Yazar "Bacak, Aykut" seçeneğine göre listele

Listeleniyor 1 - 6 / 6
  • Küçük Resim
    Öğe
    Artificial neural network-based cooling capacity estimation of various radiator configurations for power transformers operated in ONAN mode
    (Elsevier, 2024) Koca, Aliihsan; Şenturk, Oğuzkan; Çolak, Andaç Batur; Bacak, Aykut; Dalkılıç, Ahmet Selim
    Power transformers submerged in oil are universally acknowledged as very useful elements in electrical power networks. A fraction of the electrical energy involved in the conversion from high to low voltages is lost as thermal energy, which is produced inside the transformer's windings and core. The effective dissipation of heat is of paramount significance and can be accomplished through the strategic installation of radiators on the tank. This study aims to examine the total cooling capacity of a radiator employed for cooling a power transformer operating in oil, natural air, and natural mode. The investigation is conducted by varying the design parameters, specifically the number of fins per radiator and the radiator length. The study also utilizes computational fluid dynamics results to achieve a substantial convergence with experimental findings. In the context of the verification study conducted for a specific study point, it was found that the simulation outcomes at a mass flow rate of 0.15 kg/s corresponded to about 7.4 % of the experimental results. Additionally, the cooling capacity value obtained was 7.2 %. In the context of machine learning, the performance of the numerical approach was assessed by employing the Bayesian regularization method. The evaluation revealed that the margin of deviation, mean squared error, and coefficient of determination (R2) metrics yielded values of ?0.001 %, 1.32E-02, and 0.99930, respectively.
  • Küçük Resim Yok
    Öğe
    An artificial neural network-based numerical estimation of the boiling pressure drop of different refrigerants flowing in smooth and micro-fin tubes
    (De Gruyter, 2024) Çolak, Andaç Batur; Bacak, Aykut; Kayacı, Nurullah; Dalkılıç, Ahmet Selim
    In thermal engineering implementations, heat exchangers need to have improved thermal capabilities and be smaller to save energy. Surface adjustments on tube heat exchanger walls may improve heat transfer using new manufacturing technologies. Since quantifying enhanced tube features is quite difficult due to the intricacy of fluid flow and heat transfer processes, numerical methods are preferred to create efficient heat exchangers. Recently, machine learning algorithms have been able to analyze flow and heat transfer in improved tubes. Machine learning methods may increase heat exchanger efficiency estimates using data. In this study, the boiling pressure drop of different refrigerants in smooth and micro-fin tubes is predicted using an artificial neural network-based machine learning approach. Two different numerical models are built based on the operating conditions, geometric specifications, and dimensionless numbers employed in the two-phase flows. A dataset including 812 data points representing the flow of R12, R125, R134a, R22, R32, R32/R134a, R407c, and R410a through smooth and micro-fin pipes is used to evaluate feed-forward and backward propagation multi-layer perceptron networks. The findings demonstrate that the neural networks have an average error margin of 10?percent when predicting the pressure drop of the refrigerant flow in both smooth and micro-fin tubes. The calculated R-values for the artificial neural network’s supplementary performance factors are found above 0.99 for all models. According to the results, margins of deviations of 0.3?percent and 0.05?percent are obtained for the tested tubes in Model 1, while deviations of 0.79?percent and 0.32?percent are found for them in Model 2.
  • Küçük Resim
    Öğe
    Improving pressure drop predictions for R134a evaporation in corrugated vertical tubes using a machine learning technique trained with the Levenberg-Marquardt method
    (Springer, 2024) Çolak, Andaç Batur; Bacak, Aykut; Karakoyun, Yakup; Koca, Aliihsan; Dalkılıç, Ahmet Selim
    The present investigation utilized a machine learning structure to ascertain the pressure drop in vertically positioned, corrugated copper tubes during the evaporation process of R134a. The evaporator was a counter-flow heat exchanger, in which R134a flowed in the inner corrugated tube and hot water flowed in the smooth annulus. Different evaporation mass fluxes (195-406 kg m-2 s-1) and heat fluxes (10.16-66.61 kW m-2) were used with artificial neural networks at different corrugation depths. A multilayer perceptron artificial neural network model with 13 neurons in the hidden layer was proposed. Tan-Sig and Purelin transfer functions were used in the network model developed with the Levenberg-Marquardt training algorithm. The dataset, which consisted of 252 data points, related to the evaporation process, was divided into training (70%), validation (15%), and testing (15%) groups in an arbitrary manner. The artificial neural network model has been demonstrated to effectively forecast the pressure drop that occurs during evaporation. The mean squared error was computed for the Delta P values observed during the evaporation processes, yielding a value of 1.96E-03. The artificial neural network exhibited a high correlation coefficient value of 0.94479. The estimation fluctuations exhibited a range of +/- 10%, whereas the experimental and anticipated Delta P data demonstrated a divergence of +/- 10.3%.
  • Küçük Resim Yok
    Öğe
    Investigating hermetic reciprocating compressor performance by using various machine learning methods
    (SAGE Publications Ltd, 2023) Bacak, Aykut; Çolak, Andaç Batur; Dalkılıç, Ahmet Selim
    Due to their durability and efficiency, hermetic reciprocating compressors (HRCs) are used in refrigeration and air conditioning. Compressor performance and reliability concerns reduce system efficiency and raise maintenance costs. Machine learning (ML) is being used to improve hermetic reciprocating compressor performance, reliability, and energy economy. ML is used in hermetic reciprocating compressors for issue identification, performance improvement, predictive maintenance, and energy management. This research compared HRC performance factors such as mass flow rate, cooling capacity, compression power, coefficient of performance, exhaust line losses, and volumetric efficiency. Simple regression, probabilistic neural network, gradient boosted, polynomial regression, and random forest (RF) were used to examine and evaluate these parameters as outputs. Over three cycles, the Fluid-Structure Interaction (FSI) approach assessed compressor performance parameters. For compressor speeds of 1300, 2100, and 3000 rpm, mass flow rate, compression power, cooling efficiency coefficient, and exhaust line energy losses varied by 10%, 4%, 5%, and 6%. To gather ML algorithm inputs, the research used experimental, fluid-structure interaction, and ML methodologies. Experimental and FSI approaches produced 108 data points. These data points were randomly assigned, with 70% for learning and 30% for prediction. The mean convergence criterion for mass flow rate, cooling capacity, compression power, cooling efficiency coefficient, exhaust line energy losses, and volumetric efficiency parameters was 0.9966, 0.9969, 0.9572, 0.0561, 0.9925, and 0.4640 for all ML methods. Simple regression, probabilistic neural networks, gradient boosted, polynomial regression, and RF convergence criteria were 0.8978, 0.9999, 0.6016, 0.4439, and 0.7761.
  • Küçük Resim Yok
    Öğe
    Prediction of performance parameters of a hermetic reciprocating compressor applying an artificial neural network
    (SAGE Publications Ltd, 2024) Bacak, Aykut; Çolak, Andaç Batur; Dalkılıç, Ahmet Selim
    The efficiency of the compressor's thermodynamic, mechanical, and electric motors can be increased by minimizing their losses. Based on the design characteristics of the compressor, the current study proposes an artificial neural network (ANN) model to forecast the mass flow rate, cooling capacity, compression power, and discharge line loss of the hermetic reciprocating compressor as outputs. This study uses experimental and numerical results obtained by the fluid–structure interaction (FSI) method, and the whole findings are used as inputs for the machine learning method. Using 36 numerical and 3 experimental data sets, a multilayer network is created with input parameters defining mass flow rate, cooling capacity, compression power, and discharge line energy loss. The input parameters for the network model are the compressor's design parameters, which include the compressor speed, discharge valve thickness, and discharge valve length. When comparing the experimental and FSI methods for compressor speeds of 1300, 2100, and 3000?rpm, the FSI study's mass flow rate and cooling capacity parameters converged at 11.9%, 9.1%, and 9.3%, respectively, to their actual values. The convergence rate to experimental compression power data was ?0.3%, 2.6%, and 11%. The numerical-experimental deviation for discharge line energy losses is 14.9%, ?3.6%, and 7.8%, respectively, for 1300, 2100, and 3000?rpm. With the Levenberg–Marquardt (LM) ANN model, which is used in this study, the average squared error value is calculated as 1.75E-02 and the R2 value is calculated as 0.99976. ANN model can predict outputs with average deviation rates of less than 0.88%. It is seen that results obtained with the ANN method provide high convergence on the experimental and FSI results, while the results obtained with the regression model deviate from the compressor exhaust line results by more than ±20%, and the ANN method is more successful than regression.
  • Küçük Resim Yok
    Öğe
    Prediction of performance parameters of a hermetic reciprocating compressor under different discharge lift limiter heights by machine learning
    (SAGE Publications Ltd., 2024) Bacak, Aykut; Çolak, Andaç Batur; Dalkılıç, Ahmet Selim
    The research examines the complex correlation between discharge valve properties in severe temperature circumstances, ranging from 54.4 degrees C to -23.3 degrees C, in accordance with ASHRAE operational guidelines. The design parameters include examining valve thicknesses of 0.127, 0.152, 0.178, and 0.2 mm, together with lengths of 14.722, 16.222, and 17.722 mm, at compressor speeds of 1300, 2100, and 3000 rpm. An artificial neural network (ANN) is used to replicate the output properties of a hermetic reciprocating compressor, which include the ratio of cooling capacity to compression power and volumetric efficiency. One hundred and eleven numerically recorded datasets are used to train the developed ANN model. The model is trained using 77 datasets, validated using 17 datasets, and tested using 17 datasets. The LM-type ANN approach is used to train the multilayer perception neural network, which consists of a hidden layer with 15 neurons. Given the proximity of the margin of deviations (MoDs) to the 0% deviation line, the variances between the ANN and fluid-structure interaction outcomes for the cooling capacity to compression power ratio and volumetric efficiency are insignificant. The average figures for the MoD output have been calculated as -0.18% and 0.06, respectively. Not only do the data points lie on the line, indicating a 0% error, but they also fall inside the interval, indicating a 10% error. In addition, the mean squared error and correlation coefficient values for the ANN model that was created are 2.04E-03 and 0.99853, respectively.