Technical University of Cluj-Napoca

This paper presents an adaptive fuzzy control scheme capable of guaranteeing prescribed performance for stochastic nonlinear systems with unknown control directions. Unlike the majority of existing prescribed performance control schemes, the proposed scheme ensures the independence from initial errors and guarantees controllable overshoot. Moreover, the proposed prescribed function exhibits nonmonotonicity, which can be beneficial in control applications with input constraints. To address the challenge posed by unknown control directions, a novel class of multiple Nussbaum functions is introduced. Compared to the existing single Nussbaum function, the multiple Nussbaum functions can mitigate instability arising from the cancellation of multiple unknown signs. Additionally, to tackle unknown nonlinearities, a single-parameter fuzzy approximator is introduced, aiming to concurrently reduce computational complexity. Furthermore, a novel class of switching threshold event-triggered mechanisms is designed to address issues encountered in existing designs where parameter inequalities impose conservative constraints. The control scheme ensures that the tracking error converges to prescribed asymmetric boundaries with arbitrarily small residuals in a prescribed time, while also guaranteeing that all closed-loop signals are bounded in probability. The effectiveness and superiority of the control scheme are verified by simulation results.

The growing urgency of environmental challenges and the depletion of fossil fuels have accelerated the search for sustainable and renewable energy sources. Wind energy, for example, is an important source of green electricity. However, using wind power is challenging due to the variability and unpredictability of wind patterns. Consequently, the ability to predict wind power in advance is crucial. The integration of artificial intelligence within the renewable energy sector could provide a viable solution to this challenge. In this study, we investigate the potential of machine learning to improve wind power forecasting by conducting a comparison of three regression models: K-Nearest Neighbor regression, Random Forest regression, and Support Vector regression. These models are combined with a feature selection technique to forecast wind power. Additionally, we propose a novel hybrid approach that combines these machine learning models with Multiple Linear Regression to address the complexities of wind energy forecasting. The performance of the models is evaluated using the R² score, Mean Absolute Error, and Root Mean Squared Error. The dataset for this study was generated from a numerical simulation conducted at a location with a latitude of 22.55° N and a longitude of -14.33° E. The findings demonstrate that the proposed hybrid model outperforms the individual machine learning models in terms of prediction accuracy. This study provides a solid foundation for future research and development in wind energy forecasting.

Natural drug products with limited water solubility pose a challenge to the pharmaceutical industry in terms of developing an appropriate dissolution procedure. Like other flavonoids, icariside II (ICS) also faces challenges such as insufficient bioavailability caused by its poor water solubility, limiting its oral therapeutic applications as a food supplement. Consequently, new approaches are needed across the board focusing on the aqueous solubility enhancement. This work aimed to use whey protein concentrate (WPC) as a carrier, and surfactants such as Tween 80 and lecithin to improve the water solubility of icariside II. The complexation with WPC successfully increased the water solubility of ICS by approximately 258-fold. Furthermore, the incorporation of surfactants into the complex resulted in an even greater enhancement, achieving a 554-fold improvement. In addition, by repurposing of whey protein complexes, a byproduct generated during cheese production, and the application of efficient solvent recovery methods, we illustrate our dedication to sustainability. Fourier transform infrared spectroscopy; X-ray powder diffraction and differential scanning calorimetry indicated the successful complex formation procedure. Using a scanning electron microscope, the morphology of the product was analyzed. To the best of our knowledge, no studies have yet produced and investigated the aqueous solubility of mixed surfactant-based icariside II whey protein complexes (S-ICS-WPC).

Nitrogen-doped zinc oxide (N:ZnO) thin films were deposited on glass substrates via radio frequency (RF) magnetron sputtering and subsequently annealed at 300 °C, 400 °C, 500 °C, and 600 °C to assess their viability and stability as transparent conductive oxide (TCO) materials. Structural and compositional analyses were performed using X-ray diffraction (XRD), field emission scanning electron microscopy (FESEM), and X-ray photoelectron spectroscopy (XPS). XRD analysis revealed preferential crystallite orientations along the (100), (002), (101), and (110) planes. Atomic force microscopy (AFM) measurements indicated particle sizes two to four times larger than those derived from XRD, suggesting a sub-granular internal structure, as XRD probes coherently diffracting domains. XPS analysis of the N 1 s spectra identified two distinct peaks at approximately 397 eV and 407.5 eV, indicating nitrogen incorporation into the ZnO matrix. Photoluminescence spectroscopy revealed that nitrogen doping induced the formation of interstitials and defects associated with oxygen and zinc vacancies. Optical measurements showed that the (N:ZnO) thin films exhibited an average optical band gap of approximately 3.1 eV, with 80% transmittance in the visible spectrum. A linear relationship was observed between the band gap energy and the tail width. Except for the film annealed at 600 °C, all annealed films showed a reduction in peak photoluminescence intensity with increasing annealing temperature. Finally, no significant changes in the electrical performance of the p-N/n-Si diode were observed as a result of annealing-induced surface modifications. The results provide valuable insights into the optimization of (N:ZnO) thin films for use in international optoelectronic and photovoltaic research, where advancements in TCOs are critical for the development of high-performance, sustainable technologies.

Accurately detecting and localizing vineyard disease detections are essential to reduce production losses. A great variety of scientific work focuses on remote sensing methods, while with current learning-based techniques, a continuous paradigm shift is happening in this domain. Based on a thorough literature review, the need for a remote assistance survey for the detection of vine disease was motivated by the adoption of recent machine learning algorithms. Thus, in this work, the research outputs from the past few years are summarized in the domain of grapevine disease detection. A remote sensing-based distance taxonomy was introduced for different categories of detection methods. This taxonomy is relevant for differentiating among the existing solutions in this domain, the resulting methods being grouped according to the proposed taxonomy. The articles and public datasets cited are collected on the website of this project ( https://molnarszilard.github.io/VinEye/ ).

This paper revisits sparse grid integration proposed in the literature for approximating integrals that occur as choice probabilities in random coefficient discrete choice models. First, we successfully replicate their main findings for the panel mixed logit. Second, for higher variances and for a different structure of the variances of the random coefficients, in certain cases, we fail to replicate the original results. Third, for the important special case of cross‐sectional mixed logit, replication of the original results is successful when the number of alternatives is moderate but fails otherwise.

Heavy metal pollution has complex impacts on terrestrial ecosystems, affecting biodiversity, trophic relationships, species health, and the quality of natural resources. This study aims to validate a non-invasive method for detecting heavy metals (Cd, As, Zn, Cu, Cr) in micromammalian prey, which constitute the primary diet of the common genet (Genetta genetta), a mesocarnivore sensitive to habitat degradation. By focusing on prey remains (hair and bones) rather than entire fecal samples, this approach leverages the genet’s selective feeding habits to assess the bioaccumulation of contaminants in its preferred prey. Conducted in the Edough forest massif during the winter of 2021, the study analyzed 39 fecal samples, collected from the following two contrasting environments: a natural habitat and an area impacted by an open landfill. Results revealed significant levels of heavy metals, with higher concentrations in bones compared to hair, and increased accumulation in prey from the anthropized environment. Monitoring these contaminants in selective predators, such as the genet, offers a promising approach to better understanding environmental contamination and implementing measures to protect ecosystems and the species that depend on them.

The classic constitutive model of metal plasticity employs the concept of yield surface to describe the strain‐stress response of metals. Yield surfaces are constructed as level sets of yield functions, which in turn are assumed to be homogeneous, smooth and convex. These properties ensure the mathematical consistency of the constitutive model while also facilitating the calibration of the yield function. The significant progress in computing hardware and software of the last two decades has opened new possibilities for research into general‐purpose yield functions that are capable of capturing with high accuracy the mechanical properties of sheet metal. Here we investigate the modeling capabilities of yield functions defined by homogeneous, smooth and convex neural networks (HSC‐NN). We find that small‐sized HSC‐NNs are capable of reproducing a wide range of convex shapes. This type of network is then ideally suited to data‐driven frameworks based on virtual testing or on interpolation of data from mechanical tests, being easy to deploy in finite element codes. HSC‐NNs are particularly adept at fitting aggregations of plane stress and out‐of‐plane data to build yield surface models accounting for 3D‐stress states. We use them here to bring new insights into a recent cup‐drawing experiment with aluminum alloy AA6016‐T4. Finite element simulations with both plane stress and 3D models show promising results. In particular, the overall simulation run times of the HSC‐NNs employed here are found to be comparable with those of conventional yield functions.

One of the leading challenges in Water Resource Recovery Facility monitoring and control is the poor data quality and sensor consistency due to the tough and complex circumstances of the process operation. This paper presents a new principal component analysis fault detection approach for the nitrate and nitrite concentration sensor based on Water Resource Recovery Facility measurements, together with the Fisher Discriminant Analysis identification of fault types. Five malfunction cases were considered: constant additive error, ramp changing error in time, incorrect amplification error, random additive error, and unchanging sensor value error. The faults’ implementation, fault detection, and identification methods are presented and evaluated in terms of accuracy and promptitude. The models are originating from a municipal plant. The amount of required electrical energy and greenhouse gas released during the Water Resource Recovery Facility operation were assessed for the cases of nitrates and nitrites NO sensor normal and malfunctioning regimes. The environmental and economic evaluations show the benefits of detecting and identifying nitrates and nitrites NO sensor defects aimed at providing efficient and environmentally friendly operation of the Water Resource Recovery Facility. The fault-affected operation cases showed increased values, up to 10% for the total energy demand and 4% for the total greenhouse gas emissions, when they are compared to the normal operation case.

The increasing demand for high-performance materials in industrial applications highlights the need for composites with enhanced mechanical and tribological properties. Basalt fiber-reinforced polymers (BFRP) are promising materials due to their superior strength-to-weight ratio and environmental benefits, yet their wear resistance and tensile performance often require further optimization. This study examines how adding copper (Cu) powder to epoxy resin influences the mechanical and tribological properties of BFRP composites. Epoxy matrices, modified with 5%, 10%, and 15% weight fractions (wf.%) of copper powder, were reinforced with BFRP-type fabric, using a vacuum bag manufacturing method. Mechanical tests, including bending and tensile tests, showed notable improvements in tensile strength and flexural modulus due to copper addition, with higher copper (Cu) content enhancing ductility. Tribological tests using a pin-on-disk tribometer revealed reduced wear rates and an optimized coefficient of friction. Statistical analysis and 3D microscopy identified wear mechanisms such as delamination and protective copper film formation. The results highlight the significant potential of copper-modified BFRP composites for applications demanding superior mechanical and tribological performance.

The phase-shifted converter (PSFB) is one of the most used converters for applications where high power and galvanic isolation are required. To achieve zero voltage switching (ZVS) on the primary side these converters use the leakage inductance of the main transformer and the parasitic capacitances of the primary bridge MOSFET transistors. The higher this inductance is, the lower the output power that ensures ZVS. However, the resonant inductor, together with the parasitic capacitances of the secondary bridge and the transformer’s secondary winding, induce ringing and high voltage spikes at the secondary bridge output. To mitigate this problem snubber circuits are used. These converters’ maximum power depends on the energy stored in leakage inductances and parasitic capacitances, which relate to the choice of the snubber circuits used. The present paper analyzes in detail the behavior of an active clamp circuit. It determines the conditions for zero voltage turn ON of the snubber transistor thus improving the snubber’s efficiency and of the entire phase-shifted converter. All the results are validated through simulation and experimental models.

The problem of hateful expression in Romania is a recurrent one and is based on deep-seated prejudices of the Romanian society. Those prejudices are reflected in the discrepancy between the ‘legal’ realm – composed of the applicable national laws and international treaties – and the ‘real’ realm – which is made up of numerous incidents amounting to hate speech and the passive reaction of the Romanian authorities when confronted with them. Thus, the report on Romania aims at highlighting the fact that simply adopting laws and ratifying international treaties is useless in the absence of a change of mindset towards truly believing in the equality of all people.

The HADDOCK team participated in CAPRI rounds 47–55 as server, manual predictor, and scorers. Throughout these CAPRI rounds, we used a plethora of computational strategies to predict the structure of protein complexes. Of the 10 targets comprising 24 interfaces, we achieved acceptable or better models for 3 targets in the human category and 1 in the server category. Our performance in the scoring challenge was slightly better, with our simple scoring protocol being the only one capable of identifying an acceptable model for Target 234. This result highlights the robustness of the simple, fully physics‐based HADDOCK scoring function, especially when applied to highly flexible antibody–antigen complexes. Inspired by the significant advances in machine learning for structural biology and the dramatic improvement in our success rates after the public release of Alphafold2, we identify the integration of classical approaches like HADDOCK with AI‐driven structure prediction methods as a key strategy for improving the accuracy of model generation and scoring.

Durrmeyer type modifications of the Szász–Mirakjan operators were studied in several papers. We consider a general family of such operators. Basically, we investigate the relationship between the images of a function f under the operators and the solutions of a suitable linear and homogeneous partial differential equation of parabolic type with f as initial value.

Endometriosis, a chronic hormone-dependent condition affecting 10% of women globally, impacts pelvic organs and occasionally distant sites, causing pain, infertility, and sexual dysfunction. Biomarkers such as IL-8, IL-10, and BDNF influence inflammation, nerve sensitization, and pain. This study investigates their relationship with sexual quality of life, focusing on dyspareunia and related dysfunctions, as assessed using the Female Sexual Function Index (FSFI). Dyspareunia, a prominent symptom of endometriosis, is linked to lower FSFI scores in domains such as desire (mean 3.38), satisfaction (mean 3.28), and pain (mean 3.07). Elevated IL-8 tissue levels negatively correlated with desire (r = −0.649, p < 0.05) and satisfaction (r = −0.813, p < 0.01). Similarly, higher BDNF tissue levels were associated with increased pain (r = −0.435, p < 0.01) and reduced satisfaction (r = −0.252, p < 0.05). Patient factors such as higher endometriosis severity scores (mean 26.3, p < 0.05) and surgical history correlated with lower desire and satisfaction. Conversely, physical activity improved pain scores (p < 0.01) and enhanced desire and lubrication (p < 0.05), likely through reduced inflammation and better circulation. These findings highlight the complex interplay between biomarkers, individual factors, and sexual dysfunction in endometriosis, underscoring the need for personalized therapeutic approaches.

As a result of the restrictions caused by the pandemic, education has moved to another space. Online conditions largely override all the methods that can be applied offline. The online interface, as an educational platform, requires a different kind of competence. The extent to which heuristic, Socratic methods can be put into practice in online education depends primarily on the creativity of those working on the front lines. The aim of this study is to illustrate how fiction can be applied — in this case, Émile Zola’s novel The Ladies’ Delight, primarily in the teaching of commercial marketing — in order for students to be emotionally involved in the educational process and to be part of the experiential learning during a pandemic.

This study explores the experimental and theoretical optimization of process parameters to improve the quality of 3D-printed parts produced using the Fused Deposition Modeling technique. To ensure the cost-effective production of high-quality components, advancements in printing strategies are essential. This research identifies optimal 3D printing strategies to enhance the quality of finished products. Form and dimensional tolerances were assessed using a 3D Coordinate Measuring Machine, and the resulting data were analyzed via Design Expert software version 9.0.6.2. Design Expert for experimental design was utilized and an Analysis of Variance was conducted to validate the models’ accuracy. The results indicate that a 45° raster angle, combined with internal raster values between 0.5048 and 0.726, minimizes flatness, cylindricity, and dimensional deviations by optimizing deposition patterns and thermal dynamics. Internal raster values below 0.308 resulted in insufficient support and greater deviations, while higher values enhanced stability through improved interlayer adhesion. Experimental validation confirmed these parameter settings as optimal for producing precise and consistent 3D-printed parts.

Bakground: The mortality rate from community-acquired pneumonia (CAP) or coronavirus disease 19 (COVID-19) is high, especially in hospitalized patients. This study aimed to assess the disturbances of glucose and lipid metabolism with in-hospital complications and short-term outcomes for patients with pneumonia with different etiologies. Methods: This observational study comprised 398 patients divided as follows: 155 with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pneumonia, 129 participants with viral CAP, and 114 with bacterial pneumonia. Results: Fasting plasma glucose (FPG) at admission and glycemic variation during hospitalization was linked with acute kidney injury (AKI) in bacterial CAP. Compared with a value <110 mg/dL for FPG at admission, levels between 110 and 126 mg/dL are associated with mortality in both COVID-19 (OR = 3.462, 95% CI: 1.275–9.398, p = 0.015) and bacterial CAP participants (OR = 0.254; 95% CI: 0.069–0.935, p = 0.039), while a value ≥126 mg/dL was linked with mortality only in patients with SARS-CoV-2 (OR = 3.577, 95% CI: 1.166–10.976, p = 0.026). No relation between lipid biomarkers and complications or in-hospital outcomes was observed in all three participant groups. Conclusions: Patients with bacterial CAP are more prone to developing AKI due to increased FBG at admission and glycemic variations during hospitalization, while elevated FBG values at admission are associated with mortality in both COVID-19 and bacterial CAP.

This paper analyzes seven substantial distinct classes of contractive-type mappings using the technique of geodesic average perturbation within the framework of CAT(0) spaces. These classes of mappings are shown to be either saturated, in the sense that the geodesic average perturbation technique does not yield any significant new fixed point results, or unsaturated, in the sense that the technique provides genuine new fixed point results. The results establish that the class of strictly pseudocontractive self-mappings and the class of demicontractive self-mappings are saturated. Furthermore, the unsaturated category includes the class of Banach contractions, the class of Kannan contractions, the class of Bianchini contractions, the class of nonexpansive mappings, and the class of Ćirić–Reich–Rus contraction mappings. Our findings extend results from Hilbert spaces to convex (in the geodesic sense) metric spaces. This work provides an avenue for investigating fixed point results for several other important classes of contractive mappings using the geodesic average perturbation technique within the framework of geodesic spaces such as Hadamard manifolds, Hilbert balls, hyperbolic spaces, and CAT(k) spaces for some $k \in {\mathbb {R}}$.