The conventional experimental procedure involving Titanium–Nickel (Ti-Ni) shape memory alloys requires conducting dozens or even hundreds of heating and cooling cycles performed by the actuator to generate thermal hysteresis curves. This study proposes the development of a proxy model based on machine learning techniques, using experimental results, with the goal of replicating the actuator's function in this experiment. The proxy model should be capable of accurately predicting the actuator’s thermomechanical response based on time series data of heating and cooling cycles over time. It is important to highlight that this is not the traditional time series forecasting problem aimed at predicting future values, but rather a problem of predicting the dynamic responses of the actuator associated with new input profiles (temperature, mechanical stress, and strain). The proposed strategy is based on the use of deep neural network algorithms, aiming to capture the actuator’s dynamics from experimental data. The main architecture used for modeling temporal dependencies is the recurrent neural network (RNN), specifically the Long Short-Term Memory (LSTM) type, known for its ability to extract complex and nonlinear temporal patterns in time series data. To evaluate the performance of the proxy model, an experimental dataset was generated using a helical spring-shaped actuator under load. The model's predictions were compared with the experimentally obtained hysteresis curve in order to validate its generalization capability. The results demonstrate that the proposed technique is highly promising, achieving a mean squared error on the order of 1.2%.

Terahertz (THz) metasurfaces enable subwavelength electromagnetic wave control, offering a path to compact and tunable filters for spectroscopy, sensing, and communications. Here, we present a simulation-driven design for a compact multilayer THz bandpass metasurface filter [1]. The 3D device geometry consists of a silicon substrate supporting a thin gold film perforated with subwavelength annular apertures, a high-transmittance dielectric spacer (PTFE or HDPE), and an aligned array of gold rings on top. This stack can be fabricated by depositing gold through annular masks, eliminating the need for lithographic etching of the substrate presented in our previous model [2].

Full-wave electromagnetic simulations (COMSOL) guided the design process, optimizing spacer thickness and geometric parameters. Simulated transmission spectra show that devices with either PTFE or HDPE spacers yield broad passbands with high transmission. The spectral peak position and shape remain largely invariant under minor geometric deviations, indicating a robust design. The enhanced transmission is attributed to constructive interference between waves reflected at multiple interfaces in the multilayer structure. The combination of a simple mask-defined architecture, material-dependent tunability, and tolerance to fabrication imperfections makes this metasurface filter a promising candidate for THz bandpass filtering, anti-reflection coatings, spectroscopy, and biosensing. This work highlights the effectiveness of computational design in advancing THz metasurface technologies.

Foundation

This work was supported by Grants No. 22rl-056 and 24AA-2J068 of the Higher Education and Science Committee of the RA MoESCS.

References
[1] G.-M. Li, et al. Terahertz bandpass and bandstop filter based on the babinet complementary metamaterials, Optics Communications 571, 130944 (2024).

[2] K. Simonyan, et al. Broadband THz metasurface bandpass filter/antireflection coating based on metalized Si cylindrical rings, Semicond. Sci. Technol. 39 095012 (2024).

Increased use of plastic is one of the main reasons for pollution. It takes about 20 to 500 years for traditional plastic to decompose; hence, it finds its way into landfills or oceans. The solution is to produce starch-based biodegradable plastic. The main ingredient of this would be starch, which is organic and natural, and all sorts of toxic additives would be avoided to ensure that plastic remains one hundred percent organic. This would allow it to come from the soil and go back to it without causing harm, completing a full natural life cycle. This study aimed to synthesize bioplastic, and involved the design and development an automated machine capable of producing biodegradable plastic bags using starch-based polymer films. The system described in this study integrates a mixing unit followed by a film-sliding mechanism, drying, and a thermal sealing unit, all controlled via Arduino to ensure precision and repeatability. Compared to conventional machines, this design consumes less energy and supports environmentally friendly materials. It efficiently processes starch-based polymers into usable bags with acceptable mechanical properties and biodegradability. It is an alternative to traditional plastic, which is vital, as removing the use of plastic from day-to-day life appears to be an impossible task. It contributes to sustainable manufacturing and serves as a foundation for further innovations in green packaging technologies.

This work presents a comparative study on the mechanical homogenisation of Triply Periodic Minimal Surface (TPMS) lattice structures, which have attracted significant interest for their unique ability to combine lightweight design with tailored mechanical, thermal, and acoustic properties. The study investigates the effective mechanical behaviour of Representative Unit Cells (RUCs) generated using the open-source Python tool Microgen. Two homogenisation strategies are considered: (i) Finite Element (FE)-based homogenisation carried out in Abaqus, and (ii) the Mechanics of Structure Genome (MSG), a unified theory for multiscale constitutive modelling, implemented in a specialized software framework. The comparison encompasses multiple TPMS topologies, including well-studied cases used for validation as well as less-explored ones to provide new insights, namely gyroid, diamond, PMY, and F-Rhombic Dodecahedron (F-RD). RUCs are analysed across relative densities ranging from 10% to 50%. Equivalent linear elastic properties (Young’s moduli, shear moduli, and Poisson’s ratios) are derived and compared to assess the consistency, accuracy, and computational efficiency of the two approaches. Furthermore, the anisotropy of each TPMS topology across the range of relative densities is examined through the directional distribution of Young’s moduli. The outcomes are expected to clarify the strengths and limitations of FE versus MSG in capturing the effective behaviour of architected cellular solids, thus supporting the selection of homogenisation strategies for the design of lattice-based lightweight structures.

The present talk addresses the extension of hierarchical shell finite elements based on Carrera’s Unified Formulation (CUF) to a global-local approach for the investigation of Variable-Angle Tow (VAT) composite structures. VAT laminates are characterized by curvilinear fibres laid along predefined paths, enabling enhanced mechanical performance and a wider structural design space. Nevertheless, their analysis typically requires high computational effort to accurately capture the complex displacement and stress fields resulting from the variable in-plane fibre distribution. In the proposed strategy, a global analysis is first performed over the entire structural domain using low-order Abaqus shell elements with a reduced number of degrees of freedom. A subsequent local analysis employs a refined CUF model with higher-order through-the-thickness approximations in a layer-wise manner, enabling accurate and efficient capture of the high stress gradients that typically arise near geometric singularities and discontinuities. The governing equations are derived within the CUF framework using both the Principle of Virtual Displacements (PVD) and the Reissner’s Mixed Variational Theorem (RMVT). Validation against full three-dimensional finite element simulations in Abaqus demonstrates the accuracy of the proposed methodology. Comparisons in terms of degrees of freedom confirm that the global–local CUF-based approach achieves high accuracy near discontinuities at a significantly reduced computational cost relative to full 3D models. Furthermore, the differences observed between the PVD and RMVT formulations highlight the critical role of transverse stress prediction in the analysis of VAT composites.

It is important to quantitatively and graphically characterize the four core effects, the most fundamental yet disputable issues of high-entropy alloys. Yet, the traditional and commonly believed special quasirandom structure (SQS) based on the prefect random mixing structure hypothesis is insufficient as the SQS model ignores the difference of the types of different constituent atoms, the difference of the types of different crystal lattice structure, such as FCC, BCC and HCP, and the difference of the different heat treatment temperatures. In this contribution, based on crystal structure, we propose an alloy thermodynamics model based on the crystallographic structure and then establish the thermodynamic database of the end-member involved by combining computational thermodynamics and first-principle calculations. Thus, the four core effects of high-entropy alloys with various phase structures were quantitatively and graphically characterized, including the site occupying fractions (SOFs), and then the atomic distribution model construction based on SOFs, short-range ordering (SRO) cluster, diverse mechanical property, interstitial atom diffusion, and catalytic characteristic of selected high-entropy alloys. Meanwhile, such behaviors of the commonly believed SQS based on the prefect random mixing structure were also simulated and compared with those of the SOF structures. We conclude that it is quite necessary and also feasible to consider the inherent and inevitable sublattice preference of constituent atoms to simulate the structure and diverse properties of HEAs theoretically, which extends beyond the commonly believed but baseless SQS based on the random mixing hypothesis.

This study details a nonlinear buckling and post-buckling analysis of Functionally Graded Material (FGM) plates, which are increasingly utilized in advanced engineering applications like aerospace due to their superior thermal and mechanical properties. The design and integrity of these structures under complex loading conditions, however, pose significant challenges, particularly regarding their stability under compressive forces.To address this, the structural formulation is based on the First-Order Shear Deformation Theory (FSDT), which is well-suited for capturing shear deformation effects that become significant in thin to moderately thick structures. By minimizing the total potential energy of the system, a comprehensive mathematical model is derived. The subsequent solution of the highly complex nonlinear governing equations is achieved through an innovative hybrid numerical approach. This method leverages the robust Asymptotic Numerical Method (ANM), a powerful continuation technique known for its ability to efficiently trace complex equilibrium paths, and integrates it with the versatile Finite Element Method (FEM). This combination effectively handles the spatial discretization of the plate while providing a stable and accurate means to track the structural behavior beyond the critical buckling point.The presented approach is particularly effective in accurately identifying critical buckling loads and precisely locating bifurcation points along the equilibrium paths. The method's effectiveness is demonstrated by its high computational efficiency and accuracy, which are critical for the design and safety analysis of FGM components. The findings of this research provide a reliable tool for structural engineers to analyze the stability and behavior of FGM plates, contributing to the development of more resilient and efficient structural designs.

Bithiophene-based molecular systems have emerged as a cornerstone in the field of organic electronics owing to their remarkable structural versatility, strong π conjugation, and favorable optoelectronic properties. Their relatively simple chemical framework makes them ideal building blocks for more complex thiophene-based oligomers and polymers, widely used in organic semiconductors. In this work, we specifically investigate a dimer composed of two parallel bithiophene chains in order to evaluate the impact of intermolecular coupling on the electronic structure and optical response of the system. The parallel configuration is of particular interest because it mimics the π–π stacking interactions commonly observed in thin films and crystalline domains of organic materials.

Our analysis reveals that the interaction between the π orbitals of the two chains significantly modifies the energy levels of the dimer. The effective coupling reduces the electronic band gap, leading to a pronounced red shift in the absorption spectrum and enhancing the oscillator strength of the main electronic transitions. This results in improved light-harvesting ability, an essential requirement for photovoltaic and photodetector applications. Moreover, the parallel stacking promotes stronger charge delocalization along the conjugated framework, thereby lowering the reorganization energy and enhancing charge transfer efficiency. As a consequence, both electron and hole mobilities are improved, which is crucial for charge transport in optoelectronic devices.

These findings underline the central role of interchain interactions in dictating the performance of bithiophene-based materials. By controlling stacking geometry and intermolecular distances, it becomes possible to tune their electronic and optical properties, paving the way toward high-performance organic field-effect transistors (OFETs), organic photovoltaic cells (OPVs), and organic light-emitting diodes (OLEDs).

Biodegradable polymeric materials such as Polylactic Acid (PLA) and Polycaprolactone (PCL) are proven to be a good choice in the design of biopolymeric devices for tissue engineering applications for osteochondral implants such as scaffolds. The behaviour of these materials has been submitted to several studies and numerical models have been developed to predict the behaviour of such materials when implanted in the damaged tissue. When talking about amorphous polymers, there is a predominance of the degradation process of the polymeric material, and the surface erosion process. Here, a novel stable probabilistic-deterministic numerical tool developed on FreeFem++ to predict the erosion and degradation behavior of polymeric materials of biodegradable polymers is presented. The erosion model is based on a stochastic approach using cellular automated distribution. The degradation model is based on the Fick Law of diffusion of materials, whereas the surface erosion model is considered an stochastic process, and modelled using a Monte Carlo simulation technique. Furthermore, to validate the erosion mechanism, a porosity function is described, in order to compare the results with the experimental data. In order to gain more stability in the methodology, a Fictitious Domain Technique is implemented in order to describe the changes on the boundary during the erosion process.

The pollution caused by pharmaceutical contaminants has become a pressing environmental concern, as their persistent accumulation in ecosystems poses serious risks to water quality, food safety, and human health. The adsorption of pollutants onto activated carbon is a simple yet highly efficient method for water pre-treatment, widely applied at an industrial scale in both desalination and water treatment processes. The characterization of carbon adsorbents represents a broad area of research. While chemically synthesized adsorbents have predictable structural and chemical properties that are determined by their synthesis, activated carbons often exhibit considerable variability due to differences in raw materials and activation methods. Converting low-cost organic waste into high-value materials without generating secondary pollutants remains a key challenge for sustainable and eco-friendly industrial development. In this study, an inexpensive and readily available carbon precursor was selected and converted into a high-performance adsorbent. Activated carbon was synthesized at the laboratory scale from organic waste to produce a cost-effective sorbent. The precursor material was characterized before and after treatment using XRD, SEM/EDS, and FTIR analyses. Adsorption experiments were carried out to investigate the removal of pharmaceutical contaminants, with a focus on paracetamol, using both the synthesized activated carbon and the raw precursor material. This study examines adsorption kinetics, equilibrium isotherms, and the effects of critical parameters, including contact time, initial pollutant concentration, adsorbent dosage, and solution pH.