• Real-LIFE test protocol for log wood stoves developed considering all phases. • Validation with two log wood stoves performed using the Real-LIFE test protocol. • Good repeatability of emissions while applying the new test protocol achieved. • Challenging to get comparable results from four laboratories using new protocol. • Study emphasizes the importance of measuring different combustion conditions. The use of log wood stoves is common in residential homes and are tested in a type test procedure following EN 16510-1:2022 at optimal combustion condition. Since this does not represent real-life operation, a novel test protocol was developed and validated using two different log wood stoves. The new test protocol includes the ignition phase (two batches) at natural draught, followed by three batches at nominal load, two batches at partial load and one final batch at overload. Typical emission parameters such as carbon monoxide (CO), nitrogen oxides (NO X ), organic gaseous carbon (OGC) emissions were recorded as well as TPM emissions in the hot undiluted flue gas. This study shows that it is challenging to get similar emission results for the same stove in different laboratories even when using the same fuel and well-defined test protocol, differences in results are due to measurement uncertainty, differences in appliance operations and not following exactly the defined Real-LIFE test protocol. Coefficients of variation for TPM, CO, OGC and NO X were 17.8 %, 20.1 %, 30.6 % and 8.7 %, respectively for stove A and 32.7 %, 13.9 %, 19.6 % and 10.0 %, respectively for stove B based on two to three repetitions per lab. The novel test protocol showed that combustion appliances may behave differently in different combustion phases, and this emphasizes the importance of measuring different combustion conditions in official testing to ensure that the appliances work properly in the field and that the measured emissions cover the whole operating range.

Computer modeling has a strong potential to replace or supplement physical fire testing of building structures. A prepared and applicable model must be simple to prepare, capable of calculating and processing results in the shortest possible time without the need for complex computing technology, thus saving time, resources, and finances and becoming a profitable alternative. The aim of this article is to design a model that would suitably represent a non-standard fire test using the Ansys software package 2025/R2. The model was created by combining the Ansys Discovery, Ansys Mechanical and Ansys Fluent software, connected using Ansys Workbench. Several calculation settings were tested, including the proposed finite mesh. The results showed the suitability of the proposed procedure and the division of calculations in separate software. The Species Transport and Viscous Shear Stress Transport (SST) k-omega model best represented fluid flow in Ansys Fluent. Ansys Mechanical confirmed the accuracy of the model in Ansys Fluent, when the predicted temperatures matched the real medium-scale fire test with an average coefficient of determination R2 = 0.93.

Friction and friction-induced noise and vibration influence comfort and efficiency in various systems. In some systems (e.g., in electric vehicles), Noise, Vibration, and Harshness (NVH) are more noticeable and critical for comfort and efficiency. Lubricants are among the limited active means to reduce NVH in such systems. However, the absence of a standard method to assess NVH in lubricated contacts hinders lubricant development. To overcome this limitation, this study developed a tribometer to evaluate both friction and NVH characteristics under boundary lubrication conditions. The design eliminates the need for a motor to reduce system-generated noise and vibration. It also omits the load cell that alters contact stiffness. The NVH Tribometer uses a pendulum-based mechanism. A non-contact rotary encoder captures amplitude decay to evaluate friction. A microphone and an accelerometer placed near the tribo-contact record friction-induced noise and vibration. Controlled temperature and flexible tribo-contact configurations enable realistic test conditions. The tribometer was validated using a base oil and a blend of base oil with zinc dialkyldithiophosphate (ZDDP) across the velocity range up to 55 mm/s. The ZDDP blend exhibited lower friction, reduced root mean square noise and vibration levels of 34.72 dB and 0.04 milli‐g , compared to 40.78 dB and 0.63 milli‐g for the base oil. Furthermore, scanning electron microscopy (SEM) confirmed tribofilm formation, and transmission electron spectrometry (TEM) revealed a film thickness of about 35 nm, explaining reduced friction and improved NVH performance.

This paper puts forward a novel multi-agent deep reinforcement learning (MADRL) control framework based on deep deterministic policy gradient (DDPG) algorithm for tracking control of 2 degree-of-freedom (DoF) helicopter system. To handle the nonlinear dynamics and uncertainties in the helicopter system, we formulate a model free data-driven approach which exploits the potentials of reward shaping technique to realise an adaptive and cooperative control policies for efficient trajectory tracking. Specifically, in this study to ensure safe and efficient RL learning without violating the hard constraints of the 2 DoF helicopter, we harness physics informed reward shaping (PIRS) technique which augments domain specific knowledge with data-driven learning. For estimating the pitch and yaw velocities of the helicopter, this study adopts a super twisting observer (STO) based on the second-order sliding mode theory. The key distinguishing features of STO is that it can ensure finite-time convergence with minimal chattering. The efficacy of the proposed MADRL augmented with STO is experimentally validated on a laboratory scale 2 DoF helicopter for several realistic test scenarios including bounded disturbances and uncertain dynamics. The experimental results highlight that the proposed framework can offer better tracking and robustness features compared to conventional state feedback control techniques.

Systematic offset of vehicle trajectory relative to road geometry and its segment law is key for road design and traffic safety. This paper use high-precision trajectory data from real vehicle test and drone aerial photography. It focus on mountain expressway and complex interweaving area in mountain city. It analyze trajectory offset in straight section, curve section, roadside structure area and overtaking condition. It reveal segment law and key influence factors of trajectory offset in different scene. Study show that vehicle usually left offset in straight section. In curve section, trajectory offset to inner side. Roadside structure and overtaking make obvious avoidance offset. Trajectory offset is strong affected by lane position and road geometry parameter. Based on vehicle size, trajectory offset and lateral safety margin, this paper give suggested width for small passenger lane. It provide theory support for road design.

ABSTRACT This study proposes a risk-based audit framework for assessing AI-generated cyber threats within intelligent systems. The framework is designed to enable auditors to systematically identify, evaluate, and test threats associated with generative artificial intelligence through a structured and operational auditing process. The methodological approach integrates multiple components, including threat identification, attack surface mapping, risk classification, risk scoring, and scenario-based testing. This integrated process facilitates the evaluation of system behavior under controlled adversarial conditions. The findings indicate that AI-generated threats—particularly phishing, prompt injection, and synthetic content manipulation—can be effectively identified through direct system testing. Systems that lack mechanisms for output validation, monitoring, and logging controls demonstrate significant exposure to critical risks. Furthermore, the likelihood–impact scoring model supports the prioritization of audit procedures, including comprehensive testing and control validation. The study demonstrates that effective auditing of AI systems relies on realistic testing environments and evidence-based evaluation methods. The proposed framework offers a practical, replicable, and scalable approach for translating AI-generated threats into auditable risks and enforceable control mechanisms within intelligent systems.

Poly-(ether sulfone) (PES) ultrafiltration mixed matrix membranes containing zinc oxide (ZnO) nanoparticles were fabricated and evaluated for drinking water treatment, with emphasis on fouling control and pollutant rejection. Unlike most studies that use model foulants, membranes were tested with organic-rich surface water from the Guarapiranga Reservoir, enabling a realistic assessment under drinking water conditions. Membranes (0-1.0 wt % ZnO) were prepared by nonsolvent-induced phase separation and characterized (SEM, AFM, porosity, hydrophilicity, zeta potential, permeability). Crossflow experiments with raw water included resistance partitioning, Hermia model analysis, and foulant extraction. Low ZnO loadings (0.25-0.50 wt %) delivered the best performance, reducing total fouling resistance by about 84% relative to pristine PES and achieving flux recovery above 96%. Improvements were linked to a more negative surface charge (-26 to -31 mV) and favorable pore structure that promoted electrostatic repulsion and reversible deposit formation. Membranes in this range also showed higher rejection of natural organic matter, with greater removal of color, TOC, and UV254 than both pristine PES and higher ZnO loadings. By contrast, the 0.75% ZnO membrane, despite its highest pure water permeability, exhibited greater irreversible fouling and lower rejection, while the 1.0% ZnO membrane behaved similarly to unmodified PES. Combining physicochemical characterization with real water tests, the study addresses practical and scale-up barriers to applying mixed matrix membranes. Findings indicate that PES-ZnO membranes with minimal nanoparticle loadings (0.25-0.50 wt %) offer a cost-effective and scalable strategy to improve flux stability, fouling control, and pollutant rejection in drinking water production.

Eurocode-based design method to evaluate the plastic bending moment capacity of demountable steel-concrete composite beams in the fire situation

Abstract With the great help of the rapid progress in computational technology, numerical simulation of structural wind loading by computational fluid dynamics (CFD) technology has become a hot research topic in recent decades. A rectangular cross-section high-rise structure with a shape ratio of 1(length):1(width):2(height) was taken as an example to elaborate on the computation of wind pressure on high-rise structures by computational fluid dynamics (CFD) technology. The computation was done, based on the time-averaged Reynolds-averaged NS Equations (RANS) turbulence-solving method along with the RNG k - ε model. The first boundary layer height on the building surface wall should meet the requirement that Y + should be between 50 and 300, which proved to be generally around 40 in this model, in line with the requirement. The simulated wind pressure coefficient results were compared with a real wind tunnel test result of a high-rise building model from the TPU Aerodynamic Database. The results reveal that the pressure coefficient finally falls after the first reduction and afterwards rises with the vertical elevation of the building while it decreases from the middle to both sides in the horizontal direction; The windward side is covered with positive pressure while other faces were covered with negative pressure, which is manifested as wind suction; CFD simulation results demonstrate consistency with the relevant test data well especially on both the windward and leeward faces, while the discrepancy becomes relatively greater for other faces and CFD technology shows great promise in predicting the surface pressure distribution of engineering structures.

The evolution of the construction sector towards industrialization is driving the adoption of lightweight, prefabricated multilayer modular and preassembled envelope systems. This offers advantages in quality, execution speed and technology integration. However, their fire performance remains a critical challenge. While the Single Burning Item (SBI) test is commonly used to assess fire reaction, it often fails to capture the actual fire behaviour of complex façades. This study introduces a novel methodology for evaluating fire performance that goes beyond standard testing requirements to meet more stringent safety criteria. The approach identifies critical areas in heterogeneous façades where fire propagation is difficult to predict. Then it was applied to a 2.2 m × 3 m façade module incorporating photovoltaic (PV), solar thermal (ST), and hybrid PVT technologies. Intermediate-scale testing was developed as a modified version of ISO 13785-1, featuring increased burner output (from 100 kW to 300 kW), altered specimen configuration (from two-wing to single-wall), and the establishment of evaluation criteria. Two experimental tests were conducted. The first failed to meet any criteria, while the second, with design improvements and passive fire protection, met all but one (burning falling parts). The findings emphasize the importance of early-stage fire performance assessment and iterative design refinement. They also reveal that a favourable fire reaction classification does not necessarily ensure acceptable fire spread behaviour, highlighting the need for more realistic testing methods. Intermediate-scale testing is proposed as a valid, cost-effective tool for fire performance evaluation. Future research should deepen correlations with full-scale tests and explore calorimetric analysis and cost-benefit assessments of fire protection strategies. • Fire safety was introduced in the design of a multilayer modular façade system • Medium-scale fire test procedure and evaluation criteria were developed • Two tests with different passive fire protection measures were performed • Implementation of protection measures significantly improved fire behaviour

We study the response of a prototype Zero-Degree Calorimeter (ZDC) detector to irradiation equivalent to 1011 1-MeV n eq/cm2, which matches the expected exposure after one year of operation at full nominal luminosity at the future Electron-Ion Collider (EIC). The prototype, which consists of 563 channels and represents about 10 percent of the final ZDC design in terms of both channel count and detector volume, was irradiated at the NASA Space Radiation Laboratory (NSRL) at Brookhaven National Laboratory (BNL) with proton beams. We demonstrate that, despite significant radiation damage to the SiPMs and non-uniform degradation across the detector volume, the detector can be successfully calibrated on a channel-by-channel basis using cosmic-ray data. The damage profile, similar to what is expected in the experiment, varies by an order of magnitude or more across the detector. Even for the most heavily damaged channels, the signal-to-noise ratio for a MIP signal remains above 5. This study provides a realistic test of the system's performance under irradiation. It complements previous SiPM-specific irradiation studies and will inform the future operation of the ZDC and other detectors that use SiPM-on-tile technology.

The paper presents an analysis of modern hardware and software solutions aimed at increasing the reliability and safety of information and measurement systems used during tests of armored vehicles on the action of a blast wave. The research focuses on the integration and use of Catman AP, DewesoftX, MATLink and I-SPEED Software Suite 2.0 software complexes for synchronous collection, processing and storage of data under conditions of high dynamic and electromagnetic action. The proposed system architecture ensures real-time data integrity and high-precision synchronization between measurement modules and high-speed video channels. The increase in the number of attacks and system hacking requires an increase in the level of protection, which can be successfully solved on the basis of cryptographic transformations. The use of such an approach will make the developed analysis and forecasting system more secure and efficient. To guarantee the confidentiality and authenticity of data, the paper proposes a cryptographic protection mechanism based on an improved method of group matrix transformation. This approach accelerates encryption and decryption processes, ensuring mathematical stability and resistance to cyber influences. The method combines group and non-group two-operand operations with modulo-two addition, which reduces computational complexity and increases the speed of cryptographic protection during real tests. The results of the study show that the integration of specialized software with advanced cryptographic methods significantly increases the efficiency of experimental data analysis, improves protection against unauthorized access and provides comprehensive monitoring of armored vehicle survivability parameters. The proposed solutions can be implemented in modern military and dual-purpose information and measurement systems, ensuring increased measurement accuracy, operational reliability and cybersecurity during field tests.

Social network content is increasingly used as an auxiliary evidence stream for financial monitoring, risk assessment, and short-horizon decision support, yet many reported gains are hard to interpret because observability, timing, and attribution are handled inconsistently across studies. This review critically synthesizes the end-to-end pipeline that transforms social posts, interaction traces, linked artifacts, and related signals into decision-facing indicators, emphasizing evidence provenance, sampling bias, conditioning (bot/spam filtering, entity linking, timestamp alignment), and the modeling blocks typically used (text, temporal, relational, and fusion components) under deployment constraints. Across sentiment, relational, and multimodal or cross-platform signals, the analysis finds that apparent improvements often depend more on alignment discipline and conservative attribution than on architectural novelty, and that performance can be inflated by attention confounds, temporal leakage, and visibility effects. Relational indicators are most defensible for monitoring coordination and propagation patterns, while multimodal gains require clear ablations and realistic missing-modality tests. To support decision readiness, the paper consolidates assurance requirements covering manipulation, degraded observability, calibration and traceability, and provides compact reporting checklists and failure-mode mitigations. Overall, the review supports bounded claims and argues for time-aware evaluation and auditable pipelines as prerequisites for operational use.

Automatic recognition of insect sound could help us understand changing biodiversity trends around the world-but insect sounds are challenging to recognize even for deep learning, due to the broad frequency ranges and limited amount of training data. We present a new dataset comprised of 26298 audio files (226.6 hours), from 459 species of Orthoptera (310 species) and Cicadidae (149 species). InsectSet459 is the first large-scale dataset of insect sound that is easily applicable for developing novel deep-learning methods. Its recordings were made with a variety of audio recorders using varying sample rates to capture the extremely broad range of frequencies that insects produce. We benchmark performance with two state-of-the-art deep learning classifiers, demonstrating good performance but also significant room for improvement in acoustic insect classification. This dataset can serve as a realistic test case for implementing insect monitoring workflows, and as a challenging basis for the development of audio representation methods that can handle highly variable frequencies and/or sample rates.

Paper reports experimental results comparing several machine learning techniques performance when distributing massively parallel computation to a set of interconnected machines. Computational resources are intentionally heterogenous to simulate real ad-hoc network environment and provide realistic setting test results. Namely Round Robin, Q-Learning and Least Loaded algorithms-based solutions are examined for their scalability, fault tolerance and workload distribution behaviors. The novelty of the paper is a set of empirical set of coefficients and bottlenecks for each implementation that is free of infrastructure specifics or error and exemption handling tools for future considerations by engineering professionals and scholars.

Electronic Design Automation (EDA) tool chain plays a critical role in the design and manufacturing of Printed Circuit Boards (PCBs). Defects in the EDA tool chain can introduce errors into Printed Circuit Board (PCB) designs and manufacturing processes, potentially leading to system failures in these applications. Therefore, verifying the reliability of the EDA tool chain is crucial. In recent years, various approaches have been proposed based on automatically generated circuits to detect defects in the EDA tool chain. However, differential testing is often bottlenecked by two practical limitations of existing circuit schematic generators: limited coverage of structurally expressive and behaviorally diverse schematics, and the lack of functionally matched variants for controlled comparison, which reduces its effectiveness in exposing subtle defects in the EDA tool chain. Addressing these challenges, we present a circuit generator (Cir-Fuzzer) for automatically generating structurally and functionally diverse circuits, as well as functionally equivalent yet structurally varied circuit variants, to detect defects in the EDA tool chain. Specifically, Cir-Fuzzer consists of three components: diversity-enhanced synthesis and optimization component constructs structurally diverse and functionally valid circuits by applying pin-level constraints and eliminating redundant paths, facilitating the generation of realistic and effective test cases for exposing tool chain weaknesses; Functionally Equivalent Circuit (FEC) variant generator introduces slight perturbations into the original schematics without compromising circuit functionality, thereby expanding the circuit transformation space and enhancing the Cir-Fuzzer’s capability to uncover latent defects in the EDA tool chain; differential testing component leverages netlists generated from original circuits and their variants to compare simulation results across simulators, versions, and circuit variants, enabling the detection of inconsistencies and the revelation of hidden defects within the EDA tool chain. Experimental results demonstrate that Cir-Fuzzer outperforms baseline approaches in detecting more tool chain defects and has uncovered 12 real defects, 4 of which were officially confirmed or fixed by vendors.

The growing interest in automation in industrial production increases the demand for artificial intelligence in the recognition of manufactured components. Sorting tasks require automated classification methods that can be implemented using machine vision and deep learning. One of the main challenges in deep learning applications is the collection of training data, especially when many different objects must be considered. Synthetic training images offer a way to avoid the time consuming acquisition of real images. In the presented method, synthetic images of industrial components are generated automatically within a computer aided design environment. A script creates images of the objects from multiple perspectives directly in the software in which they were designed. This lightweight, CAD integrated workflow explores how artificial data can be produced during the design process. A convolutional neural network is trained using transfer learning with synthetic data only and evaluated on real images to assess the sim to real gap. The method achieved an accuracy of 79.67% on the real world test set. The results show a clear internal improvement in accuracy when increasing the number of synthetic images, even with minor variations. The findings demonstrate the feasibility of the approach while indicating that the sim to real gap remains a challenge. A model trained with this workflow could support automated classification and sorting of industrial components.

Gene duplication is a fundamental driver of species adaptation and the evolution of new functions, making the reconstruction of historical duplication events crucial for understanding evolutionary processes. Whole-genome duplications (WGDs), which duplicate all gene families simultaneously, have profoundly influenced the evolution of plants, yeast, and vertebrates. Genome-scale data, such as syntenic blocks and gene family counts, are commonly employed to infer WGDs. However, detecting ancient WGDs remains challenging, as their genomic signatures are often overshadowed by extensive rearrangements and gene losses. Phylogenetic reconciliation methods between species and gene trees offer a potential means of identifying such ancient events, but frequently assume independence among gene families. This can lead to missed detections of WGDs, where gene duplications are inherently interdependent. Phylogenomics reconciliation addresses this challenge by reconciling multiple gene families at once. Unfortunately, existing models often constrain the space of possible reconciliations, overlook gene losses resulting from fractionation, or depend on conserved synteny across multiple species. This limits the number of genes that can be analyzed concurrently.In this work, we explore a phylogenomics reconciliation model that avoids synteny reliance, explicitly incorporates gene losses, and permits flexible remapping of duplications. Reconciliation under this model is NP-hard, and existing algorithms lack the scalability for large-scale datasets. To address this need, we present novel algorithmic strategies that efficiently handle tens of thousands of gene trees-a level of scalability previously unattained. We also evaluate our approach against existing methods. Experiments on both simulations and real data show that traditional LCA-mapping can yield incorrect WGD predictions after fractionation, whereas our approach is more robust. By comparing predictions using true and reconstructed gene trees, we further show that reconstruction errors greatly affect method performance and that gene tree correction is necessary for reliable results. Real data tests also reveal that our approach can recover WGDs missed by other reconciliation methods.

Currently, invasive blood-testing methods that are widely employed for diabetes monitoring present issues including discomfort, infection risks, and delayed wound healing, which lead to poor patient compliance, creating a demand for noninvasive testing methods, including saliva-based testing. However, the implementation of saliva glucose monitoring is hindered by the necessity of achieving high sensitivity, accuracy, and rapid response times. Precise atom regulation in metal-organic frameworks (MOFs) is poised to create fast, highly sensitive, and specific electronic interfaces, enabling ultrafast sensing methods, based on which a noninvasive glucose MOF derived Cu-CuO/C sensor is proposed. The sensor employs Cu-CuO/C composites, utilizing the properties of oxygen vacancies to enhance the catalytic oxidation activity of glucose, enabling high sensitivity and rapid response. Finite element simulations revealed that the synergy between the carbon substrate and Cu-CuO heterojunction optimized charge transfer and mass transport pathways, significantly improving sensor performance. Additionally, theoretical calculations demonstrated that the Cu-CuO/C electrode effectively lowers the Gibbs free energy barrier, facilitating glucose oxidation by promoting the desorption of intermediates and enhancing electrocatalytic activity. Experimental validation of the sensor performance revealed an extremely fast response time (0.3 s) and high sensitivity (919.86 μA·mM-1·cm-2), with a detection range from 0.001 to 7.33 mM (R2 = 0.997). Selectivity tests and real sample tests in artificial and human saliva further confirmed the sensor's reliability in practical applications. Finally, Cu-CuO/C was immobilized onto screen-printed electrodes, resulting in a portable biosensor that was validated for rapid glucose detection in portable instruments, paving the way for wearable real-time continuous monitoring devices.

The European Union’s regulatory mandate requirements for vehicular components include the integrity of sealing performance, mitigating leaks from fuel tanks and transmission systems in order to guard against environmental pollution. Non-compliance can result in significant costs for the OEM and their supplier base. The majority of the reported research regarding leakage from radial lip seals focuses on static analysis of leakage under a given set of laboratory conditions. However, in practice, seal conjunctions are often subjected to significant excitations due to vehicular vibration. In the current study, the case of a front-wheel drive vehicle, equipped with three-axle accelerometers and subjected to a comprehensive road test, is used as the basis for the development of a realistic representative test rig. The test rig is developed using bespoke components from the vehicle under investigation to assess the impact of the encountered natural frequencies on sealing performance in controlled laboratory conditions, when the system is subjected to controlled excitation. Experiments are conducted to evaluate leakage at the transmission interface, focusing specifically on the sealing system’s performance. The influence of driveshaft manufacturing processes using corundum grinding and subsequent surface topography upon leakage performance are also considered. Identified modal response frequencies are imposed upon the test rig using a shaker, whilst the seal leakage is measured. The importance of shaft roughness characteristics, such as topographical skewness upon seal leakage rate under various resonant conditions, are ascertained. The results indicate potentially significant leakage rates under excitation conditions, with a non-optimised shaft roughness profile.