Documents published in Scipedia

• Numerical Simulation and Optimization of Grey Relational Analysis Models for Panel Data

  X. Liu, F. Jian, X. Tang

  Rev. int. métodos numér. cálc. diseño ing. (2025). Vol. 41, (2), 101

  
  

  Abstract

  The indicators-coupled grey relational analysis (ICGRA) models are important in clustering panel data with cross-sectional dependence. However, there is still little research on performance validation for the various ICGRA models. In this paper, we investigate the performance of the existing ICGRA models accounting for the reordering of indicators. Firstly, the robot execution failures (REF) dataset of the University of California Irvine (UCI) machine learning database is adopted to validate the robustness of four traditional ICGRA models. Then, we compared the grey relational orders for all arrangements of indicators in panel data. Simulation experiments showed that the four ICGRA models are not all robust against the grey relational order. To resolve this problem, we adopted the mean value theory and deep modeling to optimize the four models and compared them with the tetrahedral grey relational analysis (GRA) model that considers the coupling effect between indicators on the grey relational order, as well as with the k-nearest neighbor (KNN) algorithm. Results show that the classification accuracy of the averaged absolute GRA model was 97.73%, the other optimized ICGRA models and the k-nearest neighbor (KNN) method all achieved 100% accuracy, while the tetrahedral GRA model has an accuracy of 83.33%. Therefore, the average grey incidence degree for all arrangements of indicators and deep modeling significantly improves the stability of models and enhances the clustering accuracy in different cases.

  Abstract
  The indicators-coupled grey relational analysis (ICGRA) models are important in clustering panel data with cross-sectional dependence. However, there is still little research [...]

  • 3
  •  read

• Optimized Multimodal Healthcare Image Fusion Using U2Net Restormer with Dilated Dense Encoder–Decoder and Haar-Based Feature Selection

  M. Shafiq, W. Obidallah, M. Albathan, T. Kamal

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 84

  
  

  Abstract

  Multimodal medical imaging plays a pivotal role in clinical diagnostics by integrating complementary anatomical and functional information from modalities such as Computed Tomography (CT), Magnetic Resonance Imaging (MRI), Positron Emission Tomography (PET), and Single-Photon Emission Computed Tomography (SPECT). Despite notable progress, existing fusion approaches continue to face persistent challenges. Convolutional Neural Network (CNN)-based methods often suffer from information loss due to convolutional down-sampling, while Transformer architectures, though effective at capturing global dependencies, incur high computational costs and rely on large-scale pretraining. Generative Adversarial Network (GAN)-based fusion models can generate visually realistic outputs but are prone to training instability and limited reproducibility. In addition, prior studies frequently adopt inconsistent evaluation metrics, with insufficient emphasis on clinical interpretability and robustness, hindering real-world deployment across heterogeneous datasets and institutions. To address these limitations, this study proposes a U-shaped Nested Network – Restoration Transformer (U2Net–Restormer) framework with a Dilated Dense Encoder–Decoder architecture for robust multimodal medical image fusion. The framework integrates hierarchical multiscale representation learning with residual global contextual refinement. To enhance discriminative capability, an optimized Haar-based feature selection strategy is introduced to preserve high-gradient structural and functional details while reducing feature redundancy. Furthermore, an attention-driven fusion mechanism adaptively weights modality-specific contributions, enabling effective integration of heterogeneous information. The proposed method is evaluated on the Augmented Alzheimer’s Neuroimaging Library (AANLIB) multimodal brain imaging dataset, covering CT-MRI, PET-MRI, and SPECT-MRI fusion tasks. Experimental results demonstrate consistent performance gains over state-of-the-art CNN-, Transformer-, and GANbased methods, achieving Structural Similarity Index Measure (SSIM) up to 0.963, Peak Signal-to-Noise Ratio (PSNR) of 42.1 dB, Feature Mutual Information (FMI) of 0.86, and Edge Preservation Index (EPI) of 0.91, with improvements of at least 4%–6% across modalities. Subjective evaluations by radiologists and neurologists report Likert scores up to 4.8/5 for structural visibility, functional fidelity, and diagnostic value. Robustness analysis under Gaussian noise (σ= 15%) further confirms the method’s resilience. Overall, the proposed framework delivers high-fidelity, clinically interpretable multimodal fusion suitable for diverse imaging scenarios.

  Abstract
  Multimodal medical imaging plays a pivotal role in clinical diagnostics by integrating complementary anatomical and functional information from modalities such as Computed [...]

  • 61
  •  read

• Automatic Detection Algorithm of Typical Hidden Dangers in Substation Based on Dynamic Convolutional Feature Extraction

  W. Wu, M. Liang, X. Wu, W. Zhang, Y. Ruan

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 79

  
  

  Abstract

  With the advancement of smart grid construction, substation equipment, being perpetually exposed to complex environments, is prone to latent hazards such as meter malfunctions and insulator cracks. Additionally, personnel violations may trigger safety incidents, directly jeopardising grid reliability. However, traditional manual inspections suffer from low efficiency and high rates of missed defects, while existing detection methods struggle to accommodate irregular defect classifications and multi-type defect characteristics, failing to meet engineering demands for real-time response and precise identification. To address these challenges, this study proposes a substation hazard detection framework (YOLOv10_DSE) based on an enhanced YOLOv10 (You Only Look Once version 10) architecture, designed to tackle multi-type hazard detection within complex substation scenarios. Firstly, a dynamic feature extraction module (C2fDSC) was designed, employing dynamic snake convolutions to enhance adaptive sampling capabilities for small targets and irregular defects. Secondly, a self-integrated attention module head (SEAMHead) was introduced to decouple localisation and classification tasks, thereby improving multi-type hazard discrimination accuracy. Finally, a bounding box regression loss function (inner_CIoU) was adopted to optimise small target localisation and irregular shape fitting. Experiments demonstrate that on a substation dataset containing 17 defect types, this method achieves mAP@0.5 and mAP@0.95 of 73.3% and 48.2%, respectively, representing improvements of 2.6% and 1.6% over the YOLOv10 baseline. This provides an efficient and reliable solution for multi-type defect detection in substations, holding significant engineering value for ensuring the secure operation of power grids.OPEN ACCESS Received: 15/10/2025 Accepted: 17/12/2025

  Abstract
  With the advancement of smart grid construction, substation equipment, being perpetually exposed to complex environments, is prone to latent hazards such as meter malfunctions [...]

  • 95
  •  read

• An Innovative Approach to Extracting and Classifying Non-Functional Requirements Using Machine Learning and NLP

  M. Shafiq, W. Obidallah, M. Albathan, T. Kamal

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 82

  
  

  Abstract

  This study presents a hybrid automated framework based on a combination of machine learning (ML) and natural language processing (NLP) approaches for the automatic categorization and extraction of nonfunctional requirements (NFRs) from free-text software development documents. Using the PROMISE dataset, this framework systematically integrates semantic representation learning, deep feature extraction, and kernel-based classification to improve the performance of NFR classification. Unlike current CNN-based approaches with end-to-end softmaxbased classification, our proposed method fundamentally decouples feature learning from decision making. The first approach is to use Word2Vec embeddings to capture semantic context, and then use Convolutional Neural Networks (CNNs) as high-level feature extractors. An Improved Support Vector Machine with a Radial Basis Function kernel (ISVM-RBF) is applied for final classification, enabling more discriminative decision boundaries to be drawn in the high-dimensional semantic feature space. We reveal a considerable performance improvement with the CNN– Word2Vec setup, achieving as high as a 90% precision, significantly outperforming standard ML classifiers. The study points to three main findings: (i) CNN-based feature extraction is an efficient approach for finding and classifying NFRs, (ii) the semantic representation provided by word embedding methods is clearly superior to other traditional methods used in NLP, and (iii) NLP preprocessing of text is crucial for enhancing classification accuracy. Finally, ISVM-RBF adapts kernel-based classification over features derived from CNN, which enhances the robustness of the model to semantic overlaps between NFR categories and alleviates challenges posed by potentially large textual datasets required to train such models. This hybrid CNN–ISVM-RBF design constitutes the methodological novelty of the proposed method and effectively distinguishes it from current state-of-the-art methods in the literature.OPEN ACCESS Received: 03/11/2025 Accepted: 22/01/2026

  Abstract
  This study presents a hybrid automated framework based on a combination of machine learning (ML) and natural language processing (NLP) approaches for the automatic categorization [...]

  • 77
  •  read

• Study on Basic Machine Learning Techniques to Detect and Classify Cardiovascular Diseases

  Y. Rawash, H. Ting, K. Chee

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 78

  
  

  Abstract

  The heart is an essential organ required to maintain the general health of individuals. Cardiovascular diseases (CVDs) have become the leading cause of death globally, replacing cancer and diabetes. Computer-based techniques have made it easier for physicians to diagnose various cardiac conditions, including heart failure. We are currently in the “information age,” a period characterized by the generation of millions of bytes of data every day. By applying ML algorithms’ techniques, such as Random Forest (RF), XGBoost, KNN, and GaussianNB, we can evaluate and compare the performance of machine learning classifiers for heart disease prediction and transform these data into information for the estimation of heart disease. The World Health Organization has estimated that in 2019, cardiac disease was responsible for 32% of all deaths worldwide. In this paper, we use a public data set (Indicators of Heart Disease) and hyperparameters to develop four classifiers—the Random Forest, XGBoost, KNN, and GaussianNB—and compare their performance. Based on the trial data, XGBoost was the best, with an accuracy of 91.63%, a precision of 94.40%, a recall of 88.50%, an F1-score of 91.36%, a specificity of 94.75%, and an AUC score of 97.39%. This study showcases the accuracy of machine learning systems in predicting cardiac conditions and can serve as a foundation for developing a decision-support tool aimed at detecting and preventing heart disease in its early stages.OPEN ACCESS Received: 09/10/2025 Accepted: 20/01/2026

  Abstract
  The heart is an essential organ required to maintain the general health of individuals. Cardiovascular diseases (CVDs) have become the leading cause of death globally, replacing [...]

  • 70
  •  read

• Modeling and Simulation of a Novel Permanent Magnet Motor for Enhanced Vehicle Steering Performance

  M. Assen, G. Geng

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 77

  
  

  Abstract

  This paper introduces a new design for a permanent magnet motor (PMM) tailored for automotive steering applications. By utilizing new permanent magnet topology, high-performance materials, and rotor designs, the newly proposed PMM design reduces torque ripple while boosting torque density. This design is integrated within a sophisticated steer-bywire (SBW) system, enabling seamless operation and reliable feedback, which are crucial for modern vehicles. The finite element analysis (FEA) substantiates the effectiveness of the proposed PMM, resulting in a 1.2% enhancement in magnetic flux density, a 4.65% increase in static torque, and a 4.4% reduction in torque ripple. This research addresses the issues in wheel steering technology and lays the groundwork for future automotive motor technologies. And also, it provides valuable insights into the development of automotive technologies, paving the way for enhanced performance and sustainability in the automotive sector. Furthermore, these improvements not only aim to enhance driver satisfaction but also align with the industry’s transition toward greener technologies.OPEN ACCESS Received: 05/10/2025 Accepted: 15/12/2025

  Abstract
  This paper introduces a new design for a permanent magnet motor (PMM) tailored for automotive steering applications. By utilizing new permanent magnet topology, high-performance [...]

  • 132
  •  read

• Least Square Support Vector Machine Framework for Meshless and Accurate Solution of Higher Order Boundary Value Problems with Comparative Analysis of Machine Learning Techniques

  B. Mishra, S. Chakraverty

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 74

  
  

  Abstract

  This paper presents an enhanced Least Squares Support Vector Machine (LS-SVM) approach for meshless and accurate solution of higher-order boundary value problems (BVPs) that commonly arise in structural mechanics, fluid dynamics, and other engineering fields. The discussed method formulates thirdand fourth-order linear and nonlinear ordinary differential equations (ODEs) as data-driven optimization problems, eliminating the need for traditional mesh-based discretization. Leveraging a Radial Basis Function (RBF) kernel and regularization-based control of model complexity, the LS-SVM captures complex solution behaviour while maintaining stability and smoothness. The meshless nature of the model ensures geometry-independence, making it suitable for irregular or multi-point boundary conditions. A comparative analysis with established machine learning techniques, including Ridge Regression (RR), classical SVM, Random Forest (RF), and Extreme Gradient Boosting (XGB), demonstrates the competitive accuracy, robustness, and efficiency of LSSVM. The results highlight its potential as a promising solver for nonlinear and multi-point problems where meshless methods are advantageous. The results highlight its potential as a promising solver for simulation-based workflows in computational mechanics and scientific computing, where adaptability, generalization, and reliability are critical.OPEN ACCESS Received: 25/08/2025 Accepted: 21/10/2025

  Abstract
  This paper presents an enhanced Least Squares Support Vector Machine (LS-SVM) approach for meshless and accurate solution of higher-order boundary value problems (BVPs) that [...]

  • 61
  •  read

• An Improved ORB-Based Multi-View Stereo Matching Algorithm for Accurate Visual SLAM

  X. Hu, D. Xue

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (3), 72

  
  

  Abstract

  This study addresses image matching in visual Simultaneous Localization and Mapping (SLAM) to enhance synchronous positioning and mapping using a multi-view stereo matching algorithm. Utilizing a vision odometer and a binocular camera, multi-view stereo images are captured. An improved ORB (Oriented FAST and Rotated BRIEF) algorithm extracts feature points and establishes binary descriptors, with similarity computed using Euclidean distance to enable stereo matching. Mismatches are filtered using parallax constraints and corrected through triangulation. Loopback detection minimizes cumulative drift in camera position, improving spatial perception. The back-end optimization employs graph optimization theory to refine the position and pose of the binocular camera and landmarks, addressing random noise and errors. Experimental results indicate effective feature extraction, with mismatches limited to three, overall positioning error under 100 mm, and directional error within 2°.OPEN ACCESS Received: 11/08/2025 Accepted: 14/01/2026

  Abstract
  This study addresses image matching in visual Simultaneous Localization and Mapping (SLAM) to enhance synchronous positioning and mapping using a multi-view stereo matching [...]

  • 175
  •  read

• Analytical Solution for the Internal Forces of the Nuclear Power Plant’s Water Intake and Outlet Tunnel under P Waves

  Y. Duan, L. Wu, D. Wang, R. Li, D. Jiang

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (2), 70

  
  

  Abstract

  Analytical solutions provide an efficient means to evaluate the internal forces of tunnels and elucidate the relationships among key influencing parameters. This paper presents an analytical solution for the internal forces of the nuclear power plant’s water intake and outlet tunnel under P waves. In order to consider the slippage effect of the ground-tunnel and lining-lining interfaces, a spring-type tangential stiffness coefficient K iis introduced. Moreover, the proposed analytical solution can accommodate tunnels with an arbitrary number of lining layers and treat the linings as thick-walled cylinders, providing higher analytical accuracy. The validity of the proposed analytical solution is demonstrated through a comparative analysis with the traditional analytical solution and finite element simulations. The comparison results confirm the accuracy and superiority of the proposed method. A comprehensive parametric investigation is conducted to examine how the tangential stiffness coefficient K i, the flexibility ratio F, and the characteristics of the isolation layer influence the tunnel’s seismic behavior.OPEN ACCESS Received: 25/12/2025 Accepted: 28/01/2026 Published: 20/03/2026

  Abstract
  Analytical solutions provide an efficient means to evaluate the internal forces of tunnels and elucidate the relationships among key influencing parameters. This paper presents [...]

  • 105
  •  read

• Confidence Level-Based Interval-Valued p, q, r-Spherical Fuzzy Sets with Application to Industrial Waste Management

  U. Habiba, M. Sarwar, A. Alhussan

  Rev. int. métodos numér. cálc. diseño ing. (2026). Vol. 42, (2), 66

  
  

  Abstract

  Industrialization plays a substantial role in a country’s development and economic growth. As a developing nation, Pakistan heavily relies on its steel industry to drive economic progress. Pakistan Steel Mills Corporation (PkMC) is the largest steel producer in the country, manufacturing thousands of tonnes of steel annually. However, this high steel production has led to a significant increase in industrial solid waste. While various methods are available for managing industrial waste, the selection of the appropriate technology is a complex process due to the wide range of strategies and the multiple factors involved in the decisionmaking process. Existing research in the interval-valued p, q, r-spherical fuzzy sets (IVp, q, r-SFS) environment assumes 100% confidence-level from decision makers in evaluating scenarios, but real-world situations often differ from ideal situations. To address this limitation, this study incorporates confidence-level with IVp, q, r-SFS, to identify effective and sustainable waste management strategies for PkMC. Decision-makers can evaluate alternatives by assigning corresponding confidence values using this model to better capture uncertainty. This study develops and analyzes basic operational laws, averaging, and geometric operators, outlining their desirable properties. This study presents a step-by-step algorithm for the proposed MCDM methodology, demonstrating its reliability and efficiency through an industrial waste management example. Model results indicate energy recovery as the most suitable alternative. Comparative analysis further validates the effectiveness of the proposed model. This research offers helpful tips for improving decision-making in waste management and points out the importance of a robust methodological framework in tackling complex, uncertainty-driven challenges.OPEN ACCESS Received: 13/11/2025 Accepted: 19/12/2025 Published: 20/03/2026

  Abstract
  Industrialization plays a substantial role in a country’s development and economic growth. As a developing nation, Pakistan heavily relies on its steel industry to drive [...]

  • 91
  •  read