Çolak, Andaç Batur; Kılınç, Cuma

This study investigates machine learning techniques for predicting the behavior of proton exchange membrane electrolyzers, which are vital for sustainable hydrogen production. This work addresses these challenges by integrating artificial neural networks to develop predictive models capable of capturing the performance of proton exchange membrane electrolyzers with high accuracy. This research utilizes a multi-layer perceptron network architecture, optimized through rigorous data preprocessing, parameter tuning, and error minimization strategies. The dataset used was based on published PEME numerical simulation datasets and encompasses key performance indicators, including stack voltage, water transport, and electrochemical reactions. The trained artificial neural networks models achieved mean squared error values of 3.66 × 10−5 and 9.75 × 10−6, with correlation coefficients of 0.99996 and 0.99958, demonstrating near-perfect predictive accuracy. A comparative benchmarking study against alternative regression algorithms revealed that the proposed MLP models significantly outperformed Gradient Boosting and Random Forest by several orders of magnitude, thereby establishing a higher level of persuasiveness and reliability for the developed framework. Average deviation rates of 0.11% and −0.01% further validated model reliability. The novelty of this work lies in its comprehensive approach, which goes beyond isolated metrics by addressing interactions across system parameters. This integrated framework enables enhanced prediction, control, and optimization of proton exchange membrane electrolyzer’s performance, setting a new benchmark for leveraging machine learning in hydrogen energy systems. These findings pave the way for scalable, cost-effective solutions to improve proton exchange membrane electrolyzers’ efficiency and operational reliability.

Demirci, Oğuz Kürşat

Five pathways—diesel, renewable hydrogen, green ammonia, Fischer–Tropsch e-diesel, and B100 waste cooking oil fatty acid methyl ester (WCO-FAME)—were compared on a common service basis using a reduced-order engine exergy model, a partial techno-economic cost model, and 10,000 Latin hypercube simulation samples; each alternative was paired with diesel by sample index. The results were normalised to 1 kWh of net brake work (kWhb) and 1000 tonne-kilometres of freight work (1000 tkm). A carbon balance back-calculation from European Commission Vehicle Energy Consumption Calculation Tool (VECTO) monitoring data for subgroup 5-Long Haul (5-LH; a 4 × 2 tractor–semitrailer) yielded 89.31 kWhb/1000 tkm; first-order payload penalties increased this requirement to 94.79 kWhb/1000 tkm for hydrogen and 91.97 kWhb/1000 tkm for ammonia. Deterministic central-case exergy efficiencies were 43.02% for hydrogen, 39.67% for e-diesel, 39.35% for diesel, 37.56% for B100, and 33.81% for ammonia. The study-specific partial cost was expressed in 2025 EUR per 1000 tkm and included delivered fuel, maintenance, and annualised incremental vehicle capital; common fleet costs, including driver, tyre, toll, insurance, and tax costs, and externalities were excluded. Therefore, this metric was not a full freight cost or tariff. Deterministic central-case costs were EUR 25.85/1000 tkm for diesel, EUR 26.72/1000 tkm for B100, EUR 58.64/1000 tkm for hydrogen, EUR 59.84/1000 tkm for ammonia, and EUR 67.44/1000 tkm for e-diesel. Hydrogen exceeded diesel in exergy efficiency in 96.50% of the paired simulation samples; all three electrofuel pathways were costlier than diesel in every paired simulation sample, whereas B100 was cheaper than diesel in 35.62% of the paired simulation samples. Thus, hydrogen led in exergy efficiency, diesel in cost under the defined partial economic boundary, and B100 was the closest cost competitor. These are conditional model-screening results rather than a completed comparative life-cycle assessment: no quantitative life-cycle impact assessment (LCIA), environmental ranking or marginal abatement cost result is reported.

Durmaz, Lokman; Karageçili, Hasan; Erturk, Adem; Ozden, Eda Mehtap; Taslimi, Parham; Alwasel, Saleh; Gülçin, İlhami

Hamamelitannin (2′,5-di-O-galloyl-hamamelose) bears two-gallate moieties in its structure, and is a natural phenolic product in the leaves and the bark of Hamamelis virginiana. The antioxidant capacity of hamamelitannin was evaluated by a range of methods, with the following findings: the ability to reduce potassium ferric cyanide; the scavenging of N,N-dimethyl-p-phenylenediamine dihydrochloride radical (DMPD•+); the scavenging of 2,2′-azinobis-(3-ethylbenzothiazoline-6-sulphonate) radical (ABTS•+); the scavenging of 1,1-diphenyl-2-picrylhydrazyl radical (DPPH•); and the ability to reduce cupric ions (Cu2+). Additionally, reference antioxidants of α-Tocopherol, butylated hydroxyanisole (BHA), Trolox, and butylated hydroxytoluene (BHT) were used for comparison. For DPPH radical scavenging, hamamelitannin had an IC50 value of 19.31 μg/mL, while the IC50 values for BHA, BHT, Trolox, and α-Tocopherol were 10.10, 25.95, 7.05, and 11.31 μg/mL, respectively. The study found that hamamelitannin functioned similarly to BHA, α-tocopherol, and Trolox in terms of DPPH• scavenging, but better than BHT. Additionally, as a polyphenolic secondary metabolite, the hamamelitannin inhibition capability of several metabolic enzymes was demonstrated, including acetylcholinesterase (AChE), butyrylcholinesterase (BChE), carbonic anhydrase I (CA I), carbonic anhydrase II (CA II) and α-glycosidase. The Ki values of hamamelitannin exhibited 7.40, 1.99, 10.18, 18.26, and 25.79 nM toward AChE, BChE, hCA I, hCA II, and α-glycosidase, respectively.

Alic, Erdem; Das, Mehmet; Kaska, Onder

It is difficult to manually process and analyze large amounts of data. Therefore, to solve a given problem, it is easier to reach the solution by studying the data obtained from the environment of the problem with computational intelligence methods. In this study, pool boiling heat flux was estimated in the isolated bubble regime using two optimization methods (genetic and artificial bee colony algorithm) and three machine learning algorithms (decision tree, artificial neural network, and support vector machine). Six boiling mechanisms containing eighteen different parameters in the genetic and the artificial bee colony (ABC) algorithms were used to calculate overall heat flux of the isolated bubble regime. Support vector machine regression (SVMReg), alternating model tree (ADTree), and multilayer perceptron (MLP) regression only used the heat transfer equation input parameters without heat transfer equations for prediction of pool boiling heat transfer over a horizontal tube. The performance of computational intelligence methods were determined according to the results of error analysis. Mean absolute error (MAE), root mean square error (RMSE), and mean absolute percentage error (MAPE) error were used to calculate the validity of the predictive model in genetic algorithm, ABC algorithm, SVMReg, MLP regression, and alternating model tree. According to the MAPE error analysis, the accuracy values of MLP regression (0.23) and alternating model tree (0.22) methods were the same. The SVMReg method used for pool boiling heat flux estimation performed better than the other methods, with 0.17 validation error rate of MAPE.