SJEE

Time-Domain Iron Loss Estimation Using the Parametric Magneto Dynamic Model: Discretisation Sensitivity and a Skin-Depth-Based Heuristic

Timo Hörmann — 2026-07-14

This paper utilizes the Parametric Magneto-Dynamic (PMD) model, enhanced with an inverse hysteresis formulation, to estimate iron losses in the time domain under arbitrary excitation waveforms. The proposed PMD model reproduces the intricate magnetization dynamics of advanced steel grades, facilitating a systematic investigation into how metallurgical and geometric material properties - such as lamination thickness, alloy composition, and grain morphology - influence the required spatial and temporal discretisation for accurate loss estimation. The paper evaluates the consequences of discretisation parameters in the PMD model on both the accuracy of the iron loss prediction accuracy and computational overhead. The findings establish rigorous guidelines on the required discretisation schemes for purely sinusoidal excitation and for nonsinusoidal regimes with significant harmonic content across a broad frequency spectrum.

Dynamic Load Frequency Regulation in Two Area Power Systems via Pelican-Optimized PID Control

Manish Kumar Pandey — 2026-07-14

This article presents Pelican Optimization Algorithm (POA) for Load Frequency Control (LFC) in two area interconnected power systems. Extensive simulations in MATLAB indicate that the POA surpasses numerous established metaheuristic optimization algorithms, involving the Grey Wolf Optimizer (GWO), Hippopotamus Optimization Algorithm (HOA), Salp Swarm Algorithm (SSA), Particle Swarm Optimization (PSO), Teaching-Learning Optimization Algorithm (TLOA) and Whale Optimization Algorithm (WOA) regarding the mitigation of frequency deviations and improvement of dynamic response. The robustness of the POA based LFC is validated under scenarios of random load disturbances, varying system parameters, typical power system nonlinearities like generation rate constraints (GRC) and governor deadband (GDB), and with renewable energy sources. The studies highlighted in the article demonstrate superior adaptability and effectiveness of POA to optimally tune the gains of PID controller. The statistical analysis demonstrate that the proposed POA achieves convergence to the global optimum with reduced execution time (~12% less than the second best HOA) and exhibits lower mean (~15.2% less than the second best SSA) and standard deviation (~3.09% less than the second best HOA) over multiple independent runs. Stability analysis using Bode plots confirms that the proposed control strategy maintains adequate stability margins.

A Multi-Agent RAG Architecture for University Information Management: A Comprehensive Analysis of Performance, Reliability, and Cost

Turgay Tugay Bilgin — 2026-07-14

University regulations and announcements at Bursa Technical University (BTU) are dispersed across PDF files, web pages, and notice boards, causing information retrieval inefficiencies for students and administrative staff. This study presents the development and quantitative evaluation of BTU-Chatbot, an Agentic Retrieval-Augmented Generation (RAG) powered conversational assistant that consolidates fragmented institutional information into a single citation-aware dialogue interface. Ninety-seven PDF documents were processed using PyPDFLoader and regular-expression-based preprocessing, then embedded into 1536-dimensional vectors using OpenAI’s text-embedding-ada-v2 model and stored in a 29 MB ChromaDB collection. The vector-based retrieval layer selects the three most relevant passages per query using cosine similarity search. The upper layer implements a LangChain-orchestrated multi-agent ReAct loop, in which the retrieve tool accesses the vector database while the Google_search_univ tool performs domain-restricted searches limited to the “*.btu.edu.tr” domain. GPT-4o-mini, with a 128k context window, serves as the generative backbone. System reliability was measured using the RAGAS metric suite. The best performing run achieved Context Recall = 0.97, Context Precision = 0.99, and F1- RP (the F1 score of Context Recall and Precision) = 0.954, demonstrating nearperfect retrieval accuracy. The average cost per query was 6.6×10⁻⁵ USD, with 7.6 s of latency for typical 124-token exchanges. BTU-Chatbot demonstrates that an Agentic RAG pipeline can deliver source-grounded, citation-attributed answers to university-specific queries at low operating cost, although further improvements in generation faithfulness are needed before full deployment.

A Unified Transformer-BiLSTM and Graph Attention Network Framework for Explainable Multilingual Opinion Mining and Relationship Inference in Complex Social-Citation Networks

Manoharan Thangavel — 2026-07-14

The paper takes into consideration the rising requirement for accurate mining of opinion and inferring relations among entities as the amount of multilingual online information increases rapidly. Thus, the paper seeks a uniform statistical learning methodology for processing multiple languages and exploiting valuable relational discoveries. The paper introduces an innovative model which combines transformer-based context embedding, BiLSTM for capturing of sentiment flows, and GAT for examining relational data. Transformers model cross-lingual context probability distribution, BiLSTM models temporal transitions in sentiment, whereas GAT offers an attention-based statistical weighting on data structured as a graph. Examples include social networks and citation graphs. The model yields a sentiment classification accuracy and macro F1-score of 92.4% over the four multilingual datasets (English, Spanish, and Hindi), surpasses all baseline methods. The model has achieved state-of-the-art result 86.4% in relation inference with graph attention mechanism, and shows excellent reliability and generalization. In this work, we propose a statistically sound and integrated framework which incorporates contextual, sequential and relational modelling of multilingual opinion mining. Graph structured learning integrated with probabilistic encoding and temporal modeling is a novel technique for such tasks, which also shows a promising direction in terms of scalability and generalizability across different domains.

Enhancing Wi-Fi Router Placement with Unsupervised Machine Learning for Improved Network Coverage and Performance

Nahid Gudratli — 2026-07-14

The optimal placement of Wi-Fi routers is essential for ensuring strong and consistent wireless coverage, yet traditional approaches often fail to deliver comprehensive solutions. This study presents a novel application of unsupervised machine learning (ML) to improve Wi-Fi router placement by analyzing environmental factors, historical signal strength data, and user behavior patterns. Using a dataset of signal strength measurements from a multi-story building, we conducted feature engineering to identify key predictors and trained various ML models to predict the impact of router placement on signal performance. The models were assessed based on accuracy, robustness, and computational efficiency. Our results show that ML-driven placement strategies significantly reduce dead zones and enhance overall network performance. Moreover, the proposed approach streamlines the installation process and supports adaptive, realtime adjustments to changes in the environment or usage patterns, offering a scalable solution for modern network infrastructure. These findings have practical implications for network engineers and set the stage for future innovations in intelligent network design and optimization.

Two-stage Adaptive Robust Noise Cancellation System for a Hindi Speech Database

Vijay Kumar Gupta — 2026-07-14

This paper presents a two-stage adaptive noise cancellation system for noise cancellation of Hindi speech signals corrupted by babble and factory noise, under different input signal-to-noise ratio (SNR) levels. The corrupted signals with different SNR levels are passed through three different noise cancellation systems: an adaptive noise cancellation system with a fast block least mean squares algorithm (FBLMS) algorithm; a fully connected convolutional deep neural network (FCCDNN); and a two-stage system with an adaptive noise cancellation system consisting of the FBLMS algorithm followed by FCCDNN. The performance of each noise cancellation system is evaluated and compared. The parameters such as output SNR, perceptual evaluation of speech quality and shorttime objective intelligibility are used for testing the intelligibility and quality of the recovered speech signals. The proposed system shows a maximum SNR improvement of 13.997 and 15.2965 dB at an input SNR of −5 dB for factory noise and babble noise, respectively, with significant increments in both of the evaluation metrics. The proposed system outperforms speech enhancement systems based on FCCDNN and FBLMS in terms of all performance parameters.

Comparison of Theoretical and Experimental Pixel-on-Target Metrics for Early UAV Detection Using Multi-Sensor EO/IR System

Saša Vujić — 2026-07-14

This paper presents an analysis comparing theoretical model predictions and experimental measurements of the number of pixels on target (PoT) for unmanned aerial vehicles (UAVs) using multi-sensor electro-optical and infrared (EO/IR) systems. The primary objective is to evaluate the fidelity of theoretical range prediction models that use the number of PoT as a metric, comparing these calculated values against experimental field measurements. The data used in this analysis were acquired using both visible (VIS) and mid-wave infrared (MWIR) channels, with the target UAV positioned at known reference distances up to 900 m. The results quantitatively demonstrate the deviations between the two methodologies, providing critical insights into the practical limitations of performance prediction models for early drone detection systems operating under field conditions.

Analysis of Near-Field and Radiation Characteristics of a Mobile Phone in 4G and 5G Frequency Bands in the Presence of Shielding Structures

Ana Tatović — 2026-07-14

This paper presents a numerical investigation of the near-field and radiation characteristics of a mobile phone operating in 4G and 5G frequency bands. The analysis is performed at 1800 MHz, 2600 MHz, and 3500 MHz in the presence of shielding structures composed of materials with different electromagnetic properties. Three representative materials, namely aluminum, FR4, and a carbon-loaded polymer, are considered in order to evaluate the influence of conductive, dielectric, and lossy shielding configurations on the electromagnetic behavior of the device. The mobile phone model and shielding structures are developed and analyzed using the WIPL-D EM simulation software. The effects of shield material, shield distance, shield position, and phone orientation relative to the shielding plate are investigated through a comprehensive parametric study. The near-field behavior is evaluated using the electric field while the influence of the shielding structures on the antenna performance is examined through reflection coefficient, gain and radiation pattern analysis. The obtained results provide insight into the interaction between shielding structures and mobile phone antennas operating in contemporary wireless communication bands and may serve as guidelines for the design of shielding configurations that reduce nearfield electromagnetic field levels while preserving acceptable antenna performance.