Browse Items (16481 total)
Sort by:
-
AI-enhanced approaches for personalized cardiac treatment: insights from ECG data
The analysis of drug-induced alterations in the electrocardiogram (ECG) is essential in measuring cardiac safety, but manual analysis is not always accurate enough to identify subtle but important effects. This paper examines how machine learning (ML) models can be used to categorize various pharmacological treatments according to their distinct ECG patterns to establish a platform of individualized therapeutic evaluation. Using the public ECG Effects of Dofetilide, Moxifloxacin, Dofetilide+Mexiletine, Dofetilide+Lidocaine and Moxifloxacin+Diltiazem (ECGDMMLD) database, key electrophysiological features were extractedincluding heart rate variability (HRV) and standard cardiac intervals (RR, PR, QT, QRS) to train and compare three different classifiers: XGBoost, Random Forest, and a Support Vector Machine (SVM). The analysis showed that tree-based ensemble techniques were very useful in this task. The XGBoost model had a better classification accuracy of 98.1%, which was closely followed by the random forest at 97.3%. Conversely, the SVM had much lower accuracy, implying that it was not as well adapted to the complexity of the high-dimensional ECG data. These results establish that ML models, particularly XGBoost, can accurately decode complex drug-induced cardiac signatures from ECG data. This work is a powerful demonstration of the proof-of-concept of automated and data-driven analytics integration into clinical processes to enhance drug safety and promote personalized medicine. The Author(s) 2026. -
Design and Implementation of universal converter using ANN controller
This paper details the hardware implementation of a Universal Converter controlled by an Artificial Neural Network (ANN), utilizing key components such as six Insulated Gate Bipolar Transistors (IGBTs), two inductors, and two capacitors for energy storage and voltage smoothing. A Digital Signal Processor (DSP) serves as the core controller, processing real-time input and feedback signals, including voltage and current measurements, to dynamically manage five operational modes: rectifier buck, inverter boost, DC-DC buck, DC-DC boost, and AC voltage control. The pre-trained ANN algorithm generates pulse-width modulation (PWM) signals to control the switching of the IGBTs, optimizing timing and duty cycles for efficient operation. The system effectively accommodates both AC and DC inputs, ensuring stable outputs with minimal ripple by dynamically selecting the appropriate mode based on load requirements. Experimental results demonstrated that the ANN controller maintained total harmonic distortion (THD) below 5% in rectifier and inverter modes while achieving an overall efficiency of 9496% in DC-DC modes. The controllers capability to adapt to real-time feedback significantly improved power conversion quality and reduced switching losses. This study confirms the efficacy of the ANN-controlled Universal Converter in meeting the demands of modern power systems through versatile and adaptive control. The Author(s) 2025. -
An enhanced performance analysis of load based resource sharing framework for MIMO systems in 5G communication systems
Resource sharing serves as a cost-effective and dynamically adjustable method for alleviating traffic congestion in wireless networks. Advancements in multi-input multi-output (MIMO) technologies for 5G communication systems have led to the exploration of resource sharing across various cells or sectors. This approach aims to optimise network performance, focussing on coverage, capacity, and quality of service. This document presents a new load-based resource-sharing framework designed for multi-cell MIMO systems. The proposed framework utilises channel-loading data from local base stations and dynamically allocates available resources among adjacent base stations. The proposed framework facilitates dynamic resource sharing, effectively addressing traffic overload in 5G networks. The proposed LBRS achieved a delta-P value of 90.91%, a prevalence threshold value of 89.84%, a critical success index value of 91.01%, and a Mathews correlation coefficient value of 91.27% at the terminal access. At the resource transmission, the system recorded a delta-P value of 92.10%, a prevalence threshold value of 92.18%, a critical success index value of 91.65%, and a Mathews correlation coefficient value of 88.31%. The simulation results indicate that the proposed framework effectively enhances dynamic resource sharing, resulting in a notable improvement in network performance. The Author(s) 2025. -
Improving EEG based brain computer interface emotion detection with EKO ALSTM model
Decoding signals from the CNS brain activity is done by a computer-based communication device called a BCI. In contrast, the system is considered compelling communication equipment enabling command, communication, and action without using neuromuscular or muscle channels. Various techniques for automatic emotion identification based on body language, speech, or facial expressions are nowadays in use. However, the monitoring of exterior emotions, which are easily manipulated, limits the applicability of these procedures. EEG-based emotion detection research might yield significant benefits for enhancing BCI application performance and user experience. To overcome these issues, this study proposed a novel EKO-ALSTM for emotion detection in EEG-based braincomputer interfaces. The proposed study comprises EEG-based signals that record the electrical activity of the brain connected to various emotional states, which are gathered as real-time acquired EEG signals for emotion detection. The data was pre-processed using a bandpass filter to remove unwanted frequency noise for the obtained data. Then, feature extraction is performed using DWT from pre-processed data. Specifically, the proposed approach is implemented using Python software. The proposed system and existing algorithms are compared using a variety of evaluation criteria, including specificity, F1 score, accuracy, recall or sensitivity, and positive predictive values or precision. The results demonstrated that the proposed method achieved better performance in EEG-based BCI emotion detection with an accuracy of 97.93%, a positive predictive value of 96.24%, a sensitivity of 97.81%, and a specificity of 97.75%. This study emphasizes that innovative approaches have significantly increased the accuracy of emotion identification when applied to EEG-based emotion recognition systems. Additionally, the findings suggest that integrating advanced machine learning techniques can further enhance the effectiveness and reliability of these systems in real-world applications, paving the way for more responsive and intuitive BCI technologies. The Author(s) 2025. -
Changing climate and its impacts on the dynamics of future malaria transmission over certain endemic regions in India
As climate change plays a major role in evaluating the malaria disease over India, it is highly relevant to assess the spatio-temporal variability of malaria transmission dynamics over different climatic zones in India using modelling studies. In this study, VECTRI (vector-borne disease community model of the International Centre for Theoretical Physics, Trieste) model is simulated to predict the future malaria transmission dynamics over four major climatological zones of India, forced with the different climatic parameters such as temperature and rainfall and non-climatic parameter such as population density. The climate data is obtained from multi model mean of different CMIP6 global climate models under the SSP5-8.5 scenario. Results indicate that there is an overall decrease in EIR (Entomological Inoculation Rate) values of 10 to 30% are seen over most of the Indian regions with an increase in temperature about 4 to 5C and rainfall about 10 to 40%, by end of the century (2080s) when compared with the baseline period (19852014). However, few exceptions are seen over few parts of western and peninsular region where increase in EIR values are seen. This decrease (increase) in EIR values which describes the intensity of malaria transmission is predominantly controlled by temperature and rainfall during summer (winter) monsoon seasons. Such results from the VECTRI model may be useful for policymakers towards various malaria disease control programs in India and this may provide a basis for climate change impact assessments on malaria risk at a regional scale. The Author(s) 2025. -
Optimized placement and sizing of solar photovoltaic distributed generation using jellyfish search algorithm for enhanced power system performance
The strategic integration of distributed generation (DG) units into distribution power networks (DPNs) is pivotal for augmenting system efficiency and stability. This study introduces an advanced metaheuristic optimization framework leveraging the Jellyfish Search Algorithm (JSA) for the optimal placement and sizing of solar photovoltaic (PV) DG units. The formulated multi-objective function incorporates real power loss (RPL) minimization, voltage deviation index (VDI) reduction, and voltage stability index (VSI) enhancement, employing a weighted sum approach (WSA) to ensure computational rigor. The efficacy of the proposed methodology is rigorously validated on the IEEE 33-bus radial DPN under single and multiple PV system deployment scenarios. For single PV system optimized inclusion, RPL of the DPN is cut down from 210.98kW to 102.89kW, total VDI is reduced from 1.8047 p.u to 0.5331 p.u, and minimum VSI is increased from 0.6671 to 0.7559. For two PV DG units inclusion, RPL is reduced to 82.99kW, total VDI is reduced to 0.6518 p.u with a least VSI improved to 0.8848. However, better result is obtained with three units of DG placement with RPL reduced to 69.59kW, total VDI decreased to 0.3293 p.u with a least VSI of the test system increased to 0.8916. Comparative analyses against state-of-the-art metaheuristic algorithms underscore the superior convergence efficiency and optimality of JSA in addressing nonlinearity and high-dimensionality constraints. Empirical results substantiate substantial RPL reduction, bus voltage enhancement, and system stability reinforcement, establishing JSA as an avant-garde paradigm in DG optimization. This is a U.S. Government work and not under copyright protection in the US; foreign copyright protection may apply 2025. -
An efficient intelligent transportation system for traffic flow prediction using meta-temporal hyperbolic quantum graph neural networks
Intelligent Transportation Systems (ITS) necessitate scalable, real-time, and adaptive traffic flow prediction models to enhance urban mobility and alleviate congestion. Conventional Graph Neural Network methodologies encounter difficulties in managing extensive road networks, long-range temporal relationships, and computing efficiency for real-time applications. An innovative deep learning framework named Meta Temporal Hyperbolic Quantum Graph Neural Networks that integrates hyperbolic embeddings, meta learning, quantum graph, Neural Ordinary Differential Equation (NODEs) to improve the ITS Performance. Across many cities, meta learning facilitates swift adaptation with minimum retraining whereas hyperbolic graph embeddings efficiently depict hierarchical route configurations The usage of Quantum Graph Neural Networks (QGNNs) enhances graph-based scheming, enabling real-time traffic flow to forecast for extensive networks. Also, NODEs summarize ongoing traffic progress, enhancing precision under dynamic sceneries. Datasets like Los-loop and SZ-taxi datasets are validated by experiments which highlights the impact of the proposed MTH-QGNN model, acquiringamean value RMSE of 4.5 and MAE of 3.5, ensuring minimal prediction error. MTH-QGNN model constantly sustained accuracy above 80% and R2 values exceeding 83%, representing robust predictive trustworthiness. MTH-QGNN effectively captures complex spatiotemporal traffic patterns with a variance score above threshold value. The Author(s) 2025. -
One pot synthesis of a novel bioactive compound employing a deep eutectic solvent grafted MWCNT system in a solventless environment
Heterogeneous catalysis is considered as a suitable alternative to conventional organic synthesis for the selective production of industrially significant fine chemicals. The development of supported catalysts by dispersing minimal quantities of active component can reduce production costs and enhance energy efficiency. The current work reports the development of Deep eutectic solvent (DES) modified multiwalled carbon nanotube (MWCNT) system and its activity in the Knoevenagel condensation reaction. The catalytic system was developed by grinding a very low concentration (0.83mM) of DES with desired amount of MWCNT. Various interactions of the three component DES with MWCNT were analysed by X-ray photoelectron spectroscopy (XPS). The reaction favoured a novel compound selectively with yield around 92% in solvent free medium. Anti-cancerous studies of the synthesized compound demonstrated a strong IC50 value of 15.62g/ml and a statistically calculated IC50 value of 9.8g/ml. Acridine orange/ Ethidium Bromide (AO/EB) dual fluorescence staining studies revealed that the test ligand with lowest concentration of 7.8g/ml was capable to induce apoptosis in 100% of MCF-7 cells. It is evident from the studies that the synthesized compound is a strong anticancer agent with potential to be investigated further. The Author(s) 2025. -
Emerging applications of sustainable modified CdO/Ag-CdO NPs for electrochemical sensitive and selective detection of mercury (Hg+) heavy metal
The sensitivity of developed electrode has gained significant attention for potential energy storage and electrochemical sensor activities. The modified nano-CdO/Ag-CdO-carbon paste electrodes were developed for electrochemical detection of mercury (Hg+) heavy metal. The synthesized samples were well characterized through PXRD (Powder X-ray diffraction), SEM-EDAX (Scanning Electron Microscopy-Energy Dispersive X-ray Analysis), XPS (X-ray photo-electron spectroscopy), FT-IR (Fourier transform Infra-Red), and UV-Visible spectroscopy. The Ag-CdO modified electrode endowed with higher sensing current and Csp (188F/g) than pure CdO NPs (94.6F/g) measured by Linear Sweep (LS), Cyclic Voltammetric (CV) and Electrochemical Impedance Spectral (EIS) techniques. The excellent electrochemical sensing action of developed Ag-CdO electrode was examined on heavy metal Hg+ ions at 15 mM scan rate in 0.1M KCl. The linear relationship of sensing measurements with smaller concentration (15 mM) was observed with its increased current (+ 1.64 104 A/cm2 at 1 mM) at 30 mV/s. LOD of CdO and Ag-CdO electrodes (Hg+Oxid) were found at 1.91 mM & 2.41 mM (Hg+Red) respectively with maximum sensitivity at -0.006V. LOQ of CdO and Ag-CdO electrodes (Hg+Oxid) were 5.78 mM & 6.98 mM respectively with maximum sensitivity at -0.006V. The antibacterial measurements of prepared samples were examined for their susceptibility to inhibit the growth of gram-negative (Escherichia coli) and gram-positive (Staphylococcus aureus) bacteria. Thus, the synthesized Ag-CdO electrode provides a new insight for determining the concentrations of critical pollutants and processing the various nanoparticles for sensing of cyanogenic heavy metals. The Author(s) 2025. -
Design of a novel shunt active harmonic compensator with AUV-PQ-SRF reference current extraction, OSV-MPC and SMC techniques
Harmonic distortion makes it difficult to maintain good Electrical Power Quality (EPQ) in distribution networks with many nonlinear loads. Three significant advances are combined in this papers innovative Shunt Active Harmonic Compensator (SAHC) design: (i) a new technique for extracting reference currents, called AUV-PQ-SRF, which combines the Unit Vector, PQ, and SRF techniques in a unique way to improve harmonic detection; (ii) an OSV-MPC strategy that improves reference current tracking accuracy by doing away with traditional pulse width modulation; and (iii) a Sliding Mode Controller (SMC) for dynamic and reliable DC link voltage regulation under a range of load conditions. The accuracy, robustness, and response time issues with traditional methods are addressed by the suggested approach. Results from simulations conducted in accordance with IEEE-519-2022 standards show a considerable decrease in total harmonic distortion (THD), along with increased power factor and real and reactive power compensation. This study provides a thorough and useful solution for dynamic power quality issues, setting a new standard in active filtering. The Author(s) 2025. -
An enhancement of machine learning model performance in disease prediction with synthetic data generation
The challenges of handling imbalanced datasets in machine learning significantly affect the model performance and predictive accuracy. Classifiers tend to favor the majority class, leading to biased training and poor generalization of minority classes. Initially, the model incorrectly treats the target variable as an independent feature during data generation, resulting in suboptimal outcomes. To address this limitation, the model was adjusted to more effectively manage target variable generation and mitigate the issue. This study employed advanced techniques for synthetic data generation, such as synthetic minority oversampling (SMOTE) and Adaptive Synthetic Sampling (ADASYN), to enhance the representation of minority classes by generating synthetic samples. In addition, data augmentation strategies using Deep Conditional Tabular Generative Adversarial Networks (Deep-CTGANs) integrated with ResNet have been utilized to improve model robustness and overall generalizability. For classification, TabNet, a model tailored specifically for tabular data, proved highly effective with its sequential attention mechanism that dynamically processes features, making it well suited for handling complex and imbalanced datasets. Model performance was evaluated using a novel approach of training synthetic data and testing on real data (TSTR). The framework was validated on the COVID-19, Kidney, and Dengue datasets, achieving impressive testing accuracies of 99.2%, 99.4%, and 99.5%, respectively. Furthermore, similarity scores of 84.25%, 87.35%, and 86.73% between the real and synthetic data for the COVID-19, Kidney, and Dengue datasets, respectively, confirmed the reliability of the synthetic data. TabNet consistently showed substantial improvements in F1-scores compared to other models, such as Random Forest, XGBoost, and KNN, emphasizing the importance of selecting the right synthetic data augmentation techniques and classifiers. Additionally, SHapley Additive exPlanations (SHAP)-based explainable AI tools were used to interpret model performance, providing insights into feature importance and its impact on predictions. These findings confirm that the proposed approach enhances the accuracy, robustness, and interpretability, offering a valuable solution for addressing data imbalance in classification tasks. The Author(s) 2025. -
Linear and weakly nonlinear stability analyses of Darcy Bard convection with feedback control
In this paper, the effect of feedback control on the criterion for the onset of DarcyBard convection in a horizontal Boussinesq Newtonian fluid is studied theoretically. The bounding isothermal lower and upper surfaces are considered to be rigid. The single term Galerkin method and the Maclaurin series expansion are combined with the Newton-Raphson method of three variables to perform a linear stability analysis in order to determine eigen value. To make a weakly nonlinear stability analysis of the system, a Vadasz Lorenz model is constructed. The models various properties are discovered to be identical to those of the standard Lorenz model. The exhibits both dissipative and conservative characteristics and the bounded nature of its solution is demonstrated by the trapping region, which takes the form of an ellipsoid. The Hopf-Rayleigh number determined from the autonomous dynamical system predicts the onset of chaos. The influence of the controller gain parameter and the Biot number on the onset of convection has been analyzed. Results from the study reveal that the controller gain parameter stabilizes the system and further delays the onset of chaos. Overall, the study establishes that an increase in the Biot number promotes long-term periodic motion over chaotic behavior, while an increase in the controller gain parameter enlarges the trapping region, thereby contributing to improved system stability. The Author(s) 2025. -
Simultaneous photovoltaic distributed generation and capacitor optimization for enhancing performance indices of radial power distribution system
This paper presents an effective metaheuristic framework using the Osprey Optimization Algorithm (OOA) for the simultaneous allocation of distributed generation (DG) units and capacitor banks (CB) in radial distribution systems (RDS). The method optimizes the locations and sizing for DG units and CB to minimize active power losses (APL), to reduce voltage deviation (VD), and to enhance voltage stability. The performance of the proposed approach is tested on IEEE 69-bus and 118-bus benchmark RDSs and the real-time Tala Egyptian RDS. The OOA achieved superior results compared to popular heuristic algorithms such as antlion optimizer (ALO), hunter-prey optimizer (HPO), and whale optimizer algorithm (WOA). Specifically, for three units of DG and single capacitor integration in the 69-bus system, OOA reduced the total APL by 75.1%, lowered the total voltage deviation (TVD) by 1.4835p.u., and improved the total voltage stability index (TVSI) by 3.0229. With optimal assimilation of three units of DG and capacitors each, APL reduction, TVD minimization, and TVSI improvement further extended to 79.9%, 1.5013p.u., and 2.2787, respectively. Furthermore, OOA validation on a variable-load 69-bus RDS and the real 37-bus Tala Egyptian RDS demonstrated consistent and superior performance, showcasing its robustness. Statistical analyses also confirm OOAs efficiency and ability to solve DG planning in the distribution networks. The Author(s) 2025. -
f(Q) gravity as a possible resolution of the H0 and S8 tensions with DESI DR2
The symmetric teleparallel framework brings about the possibility of alleviating cosmological tensions. The current burning issue in cosmological studies is the increase in discrepancies in measurements from several surveys. Here, we have focused on and tensions, which are important factors in describing the evolution of the Universe from primordial perturbation to late-time acceleration. Additionally, the consistency of the sound horizon is verified against the Planck results. The gravity model is constrained using recently obtained data. Implementing gravitational wave data to study late-time acceleration is one of the key features of our study. Since standard sirens show promising results, the implementation of gravitational waves to probe dark energy is an interesting study. Through our work, we introduce this possibility by performing statistical MCMC analysis for late-time cosmological evolution. Also, the and tensions are explored utilizing gravitational wave data alongside other prominent datasets, such as the latest DESI BAO, redshift space distortion, cosmic chronometers, Pantheon+SH0ES, and cosmic microwave background data. With the results obtained, we analyzed the profile of cosmological parameters. Finally, the study presents the tension of the model with observations, which is found to have a much lower magnitude compared to the current trend. Thus, the considered f(Q) model alleviates tension, making it the best candidate for further investigation. The Author(s) 2025. -
Robust and imperceptible image watermarking using chaotic map-integrated quantum-inspired multi-objective cuckoo search optimization
With the rapid growth of multimedia data transmission for secure and reliable communication has become critical due to increasing cyber threats. This paper presents a Chaotic Map-integrated Quantum-Inspired Multi-Objective Cuckoo Search (CMQICS)-based watermarking approach to achieve high imperceptibility, robustness, and embedding efficacy. The proposed approach integrates quantum-inspired cuckoo search optimization with chaotic mapping to enhance watermark embedding. A multi-image watermarking scheme is also used to strengthen payload capacity while minimizing visual distortion. The embedding process operates in the frequency domain using a hybrid Discrete Cosine TransformTwo-Dimensional Discrete Wavelet Transform (DCT-2DWT) combined with a zig-zag scanning strategy. This ensures attack resilience. The experimental results show that CMQICS achieves a Peak Signal-to-Noise Ratio (PSNR) of approximately 89 dB, a Structural Similarity Index Measure (SSIM) of 0.99, and an average embedding time of around 1s. Randomness analysis further validates the security of the embedded watermark. The comparative evaluations states that the CMQICS outperforms existing approaches. The Author(s) 2025. -
Pearson correlation-based clustering with collaborative task allocation in 5G Industrial Internet of Things divergent health networks
Simultaneous task allocation is crucial for enhancing service quality in Industrial Internet of Things (IIoT) environments. The distribution and management of tasks remain among the biggest challenges in the IIoT era. Efficient allocation strategies are needed to enable transparent network configurations and maximize task throughput. Although recent methods address the dynamic management of objects, they often overlook the correlations between tasks and their associated functionalities. This paper introduces a novel Connected Harmonical Adaptive Task Allocation (CHATA) model for IIoT health networks to ensure fair task distribution. CHATA leverages similarity measures of object functionalities to identify the most suitable object to perform each task. Simulations conducted in NS-3 demonstrate that CHATA achieves up to 90% allocation efficiency in 5G Radio Access Technologies IIoT health environments and significantly outperforms recent approaches in task assignment performance. The Author(s) 2025. -
A scalable scheduling and resource management framework for cloud-native B2B applications
In modern cloud computing environments, customers increasingly depend on on-demand resource provisioning to handle dynamic workloads. However, fluctuations in job arrival rates can result in prolonged queue times, which negatively affect overall system performance. Although existing scheduling algorithms provide efficient job management, they often fail to account for the combined impact of queue delays and the need for flexible resource provisioningparticularly in business-critical applications. In order to tackle these issues, the paper proposes a new Optimized Job Scheduling and Resource Scaling (OJSRS) algorithm designed to improve job execution efficiency and support elastic resource management in cloud environments. The OJSRS algorithm integrates two key components: Tree-based Job Scheduling (TJS) and Automated Resource Scaling and Scheduling (ARSS). The TJS component constructs a hierarchical structure that concurrently maps incoming jobs to the most suitable Virtual Machines (VMs), thereby minimizing queue delays. Meanwhile, ARSS adjusts resource allocation dynamically, increasing or decreasing capacity according to workload requirements and cloud service provider policies, enabling responsive and adaptive provisioning. Experimental results show that the OJSRS algorithm increases resource utilization by approximately 510% and accelerates job completion through proactive resource scaling. This approach provides a significant performance advantage for cloud-native business applications that require both efficiency and scalability. The Author(s) 2025. -
Interface improvement and multiscale assessment of recycled concrete aggregates with epoxy resin polymer
Recycled concrete aggregate (RCA) exhibits challenges like weak bonding, high porosity, and inferior strength compared to natural aggregates. This study evaluates the effect of epoxy resin polymer treatment on RCA on enhancing compressive and split tensile strengths in concrete, replacing natural aggregates with untreated RCA (UTRAC) and treated RCA (ERTAC) at 25%, 50%, 75%, and 100% levels. The tests were conducted at 3, 7, and 28 days. UTRAC showed reductions of up to 26.32% in compressive strength and 35.38% in tensile strength at 100% replacement; ERTAC outperformed control concrete (CC) with gains of up to 26.32% in compressive strength (at 25%) and 122.73% in tensile strength (at 100%), identifying 25% as the optimum replacement ratio. SEM and XRD analyses confirmed improved particle packing, reduced porosity, and stronger interfacial transition zones (ITZ) in ERTAC. The Author(s) 2026. -
High-efficiency stepdown/step-up converter for series-connected energy storage system
This work introduces a novel stepdown/step-up converter designed to optimize the run time of series-connected Battery, whose voltage drops progressively with increased usage, eventually falling below the necessary operating levels. The proposed converter automatically transitions between stepdown, step-up, and stepdown/step-up modes based on a comparison of input and output voltages, with the stepdown/step-up mode restricted to the narrowest range to minimize its lower efficiency in power conversion. It supports an input voltage range from 2.5 to 8V and incorporates a capacitive coupling level shift circuit to maintain the gate-source voltage of the power transistor under 5V, protecting against gate oxide layer damage. Fabricated with 180nm BCD technology, the converters compact size is 1.44mm by 0.73mm. Testing reveals that this converter achieves up to 93% power conversion efficiency, an 11% improvement over conventional models, and supports an output current up to 500mA, a 67% increase, enhancing the performance and longevity of Battery in compact electronic devices. The Author(s) 2025. -
ANN and machine learning based predictions of MRR in AWSJ machining of CFRP composites
This study investigates the effectiveness of Abrasive Water Suspension Jet (AWSJ) Machining, a non-conventional erosion-based method, for machining carbon fiber-reinforced polymer (CFRP) composites. The focus was on analyzing key process parametersabrasive size, feed rate, and standoff distance (SOD)under submerged cutting conditions and their impact on material removal rate (MRR), kerf width, and surface roughness. Experimental trials were conducted, and advanced computational techniques, including Response Surface Methodology (RSM), Random Forest (RF), Extreme Gradient Boosting (XGBoost), and Artificial Neural Networks (ANN), were used for parameter optimization and predictive analysis. The results showed that submerged cutting significantly improved machining quality by reducing surface roughness and ensuring uniform kerf widths. Increasing the jet diameter in underwater conditions stabilized the nozzle, leading to smoother and more precise cuts. Among the predictive models, XGBoost demonstrated the highest accuracy and efficiency in forecasting MRR, while Random Forest and ANN provided competitive performance. The integration of RSM and machine learning (ML) techniques enabled effective optimization of machining parameters, showcasing the potential for cost-effective and high-precision CFRP machining. These findings are particularly relevant for industries like aerospace and automotive, where machining efficiency and precision are crucial. The Author(s) 2025.