99.65%, and 99.76% were achieved in the VGG16 model, ResNet50 model, and proposed model, respectively. According to the results, the models classified the Covid-19 and normal lung x-ray images with high accuracy, and the proposed model showed a high success rate at a lower time complexity than the other models.

Can, Erol; Toksoy, Mustafa Serdar

For electrical machines and power circuits to operate at high performance, it is important to control them at the optimum point. So, in this study, a flexible closed-loop (FCL) PID and PID FUZZY study is presented. Different from other studies, different controls can be activated for different positions and durations within the total operating time of the motor. At the same time, by applying four inputs for fuzzification in the designed fuzzy logic control section, 3 different outputs are provided to the PID control in three defuzzification sections. So, Dynamic Fuzzy logic interactivity is ensured with PID on load. This, unlike other PID FUZZY controllers, provides a more effective application of error and error change to the system. In the second chapter, DC motors and controls are explained. In the third chapter, designed controls and dc motor applications are made. The results obtained in practice reveal the effectiveness of the proposed system after compared with conventional P, PI and PID.

Can, Erol; Sayan, Hasan Hüseyin

The usage of multi-level inverters in converting direct current to alternating current is one of the most demanded and used methods. While creating multi-level voltage with inverters, the operating times of the switches that constitute the first step of the alternating voltage determine the gap of the step at the zero-crossing point. This situation can be preferred to create an extra level. Therefore, this article examines the effect of variation in operating times and frequencies of first-step PWMs on this zero-crossing point. First, the inverter structure providing load voltage is given in the study. Then, the transition of PWM run times is tried for different first-level voltages on the loads. The results obtained show the operation times of PWM at the transition point are quite effective on the load currents and voltages.

Can, Erol

It is important to find the optimum point in terms of energy quality in the studies for electrical energy conversions with the converter and inverter circuits. In such studies, the authors tried to find the optimum operating point by using control systems such as PID and Fuzzy Logic in closed-loop controls. When a large number of variables are involved in finding the optimum point, closed-loop control methods such as these may be insufficient to find the optimum point. Therefore, in this study, the Adaptive Taguchi Method (ATM) with the maximum-minimum value is used for the estimation of variables providing the optimum point for energy quality by using it in a multilevel inverter with a double DC-DC converter. While in traditional Taguchi Methods, dependent variables predict the results by revealing their effects on independent variables, in ATM, besides the effects of the independent variables, the effects of different dependent variables on each other are also estimated. First, the system to which the ATM will be applied is introduced. Then, the principles of applying the ATM are explained. In the known Taguchi method, variable values to be found with 34 = 81 trials, variables are estimated with 27 trials, while the values of two different dependent variables in the system can be found with 54 trials instead of 2×34= 2×81 trials. The resulting values to be estimated with the proposed ATM method are estimated with 27 trials instead of 2×81 or 54 trials. Finally, the observation results to be used for estimation are analyzed and evaluated. By trying the prediction results, it is seen that the proposed system is quite effective because the estimated result value that is %3.71 in the experiments gives a lesser distortion value than the values used for prediction.

Dalkiliç, Hüseyin Y.; Gharehbaghi, Amin

This paper documents a novel numerical model for calculating the behavior of unsteady, one-dimensional groundwater flow by using the finite volume method. The developed model determined water table fluctuations for different scenarios as follows: Drainage and recession from an unconfined aquifer, and water table fluctuations above an inclined leaky layer due to ditch recharge with a constant and variable upper boundary condition. The Boussinesq equation, which is the governing equation in this domain, is linearized and solved numerically in both of the explicit and fully implicit conditions. Meanwhile, the upwind scheme is employed to discretize the governing equation. The computed outcomes of both the explicit and implicit approaches agreed well with the results of analytical solution and laboratory experiments. The main reason is that in the first half of simulation process explicit scheme obtains slightly better results and in the second half of the simulation process fully implicit scheme predicts more reliable outcomes that are hidden in the neighbor node points. As a final point, the numerical outcomes confirm that the developed model is capable of calculating satisfactory outcomes in engineering and science applications.