Sheet metal blanking optimization is traditionally performed using design-of-experiments (DOE) techniques that evaluate only discrete parameter combinations, limiting the identification of optimal operating conditions within the continuous design space. This study proposes a surrogate-assisted Artificial Neural Network (ANN) and NSGA-II framework for continuous multi-objective optimization of sheet metal blanking. A unified ANN surrogateab model was developed using 100 experimental observations obtained from Taguchi L25 orthogonal array experiments on four engineering materials (SS316 steel, C110 copper, AA1100 aluminium, and C26000 brass). The model predicts burr height from sheet thickness, punch–die clearance, and material type, while analytical formulations were integrated to evaluate blanking force, shearing energy, and specific work, resulting in a four-objective optimization framework. A systematic architecture search involving 420 ANN models identified the optimal 6-12-1 network, achieving an MSE of 6.71 × 10−6, RMSE of 0.00259 mm, and R2 of 0.995. Model robustness was further verified using repeated five-fold cross-validation, sensitivity analysis, and comparative evaluation against six alternative surrogate learning models. The trained ANN was coupled with NSGA-II to generate 140 continuous Pareto-optimal solutions beyond the discrete DOE grid, enabling identification of balanced knee-point solutions through normalized Euclidean distance from the utopia point. Experimental validation of twelve representative Pareto solutions demonstrated excellent agreement between predicted and measured burr heights, with a mean absolute error of 0.0023 mm and a mean absolute percentage error of 3.31
This work explores how small additions of boron influence the microstructure and mechanical behavior of arc-welded mild steel joints. Welds produced without boron were compared with those containing 45 ppm and 47 ppm boron. Mechanical performance was evaluated using tensile, hardness, and Charpy impact testing, while microstructural features were examined through FESEM, EBSD, and XRD analyses. The results show that even trace amounts of boron can noticeably improve weld performance. The joint containing 47 ppm boron exhibited the highest strength and impact resistance, with a moderate increase in tensile strength and a significant improvement in absorbed energy compared to the boron-free weld. These improvements are closely linked to the formation of finer microstructures, particularly acicular ferrite, along with an increase in high-angle grain boundaries. Fracture surface analysis revealed a clear shift from brittle behavior in the boron-free weld to a more ductile failure mode in boron-modified joints. Finite element simulations further supported the experimental findings by indicating a more uniform stress distribution in the presence of boron. Although the difference between 45 and 47 ppm is small, the study provides useful insight into the role of boron near its upper practical limit in welding applications, while also acknowledging the challenges associated with controlling such low concentrations.
Mobile robots deployed in real-world environments are exposed to variable terrain conditions, payload changes, and repeated acceleration and deceleration cycles that induce fluctuating mechanical loads. These operating conditions lead to cumulative mechanical fatigue, which is a major cause of actuator and structural component failure in robotic systems. Conventional robot controllers primarily focus on trajectory tracking and stability and do not explicitly account for mechanical health or fatigue accumulation. This paper proposes a predictive fatigue-aware adaptive control framework for mobile robots that integrates low-cost load estimation, real-time fatigue accumulation modeling, lightweight prediction, and supervisory adaptive motion control. The proposed approach estimates mechanical load using motor current, inertial measurements, and encoder data, and computes a fatigue index that serves as a proxy indicator of cumulative operational mechanical stress (not a direct measurement of material fatigue damage). A prediction module anticipates future load peaks and enables proactive adaptation of motion constraints, such as velocity and acceleration limits. The framework is validated through extensive simulation and experimental implementation on a differential-drive mobile robot platform. Results demonstrate a significant reduction in peak motor current and vibration levels, indicating reduced mechanical stress. This mechanical protection is achieved at the cost of increased task completion time—rising from approximately 45–55 s under rough terrain and from 60 to 74 s under high-payload conditions—and a modest increase in trajectory tracking RMSE of 6–13 mm across all tested conditions. This explicit trade-off between mechanical protection and task performance is a direct consequence of the imposed velocity and acceleration limits and is discussed in full in the results section. Experimental validation under rough-terrain and high-payload conditions demonstrated that the proposed predictive fatigue-aware adaptive controller reduced peak motor current by approximately 35
Friction stir welding (FSW) subjects the welding tool to severe thermo-mechanical loading, resulting in complex fatigue conditions that limit tool life. Current tool design approaches remain largely empirical, leading to considerable uncertainty in lifetime prediction under realistic process conditions. This study presents a fatigue-based methodology for predicting FSW tool damage by combining experimentally measured process loads with linear damage accumulation. Process forces and torque were acquired during welding of AA 6060 T66 and AA 5754 H111 and converted into probe stresses. Fatigue data for H13 tool steel were experimentally determined at 530 °C for rotating bending, torsion and axial tension/compression. Owing to the pronounced scatter of the axial data, the experimental bending and torsional S–N curves were implemented in the lifetime assessment, whereas a hybrid axial S–N curve combining the experimentally determined median endurance limit with the FKM-based finite-life slope was used for axial loading. The workflow was evaluated for precipitation-hardenable and strain-hardened aluminum alloys. For AA 6060, its application reduced the overestimation of tool life to a factor of 2.1, while AA 5754 still exhibited an overestimation factor of 5.1. The remaining deviations highlight the limitations of linear damage accumulation under non-stationary welding conditions. Overall, the results demonstrate that experimentally supported fatigue data improve the prediction accuracy of FSW tool life and provide a first-order basis for predictive tool assessment.
The increasing demand for sustainable and cost-effective automotive spare-part production has encouraged the adoption of near-net-shape manufacturing techniques. This study presents the CAD modelling, finite element analysis, floor-mould casting, and electrochemical evaluation of an internal combustion engine rocker arm produced from Aluminium Alloy 6061. The rocker arm geometry was developed from OEM-derived field measurements and analysed under a representative static valve load of 500 N. The FEA results showed a maximum von Mises stress of 96.3 MPa at the neck region and a minimum factor of safety of 2.87, indicating that the component remained within the elastic design limit under the assumed loading condition. The component was fabricated by sand casting with minimal machining, showing the feasibility of near-net-shape production. Electrochemical testing in 3.5 wt
This study introduces Multi-Objective Chaos Game Optimization (MOCGO), a metaheuristic algorithm developed for the purpose of handling multi-objective structural optimization problems with constraints. In doing so, MOCGO utilizes a single-objective algorithm known as Chaos Game Optimization (CGO) in conjunction with chaos exploration dynamics, a diversity-preserving archive, and a grid-selection process to create an effective set of Pareto optimal solutions. This algorithm has been tested on six different trusses (10-bar to 120-bar) based on the problem of minimizing the weight and compliance of a structure while satisfying the stress constraint criteria. As measures of fitness, MOCGO performs significantly better than other algorithms, including NSGA-II, MOGOA, MOALO, and MOAVOA. These results demonstrate MOCGO as an efficient and robust approach for constrained multi-objective structural design.
Magnesium slag (MS) is a significant waste product of magnesium production. The sustainable reuse of MS presents a highly promising approach to the development of eco-friendly construction materials. The use of supplementary materials to reduce cement consumption significantly decreases the amount of CO2 produced by ordinary Portland cement concrete (OPCC). The present study systematically analyzes and summarizes the current state of research on the mechanical, durability and micro-structural properties of MS as a building material based on the available literature. The effects of the utilization method, dosage, and curing conditions of MS on the mechanical and durability properties were discussed. Research findings on the strength and durability properties of MS as a building material remain relatively limited, and the conclusions reported in the literature show a lack of full consistency. Overall, MS can effectively serve as a supplementary binder, where an optimum substitution level of approximately 30
Zirconium and its alloys are widely used in nuclear applications due to their low neutron absorption, corrosion resistance, and favorable mechanical properties. However, hydrogen uptake and irradiation-induced defects (vacancies and self-interstitial atoms (SIAs)) can significantly degrade performance. In this work, molecular dynamics simulations are employed to investigate the coupled effects of hydrogen and point defects on the thermodynamic, elastic, and thermal transport properties of zirconium within the low-concentration regime (0–2
Biomass briquettes improve handling, storage, and energy density, yet sugar cane residue briquettes suffer from high mineral content that drives ash deposition, slagging, corrosion, and equipment wear. This study evaluates water-leaching pretreatment followed by briquetting with cassava starch binder to reduce problematic inorganics and enhance the physical and combustion properties of sugarcane residue briquettes. Raw sugar cane residue was collected and size-reduced to 1, 5, and 10 mm, then leached in ultrapure water at 25, 50, and 100 °C for 20, 40, and 60 min using a 3 × 3 × 3 full factorial design (27 runs; triplicates). Leached solids were blended with cassava starch binder (20:3, w/w) and densified on a hydraulic press at 18 MPa (Ø 50 mm × 50 mm). Physical and combustion properties were measured according to ASTM/ASABE protocols. Elemental analyses were also performed. ANOVA and regression were used to assess effects of particle size, leach temperature, and time. Leaching and densification measurably upgraded fuel quality: HHV increased from 16.85 to 17.99 MJ kg−1 (+ 6.94
Stress concentration at geometric discontinuities significantly influences the structural integrity and fatigue life of mechanical components. Stepped circular bars with shoulder fillets are commonly used in engineering systems where load transfer occurs between sections of different diameters, making accurate prediction of stress concentration essential. This study presents an automated MATLAB–ABAQUS framework for high-throughput parametric evaluation of stress concentration factors in stepped circular bars subjected to axial loading. MATLAB was used to generate design-of-experiments parameters and automatically create Python scripts for ABAQUS to perform finite element simulations and extract stress results. A 24 full factorial design of experiments was conducted to investigate the effects of section length, smaller diameter, fillet radius, and larger diameter on the stress concentration factor. Statistical analysis using ANOVA showed that the smaller diameter and fillet radius are the most influential parameters. A regression-based predictive model and a machine learning surrogate model were developed to estimate stress concentration factors directly from geometric parameters, achieving a coefficient of determination R2 of approximately 0.91. Validation with classical Peterson stress concentration charts showed agreement within 5–8
Thermal management of high-power microelectronics requires coupled design of coolant, channel geometry and operating point, yet mesh-based conjugate CFD is too costly for broad parametric search. This work develops a mesh-free surrogate for three-dimensional steady laminar conjugate heat transfer in microchannel heat sinks using a single physics-informed neural network (PINN); the contribution is primarily methodological. One fully connected network represents the field map (x, y, z, φ, Re) → (u, v, w, p, Tf, Ts). Temperature continuity at the solid–fluid interface is enforced by reading the two temperature outputs as one continuous field, and heat-flux continuity through a dedicated interface loss with an inverse-residual adaptive weight. The framework is demonstrated on a Cu-Al₂O₃/water hybrid nanofluid in a silicon sink (ks/knf ≈ 220). Against the finite-volume benchmark of Qu and Mudawar, the network reproduces velocity to within a pointwise L₂ error of 1.2
Due to the constant reduction in the size of interconnects (based on Al, Cu, etc.) in microelectronic structures, the associated thermal load is constantly increasing. Therefore, it is important to study the heating dynamics of interconnect in pulsed modes. In this work, the temperature modes of operation of an Al-Si binary structure were simulated using the finite element method. The results of numerical modeling of thermal behavior of Al-Si structures were compared with experimental data obtained at various pulse energies (Wτ = 0.5–5.5 J) and duration up to 500 µs. Comparison showed that discrepancy between experimental results and numerical modeling results exceeded 10
Kinked geometries are common in lightweight aerospace structures, deployable robotic architectures, and composite assemblies, yet they are typically treated as geometric discontinuities that reduce stiffness or perturb vibration characteristics. Their influence on modal strain-energy distribution and vibration response has received comparatively little attention. This work systematically investigates the free-vibration behaviour of cantilevered kinked beams fabricated from monolithic aluminum and particle-reinforced composites (Al–TiC and Al–SiC) using Timoshenko beam theory. A finite-element formulation incorporating reduced-order viscoelastic damping is developed to examine the influence of kink angle (0°–180°) and normalized kink position on spectral characteristics and strain-energy distribution. The results reveal pronounced non-monotonic frequency evolution, strong mode veering, and substantial redistribution of modal strain energy. As the kink angle increases into the obtuse regime, the kink vertex transitions from a passive geometric junction into an active energy-localization region, producing significant concentration of modal strain energy and higher predicted effective damping. Mesh-convergence studies and validation against the analytical straight-beam solution demonstrate excellent agreement, with a fundamental frequency error of only 0.0077
Titanium metal matrix composites (TiMMCs) offer exceptional specific strength and thermal stability; however, their machinability is severely hindered by abrasive ceramic reinforcements. The aim of this study synthesizes Ti6Al4V-based TiMMCs reinforced with hybrid B4C, ZrO2, SiC, and MoS2 particles via powder metallurgy and optimizes multi-stage machinability through sequential chemical etching, grinding, and polishing, evaluating effects on surface roughness (Ra) and material removal rate (MRR) via Taguchi L9 orthogonal arrays.The result shown as a morphological analysis revealed homogeneous reinforcement distribution, while XRD confirmed in-situ TiC and TiB2 formation via Ti-B4C reactions, enhancing mechanical properties. The optimal B3 composition (7.5 wt.
The purpose of this investigation is to assess the outcome of Soret and Dufour effects on viscoelastic hybrid nanofluid flow across a sheet with convective conditions. The Levenberg–Marquardt technique is notable for its novel approach and convergent stability in the field of artificial neural networks. Using regression plots, state transition measures, histogram representations, and mean squared errors, this proposed model generates a numerical approach. The thermal–solutal convective flow of viscoelastic hybrid nanofluid based on AA7072–AA7075-ethylene glycol–water that is appropriate for complex industrial heat transfer systems where simultaneous mass and heat transport is essential. Heat exchanger design and optimization, cooling systems for metallurgical and chemical processing facilities, polymer production, and energy systems needing improved thermal performance under intricate flow circumstances are all areas in which it is especially helpful. The model helps enhance thermal efficiency, regulate concentration gradients, and guarantee stable operation in high-performance industrial applications by taking into consideration Soret-Dufour effects in addition to viscoelastic behavior. This study investigates mass and heat transmission enhancement in a laminar, steady, and incompressible flow of AA7072–AA7075/EG–H₂O Boger hybrid nanofluid across a sheet. Dufour–Soret effects, convective boundary conditions, thermal radiation, magnetic fields, and Darcy–Forchheimer porous resistance all affect the flow.
Brake squeal remains a critical noise–vibration–harshness (NVH) problem in automotive disc brakes because it arises from friction-induced mode coupling and is highly sensitive to contact, boundary, and operating conditions. This study aims to develop and validate a finite-element-based complex eigenvalue analysis (FE–CEA) workflow for predicting high-frequency squeal and identifying an effective structural countermeasure using a centre-slotted brake pad. A three-dimensional finite element model of a hydraulic disc brake assembly was developed and validated using experimental modal analysis (EMA). The validated model was then used to perform complex eigenvalue analysis after static preloading and contact linearisation. Parametric studies were carried out to examine the effects of friction coefficient and brake pressure, followed by numerical screening of ten rotor and pad geometric configurations. The most effective design was experimentally validated using dynamometer tests under controlled operating conditions. The FE model showed good agreement with EMA results, with rotor mode-shape correlation values of MAC ≥ 0.85 and an assembled-system frequency mismatch below approximately 3.5% . The parametric study identified a critical friction coefficient near μ≈ 0.27 for instability onset, while increasing brake pressure mainly amplified the instability growth factor. Among the ten configurations, the 10 mm centre-slotted pad suppressed the dominant unstable mode by weakening rotor–pad mode coupling. Experimental validation at 300 rpm, 4 bar, and 45 ^∘C to 55 ^∘C confirmed that the centre-slotted pad eliminated the 6.31 kHz squeal tone and reduced the peak sound pressure level by more than 20 dB(A) over five repeated tests. The results demonstrate that FE–CEA, when supported by modal and acoustic validation, can be used as an effective screening tool for brake squeal mitigation. The proposed centre-slotted pad provides a practical structural modification for suppressing high-frequency squeal while reducing reliance on trial-and-error prototyping.
This research proposes a study on AI-assisted optimized Job-Shop Production System (JPS) to improve performance, reduce cost and time, and detect machine failure. In the current industrial landscape, optimizing Job Shop Production Systems (JPS) is critical for achieving higher efficiency, reduced costs, and improved resource utilization. This study presents a comprehensive review of both conventional and artificial intelligence (AI)/machine learning (ML)-based approaches applied to JPS, with particular emphasis on job sequencing and machine failure prediction. The system is modeled as a multi-stage job-shop with diverse tasks and constraints, highlighting challenges in sequencing, resource allocation, and machine reliability. This work addresses job-shop scheduling in a T-shirt manufacturing system using Grey Wolf Optimization (GWO). While conventional GWO provides feasible job sequencing, it suffers from premature convergence. An Improved Grey Wolf Optimization (IGWO) algorithm is therefore proposed to enhance exploration and convergence efficiency. Simulation results show that IGWO achieves better job sequencing with reduced make-span, lower production cost, and improved resource utilization compared to standard GWO.
Intrusion detection is essential in contemporary manufacturing systems. These are vulnerable to various cyber threats due to the integration of cyber-physical systems and continuous data exchange. Traditional intrusion detection systems include statistical models and standard machine learning (ML) approaches. They struggle with high-dimensional sensor data, imbalanced datasets, and fast-changing attack patterns. To overcome these challenges, we propose a hybrid intrusion detection model. It combines Dynamic Feature Convolution (DFC) with a Transformer-based temporal modelling structure. The DFC component uses gated convolutional layers with sigmoid and tanh activations to learn localized temporal features. The Transformer component applies self-attention mechanisms to capture long-term dependencies. This hybrid model learns both local feature dynamics and global temporal dependencies in industrial time series data. We evaluated the proposed model using the publicly available Water Distribution (WADI) dataset. This dataset simulates realistic industrial processes under both normal and attack scenarios. Experimental results demonstrate robust detection performance. Over five independent training runs and 5-fold cross-validation, the model achieved an average accuracy of 97.34
This research focuses on how basalt macrofibers and microfibres affect the durability and Mechanical properties of M60 high-strength concrete. Compressive, split tensile, flexural strength, and acid resistance tests were employed to assess the performance of fiber-reinforced concrete at various fibre volume fractions. The findings demonstrate that the addition basalt fibres greatly enhanced the overall behaviour of concrete compared with the conventional control mix. The greatest compressive strengths were attained at a fibre dosage of 0.75
This paper reviews the latest advances in graphene-reinforced carbon fiber composites (GRCFCs), focusing on their mechanical, thermal, electrical and interfacial properties. Various graphene preparation methods are discussed, including chemical vapour deposition, mechanical and chemical exfoliation, reduction of graphene oxide (GO), with critical analysis of how each method influences graphene quality and subsequent composite performance. GRCFC fabrication techniques including chemical grafting, electrophoretic deposition and wet spinning, are then discussed for their impact on composite properties. Furthermore, the role of matrix selection, dispersion strategies, and associated challenges are explored. This review highlights the enhancement in tensile strength, Young’s modulus, and overall multifunctional performance arising from improved interfacial bonding and microstructural design, particularly with the incorporation of polyacrylonitrile (PAN)-based systems. A comparative analysis of structure-property relationships is presented to elucidate the synergistic effects of graphene within carbon fiber matrices. In addition, key applications of GRCFCs in aerospace, automotive and energy sectors are outlined, emphasizing their role in mass-efficient structural components, fuel efficiency, and energy storage systems. The potential of flash Joule heating as an emerging recycling strategy is also discussed, demonstrating its capability to convert waste composites into turbostratic flash graphene with catalytic applications. At the end, this paper provides an in-depth analysis of the current challenges, including dispersion uniformity, interfacial compatibility and manufacturing complexity along with future perspectives for the scalable development and industrial adoption of GRCFCs. Various GRCFC preparation technologies and their properties are discussed. Interfacial mechanisms for enhancing mechanical, thermal, and electrical properties are reviewed. Recent advances in GRCFC applications in aerospace, automotive, and energy sectors are highlighted. The strategic prospects for sustainable recycling using flash Joule heating are explored. The challenges and future development directions of GRCFCs are discussed.
