Yazar "Yildirim, Ramazan" seçeneğine göre listele
Listeleniyor 1 - 16 / 16
Sayfa Başına Sonuç
Sıralama seçenekleri
Öğe Analysis of lipid production from Yarrowia lipolytica for renewable fuel production by machine learning(Elsevier Sci Ltd, 2022) Cosgun, Ahmet; Gunay, M. Erdem; Yildirim, RamazanIn this work, biomass and lipid productivities of Yarrowia lipolytica were analyzed using machine learning techniques. A dataset containing 356 instances was constructed from the experimental results reported in 22 publications. The dataset was analyzed using decision trees to identify the features (descriptors) that lead to high biomass production, lipid content and lipid production. C/N ratio and fermentation time were found to be the most influential features for biomass production while the use of glucose and medium pH seemed to be more important for high lipid content. For the lipid production case, five generalizable paths leading to high values of this output were identified. One of those paths required pH to be < 6.3, high glucose and (NH4)(2)SO4 concentrations, lower concentration for yeast extract and the yeast strain not be H-222. Another one needed a pH greater than 6.3, a C/N ratio smaller than 75, a time greater than 14 h, and a strain other than W29. The same dataset was also explored deeper using association rule mining to determine the effects of individual features on output variables. It was then concluded that machine learning methods are very useful in determining the optimal conditions of biomass growth and lipid yield for Yarrowia lipolytica to produce renewable biofuels.Öğe Analysis of past experimental data in literature to determine conditions for high performance in biodiesel production(Wiley, 2016) Tapan, N. Alper; Yildirim, Ramazan; Gunay, M. ErdemIn this study, published experimental works on catalytic transesterification were analyzed to determine the most important variables affecting fatty acid conversion and the most suitable ranges of these variables for high performance. A database of 1324 data points was constructed from the experimental results in 31 representative papers published between 2008 and 2014, and this database was analyzed using artificial neural network (ANN) and decision tree (DT) techniques. It was found from ANN analysis that the most important variable for high fatty acid conversion was reaction time (with about 40% relative importance) followed by catalyst loading, alcohol:oil molar ratio, operating temperature, and support type with similar relative importance (about 10% each). DT analysis revealed 14 combinations of conditions leading to high performance, and some of these seemed to be generalizable for the use for the future studies; some heuristics were also derived from these generalizable conditions. (c) 2016 Society of Chemical Industry and John Wiley & Sons, LtdÖğe CO2 capture over amine-functionalized MCM-41 and SBA-15: Exploratory analysis and decision tree classification of past data(Elsevier Sci Ltd, 2019) Yildiz, Merve G.; Davran-Candan, Tugba; Gunay, M. Erdem; Yildirim, RamazanThis study aims to extract knowledge for CO2 capture by amine-functionalized mesoporous silica (MCM-41 and SBA-15) through exploratory analysis and decision tree classification of the data reported in over 100 papers published between the years 2002 and 2017. A database containing 1039 data points showing the effects of 15 input variables (grouped in four as support properties, preparation method, amine properties and operational variables) over two performance variables as CO2 adsorption capacity and amine efficiency (CO2 captured/amino groups involved) was constructed. Box and whisker plots were applied (as a part of exploratory data analysis) to determine how various input variables influence the performance variables. Moreover, decision tree classification was used to determine the relative significance of the input variables and the possible combinations of these variables leading to high performance (to deduce heuristics for high CO2 uptake). It was found from the exploratory data analysis that amine density was the most significant variable affecting the adsorption capacity whereas remaining pore volume and adsorption temperature were the most influential variables in case of amine efficiency. Furthermore, various combinations of input variables leading to high CO2 capture performance were revealed through the decision tree analysis, all of which may be used as guidelines for future studies in this area.Öğe Constructing global models from past publications to improve design and operating conditions for direct alcohol fuel cells(Inst Chemical Engineers, 2016) Tapan, N. Alper; Gunay, M. Erdem; Yildirim, RamazanThis work aims to analyze past publications on direct alcohol fuel cells (DAFC) in the literature using two data mining tools (artificial neural networks and decision trees) and to develop global models to predict the conditions leading to high performance of DAFC. The database constructed for this purpose contains 4682 data points over 271 polarization (IV) curves obtained from 36 publications in the literature. Decision tree classification models were used to develop heuristics to select the suitable fuel cell design and operational conditions to improve the maximum power density while artificial neural network models (ANN) were developed to test the predictability of IV curves at the conditions where experimental results were not available. The same ANN models were also used to determine the relative importance of design and operational variables to provide some insight to determine the variable to be manipulated. All these analyses were quite successful deducing some useful heuristics and models for the future studies from the continuously growing experience accumulated in the literature. (C) 2015 The Institution of Chemical Engineers. Published by Elsevier B.V. All rights reserved.Öğe Decision tree analysis of past publications on catalytic steam reforming to develop heuristics for high performance: A statistical review(Pergamon-Elsevier Science Ltd, 2017) Baysal, Meltem; Gunay, M. Erdem; Yildirim, RamazanIn this study, a database containing 5508 experimental data points was constructed for the steam reforming of methane using 81 papers (out of 453 initially screened) published between 2004 and 2014. The database was reviewed and analyzed with the help of decision trees to extract trends, heuristics and correlations, which are not visible to the naked eyes, through the vast experimental works accumulated in the literature over the years. The performance variable was selected as CH4 conversion while 21 variables related to catalyst preparation and operational conditions were used as input variables. It was found from a simple analysis of the literature that Ni, Rh, Ru and Pt are the most frequently used active metals, and they are generally applied over the supports of Al(2)0(3), CeO2 and ZrO2 usually using impregnation methods. A decision tree analysis was also applied to the database to determine the ranges of the catalyst preparation and operational conditions leading to high CH4 conversion. It was found for the Ni based catalysts that, even though the reaction temperature higher than 970 K is always required to achieve high CH4 conversion, some additional set of conditions are also needed; the combination of other variables especially support type and the feed composition seems to determine the catalytic performance. (C) 2016 Hydrogen Energy Publications LLC. Published by Elsevier Ltd. All rights reserved.Öğe Developing global reaction rate model for CO oxidation over Au catalysts from past data in literature using artificial neural networks(Elsevier, 2013) Gunay, M. Erdem; Yildirim, RamazanIn this work, the literature for CO oxidation kinetics over Au based catalysts was analyzed using artificial neural networks to test the possibility of developing global reaction rate models representing the entire literature. A database was constructed using the data obtained from nineteen papers published between the years 1997 and 2011; then, the reaction rate was modeled as a function of catalyst preparation and operational variables by using neural networks. Next, global reaction rate equations in the form of power law were developed for each support type by the help of the neural network model, and the order of reaction with respect to each reactant and the parameters of Arrhenius relation were estimated. These power law models were successfully validated by using the information reported in the literature; hence, it was concluded that they can be used for the initial estimation of the reaction rates in the absence of more specific rate equations. (C) 2013 Elsevier B.V. All rights reserved.Öğe Explainable machine learning analysis of tri-reforming of biogas for sustainable syngas production(Pergamon-Elsevier Science Ltd, 2025) Cosgun, Ahmet; Gunay, M. Erdem; Yildirim, RamazanIn this work, tri-reforming (TRM) of biogas was investigated using a variety of machine learning (ML) tools for knowledge extraction. For this purpose, a comprehensive database including 1183 data entries with 41 descriptors and 3 performance measures (CH4 conversion, CO2 conversion, and H2/CO ratio) was compiled from 29 articles published between 2004 and 2024. Random forest (RF) models were constructed to predict the values of performance measures that can be obtained under unknown conditions; the models were usually quite successful in the majority of the cases with the training/testing R2 values of 0.99/0.87, 0.99/0.91 and 0.96/0.58 for CH4 conversion, CO2 conversion, and H2/CO ratio respectively. To bring some explainability to the predictive models, the SHapley Additive exPlanations (SHAP) analysis was performed to determine the importance of descriptors and their effects on the performance measures. Among many results, SHAP analysis of CH4 conversion revealed the most important variable to be the reaction temperature, followed by calcination time, H2O and O2 percentages in the reaction stream, and W/F ratio. Lastly, to improve the explainability of ML even more, DT classification analysis was successfully used to generate heuristic rules that describe the combinations of individual descriptors leading to different levels of the target variables.Öğe Exploring the critical factors of algal biomass and lipid production for renewable fuel production by machine learning(Pergamon-Elsevier Science Ltd, 2021) Cosgun, Ahmet; Gunay, M. Erdem; Yildirim, RamazanIn this work, the algal biomass productivity and its lipid content were explored using a database containing 4670 instances extracted from the experimental results reported in 102 published articles. First, the influences of critical factors such as microalgae species, cultivation conditions, light intensity, CO2 amount, nutrient concentrations, reactor type, stress conditions, cell disruption methods, and lipid extraction solvents on the biomass and lipid production were reviewed. Then, the database was analyzed using machine learning techniques; decision trees were utilized to determine the combination of variables leading to high biomass and lipid content while association rule mining was used to find the specific conditions leading to very high biomass and lipid levels. Decision tree analysis discovered 11 different combinations of variables leading to high biomass productivity and 13 combinations for high lipid content; whereas, association rule mining analysis helped to identify the levels of specific factors for very high biomass and lipid production. It was then concluded that machine learning methods can help to determine the best conditions for optimum biomass growth and lipid yield for microalgae to manufacture renewable biofuels, and this can guide the planning of new experimental works. (C) 2020 Elsevier Ltd. All rights reserved.Öğe Knowledge extraction for water gas shift reaction over noble metal catalysts from publications in the literature between 2002 and 2012(Pergamon-Elsevier Science Ltd, 2014) Odabasi, Cagla; Gunay, M. Erdem; Yildirim, RamazanIn this work, a database (containing 4360 experimental data points) on water gas shift reaction (WGS) over Pt and Au based catalysts was constructed using the data obtained from the published papers between the years 2002 and 2012. Then, the database was analyzed using three data mining tools to extract knowledge in three areas: Decision trees to determine the empirical rules and conditions that lead to high catalytic performance (high CO conversion); artificial neural networks (ANNs) to determine the relative importance of various catalyst preparation and operational variables and their effects on CO conversion; support vector machines (SVMs) to predict the outcome of unstudied experimental conditions. It was concluded that, all three models were quite successful and they complement each other to extract knowledge from the past published works and to deduce useful trends, rules and correlations, which are not easily comprehensible by the naked eyes. Copyright (C) 2014, Hydrogen Energy Publications, LLC. Published by Elsevier Ltd. All rights reserved.Öğe Machine learning for a sustainable energy future(Royal Soc Chemistry, 2025) Oral, Burcu; Cosgun, Ahmet; Kilic, Aysegul; Eroglu, Damla; Gunay, M. Erdem; Yildirim, RamazanEnergy production is one of the key enablers for human activities such as food and clean water production, transportation, telecommunication, education, and healthcare; however, it is also the main cause of global warming. Hence, sustainable energy is critical for most United Nations (UN) Sustainable Development Goals (SDGs), and it is directly targeted in SDG7. In this review, we analyze the potential role of machine learning (ML), another enabler technology, in sustainable energy and SGDs. We review the use of ML in energy production and storage as well as in energy forecasting and planning activities and provide our perspective on the challenges and opportunities for the future role of ML. Although there are strong challenges for both sustainable energy supply (like conflict between the urgent energy needs and global warming) and ML applications (like high energy consumption in ML applications and risk of increasing inequalities among people and nations), ML may make significant contributions to sustainable energy efforts and therefore to the achievement of SDGs through monitoring and remote sensing to collect data, planning the worldwide efforts and improving the performance of new and more sustainable energy technologies.Öğe Machine learning for algal biofuels: a critical review and perspective for the future(Royal Soc Chemistry, 2023) Cosgun, Ahmet; Gunay, M. Erdem; Yildirim, RamazanIn this work, machine learning (ML) applications in microalgal biofuel production are reviewed. First, the basic steps of algal biofuel production are summarized followed by a bibliometric analysis to demonstrate the major research trends in the field. Also, the major challenges related to the commercialization of technology are identified. Then, ML applications for various steps in the value chain are reviewed and analyzed systematically. Finally, a future perspective on the contribution of ML in the field is provided. Our analysis indicates that ML applications should focus on screening and selecting suitable strains, preferably together with some other value-added products, requiring close collaborations among the researchers in the field to construct an extensive microalgal strain database. Optimization of cultivation conditions appears to be another area where ML can be helpful. Although most published ML works on cultivation are not usually suitable to extract generalizable knowledge (due to the nonstandard nature of strains, wastewater, and irradiation), standard testing and methodologies related to reporting protocols should also be built through collaboration to build comparable and generalizable ML models.Öğe Machine Learning for Catalytic Reaction Systems: A Framework for Complex Chemical Processes(Amer Chemical Soc, 2026) Gunay, M. Erdem; Yildirim, RamazanMachine learning (ML) and artificial intelligence (AI) have been recognized as transformative tools in chemical engineering, offering opportunities for accelerated discovery, design, and optimization. In this work, the field of catalysis is selected as a case study to demonstrate how AI/ML can be employed to complement traditional experimental and computational approaches. The integration of ML across the full scope of catalytic reaction systems is investigated from initial catalyst screening and material design to reactor development, process monitoring, and real-time control. Co-occurrence analysis for catalysis and ML was performed using 12,743 author keywords from 3924 papers, while smaller subsets for author keywords were used to analyze the co-occurrence of ML for specific aspects of catalysis; the number of papers used to review and analyze the basic trends and findings was 174, as given in references. Applications of ML in analyzing catalytic performance, characterizing structures through spectroscopic data, developing kinetic and mechanistic models, and addressing transport limitations are highlighted. Emerging strategies such as physics-informed ML, hybrid frameworks, and generative AI are regarded as particularly promising for overcoming data scarcity, interpretability, and scalability challenges. A comprehensive view of the role of ML in catalysis is presented by tracing the evolution of the field and identifying future opportunities. Consequently, a roadmap is suggested for applying similar approaches to other complex chemical engineering processes.Öğe Machine Learning-Based Analysis of Sustainable Biochar Production Processes(Springer, 2024) Cosgun, Ahmet; Oral, Burcu; Gunay, M. Erdem; Yildirim, RamazanBiochar production from biomass sources is a highly complex, multistep process that depends on several factors, including feedstock composition (e.g., type of biomass, particle size) and operating conditions (e.g., reaction temperature, pressure, residence time). However, the optimal set of variables for producing the maximum amount of biochar with the required characteristics can be determined by using machine learning (ML). In light of this, the purpose of this paper is to examine ML applications in biochar processes for the production of sustainable fuels. First, recent developments in the field are summarized, and then, a detailed review of ML applications in biochar production is presented. Following that, a bibliometric analysis is done to illustrate the major trends and construct a comprehensive perspective for future studies. It is found that biochar yield is the most common target variable for ML applications in biochar production. It is then concluded that ML can help to detect hidden patterns and make accurate predictions for determining the combination of variables that results in the desired properties of biochar which can be later used for decision-making, resource allocation, and fuel production.Öğe Machine learning-based exploration of biochar for environmental management and remediation(Academic Press Ltd- Elsevier Science Ltd, 2024) Oral, Burcu; Cosgun, Ahmet; Guenay, M. Erdem; Yildirim, RamazanBiochar has a wide range of applications, including environmental management, such as preventing soil and water pollution, removing heavy metals from water sources, and reducing air pollution. However, there are several challenges associated with the usage of biochar for these purposes, resulting in an abundance of experimental data in the literature. Accordingly, the purpose of this study is to examine the use of machine learning in biochar processes with an eye toward the potential of biochar in environmental remediation. First, recent developments in biochar utilization for the environment are summarized. Then, a bibliometric analysis is carried out to illustrate the major trends (demonstrating that the top three keywords are heavy metal, wastewater, and adsorption) and construct a comprehensive perspective for future studies. This is followed by a detailed review of machine learning applications, which reveals that adsorption efficiency and capacity are the primary utilization targets in biochar utilization. Finally, a comprehensive perspective is provided for the future. It is then concluded that machine learning can help to detect hidden patterns and make accurate predictions for determining the combination of variables that results in the desired properties which can be later used for decision-making, resource allocation, and environmental management.Öğe Recent advances in knowledge discovery for heterogeneous catalysis using machine learning(Taylor & Francis Inc, 2021) Gunay, M. Erdem; Yildirim, RamazanThe use of machine learning (ML) in catalysis has been significantly increased in recent years due to the astonishing developments in data processing technologies and the accumulation of a large amount of data in published literature and databases. The data generated in house or extracted from external sources have been analyzed using various ML techniques to see patterns, develop models for prediction and deduce heuristic rules for the future. This communication aims to review the works involving knowledge discovery in catalysis using ML techniques; the basic principles, common tools and implementation of ML in catalysis are also summarized.Öğe Statistical review of dry reforming of methane literature using decision tree and artificial neural network analysis(Elsevier Science Bv, 2018) Sener, Ayse Neslihan; Gunay, M. Erdem; Leba, Aybuke; Yildirim, RamazanThe aim of this work was to extract knowledge for dry reforming of methane (DRM) reaction from experimental data using data mining tools such as decision trees and artificial neural networks. An extensive database containing 5521 data points depending on 63 catalyst preparation and operational variables was constructed from 101 papers published between 2005 and 2014; the output variables were CH4 conversion, CO2 conversion and H-2/CO ratio of the product stream. Then, the database, as a whole or as subsets for different base metals were analyzed using decision trees (DT) to develop heuristics for high performance and artificial neural networks (ANN) to determine relative importance of input variables and predict the performance under unstudied conditions; mostly CH4 conversion, which is the most frequently reported output variable, were used in analysis. The testing accuracy of the decision tree was about 80% leading to four heuristics (i.e. four possible courses of action) for high CH4 conversion over Ni based catalyst. The first decision point to separate these heuristics is the reaction temperature as can be expected. This is followed by the other variables such as support type, W/F and reduction temperature. ANN analysis revealed that operational variables have higher relative importance (55%) compared to catalyst preparation variables (45%). The most important operational variable was found to be the reaction temperature while the active metal and the support are the most important catalyst preparation variables. ANN model was also tested to predict the data, which was not seen by the model before, and the data in 65 papers out of 101 were predicted within 15% error while 76 papers had the error rate of less than 20%.
