ECTI Transactions on Electrical Engineering, Electronics, and Communications

Simultaneous Network Reconfiguration and Optimal Allocation of Multiple DGs in Radial Distribution System to Reduce Real Power Loss

2025-10-16T23:53:55+07:00

This work addresses the challenging issue of minimizing actual power loss in radial distribution systems (RDS) by integrating network reconfiguration and multiple distributed generations (DG). A critical survey demonstrates that NR can reduce power losses on its own; however, the implementation process is challenging. While the incorporation of DGs can reduce losses, voltage instability and increased losses can result from poor placement or sizing. Consequently, it is crucial to allocate DGs optimally. The authors propose two modern optimization methods—the Rao-1 algorithm, which lacks metaphors, and the Sine Cosine Algorithm (SCA)—to optimize DG and NR allocation in order to resolve these challenges. This investigation employs Modified Load Flow (MLF) to ascertain the optimal network reconfiguration and loss sensitivity analysis to identify DG. Rao-1 and SCA are employed to size DGs in order to enhance voltage profiles and minimize actual power losses in IEEE 33-bus and 69-bus test systems. In comparison to SCA, Rao-1 enhances efficiency, convergence, voltage profile, and efficiency, particularly when multiple DGs are incorporated with NR. Rao-1’s exceptional performance in both systems is supported by statistical analyses. The proposed approach reduces loss by 74.9% for the IEEE 33-bus system and 83.8% for the IEEE 69-bus system when contrasted with the Enhanced Sine Cosine Algorithm (ESCA).

Enhanced Data-Driven Load Forecasting Framework for High PV Penetration Electrical System: A Case Study of SUT’s Campus

2026-01-01T08:59:48+07:00

This research proposes an approach to develop a one-day-ahead electrical load forecasting model using deep learning techniques. It compares deep learning models with basic models, namely Moving Average, DNN, LSTM, Bi-LSTM, CNN-LSTM, and Attention-LSTM, to improve accuracy and reliability through Mutual Information (MI) and Shapley Additive Explanation (SHAP) analysis. This approach utilizes data collection and feature engineering to optimize the context of an electrical loads with integrated solar power generation, Using Suranaree University of Technology (SUT) campus electrical system as a case study. Three accuracy indices are examined: Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Mean Square Error (RMSE). Among the NN-based models, Bi-LSTM and LSTM tend to provide the best overall forecasting performance from the input data, and the Bi-LSTM model is the most accurate model with the lowest metric value among all models. The lag load feature, or the value of the electrical load in the past one day, is the feature that is most related to the forecasting target and has the most impact on the model’s decision.

Comprehensive Study of Switching Overvoltages in Transmission Line of Northern Laos Power Grid Interconnected Across Pakbeng–Tha Wang Pha Region

2025-12-04T11:58:58+07:00

Switching and temporary overvoltages pose critical challenges to the reliable operation of extra-high-voltage (EHV) transmission networks. This study analyzes the 500 kV Pakbeng–Tha Wang Pha (PKB–TWP) cross-border interconnection between Laos and Thailand, an essential corridor for regional hydropower integration. A probabilistic electromagnetic transient (EMT) model was developed in PSCAD/EMTDC to evaluate switching overvoltages (SOV) and temporary overvoltages (TOV) under realistic operating conditions. Monte Carlo simulations incorporating Gaussian breaker pole scatter and representative TOV scenarios-such as the Ferranti effect, load rejection, and transformer inrush-were performed. Uncontrolled line energization produced surges up to 2.251 p.u., while reclosing after three-phase-to-ground faults reached 3.445 p.u., both exceeding the ≈2.0 p.u. insulation coordination threshold. The coordinated application of 444 kV metal-oxide surge arresters (MOSA) and 110 Mvar shunt reactors successfully reduced all surges below 2.0 p.u., ensuring compliance with IEC 60071-2 withstand requirements. Temporary overvoltages were contained within an acceptable level: 1.157 p.u. (Ferranti effect), 1.317 p.u. (90% load rejection), and 1.71 p.u. (transformer inrush). Spatial analysis identified resonance hotspots near midline sections, emphasizing the importance of distributed monitoring. Overall, the PKB–TWP interconnection was verified to be technically robust; the proposed methodology offers a practical framework for future EHV insulation coordination in Southeast Asia.

Distribution System Performance through Charging Station Integration Feeder Reconfiguration and Distributed Generation to Reduce Greenhouse Gas

2026-01-20T11:04:21+07:00

This paper presents an analysis of distribution system
performance through charging station, integration
feeder reconfiguration, and distributed generation to
reduce greenhouse gas emissions. The test system is
a 33-bus distribution network, and the simulations are
performed using the MATLAB. The case study is divided
into five scenarios. Case 1 is the base case. Case 2
involves the installation of 10 charging stations. Case
3 applies feeder reconfiguration. Case 4 includes the
installation of three distributed generation units. Case 5
includes the installation of three battery energy storage
system units. The results show that installing EV
charging stations causes a voltage drop in the distribution
system. The feeder reconfiguration technique helps
balance the load within the system, resulting in reduced
power loss. The installation of distributed generation
raises the voltage level at the load, thereby reducing
electricity consumption from the grid. This leads to
lower power losses and contributes to a reduction in
greenhouse gas emissions. Power loss is reduced by
24.97% in Case 3 and by 59.09% in Case 4, compared with
the Case 2 (the charging-station installation case).

Lithium-Ion Battery State of Health Estimation Using Resampling-Based Data Simplification Deep Learning Techniques

2025-12-15T21:01:59+07:00

Lithium-ion batteries (Li-ion) are widely used in various applications due to their high efficiency and reliability. However, as these batteries are continuously used, their performance gradually degrades over time, making accurate estimation of the State of Health (SOH) of increasing significance. One of the key challenges in SOH estimation lies in the nature of the measurement data collected during charge and discharge processes, which is typically time-series data with large volume and complexity. This results in increased computational load and reduced efficiency of the estimation models. To address this issue, this research proposes a data simplification method using a resampling technique aimed at identifying the optimal sampling level that maintains estimation accuracy while reducing computational cost. Four deep learning (DL) models Long Short-Term Memory (LSTM), Bidirectional Long Short-Term Memory (Bi-LSTM), Attention-based LSTM (ATT-LSTM), and Gated Recurrent Unit (GRU) are employed in this work. These models are trained and evaluated using public battery datasets that contain complete charge-discharge cycles. The proposed methods had been tested with the Oxford Battery Dataset, which the experimental results demonstrate that the proposed approach achieves higher estimation accuracy while significantly reducing computation time.

A High-Efficiency Low Dropout Regulator in 180 nm CMOS for Power-Constrained SoC Applications

2025-12-16T20:39:37+07:00

Low Dropout Regulators (LDOs) are one of the most vital components in today's System-on-Chip (SoC) platforms, particularly in battery-powered and noise-sensitive applications such as biomedical and analog sensor interfaces. Their fundamental function is to deliver an uncompromised regulated output voltage with as minimal input-to-output voltage difference, or dropout voltage, as achievable while maintaining low power consumption and high noise rejection. This article presents the design and simulation of a low-voltage SoC power management integrated LDO with 180 nm CMOS technology. The adopted design features a PMOS pass transistor, resistive feedback network for accurate voltage regulation, and two-stage differential error amplifier with Miller compensation for frequency stability. Powered by a 1.8 V supply, the regulator produces a stable output of 1.6 V with a dropout voltage of at least 200 mV. The LDO has a load current of 1 µA to 10 mA and exhibits a phase margin of 60°, gain-bandwidth product of 7.57 MHz, and a quiescent current of as low as 10 µA. The LDO also displays satisfactory Power Supply Rejection Ratio (PSRR) performance, −54 dB at 1 kHz, with the ability to supply solidity against supply noise and load transients. Simulation results validate the LDO's capability to deliver stable voltage with negligible overshoot and fast settling during load transitions. The compact area involves no charge pumps or outside components, making it ideal for area- and power-constrained SoC integration. Overall, the proposed LDO compromises between power efficiency, noise performance, and integration simplicity.

Wild Gibbon Optimized Sparse Attentive Convolutional Transformer Network for Fault Diagnosis in Electric Vehicles

2025-12-30T09:25:58+07:00

Fault Detection and Diagnosis (FDD) is critical for maintaining the security and dependability of Electric Vehicles (EVs). The electric motor drive and battery system, along with the EV’s powertrain and energy storage, are essential parts that are susceptible to a variety of malfunctions. Henceforth, this paper presents a fault diagnosis framework dependent on deep learning (DL) and nature-inspired optimization to classify faults with high accuracy. The application uses the NEV Fault Testing Dataset, which contains critical operational signals, including voltage, current, motor speed, temperature,
vibration, and humidity signals. Data normalization is applied for ensuring uniformity across the dataset while improving learning capability of model. Exploratory Data Analysis (EDA) is employed for identifying hidden patterns in the dataset and examining the contribution of each variable to the features’ distribution. Feature engineering is used for extracting meaningful variables that influence fault-related behavior. The proposed novel model, Wild Gibbon Optimized Sparse Attentive Convolutional Transformer Network (WG-Sparse ACTNet), integrates sparse convolutional methods and attention mechanisms for effective and accurate fault classification while the Wild Gibbon Optimization Algorithm (WGOA) is employed for hyper-parameter tuning to further
enhance model accuracy. This model is implemented in Python and evaluated using standard performance metrics, which achieved an accuracy of 99% and a precision, recall, and F1-score of 98%, respectively.

Power Quality Improvement in Grid-Connected PV Systems Using Sequence Control

2025-11-03T11:16:14+07:00

A modified control strategy for grid-connected inverters is presented to enhance the integration of renewable energy sources while effectively addressing challenges arising from unhealthy voltage and current conditions. The study introduces an improved instantaneous reactive power theory–based control scheme that utilizes positive-sequence components of grid voltages and negative-sequence components of load currents to regulate the grid-tied converter. To further enhance dynamic performance, a proportional–integral controller is incorporated to maintain dc-link voltage stability under abnormal grid and load scenarios. The proposed control approach is implemented and validated in MATLAB/Simulink across a range of operating conditions, including steady-state, dc offset, grid harmonics,
unbalanced grid voltages, and non-linear as well as unbalanced loads, both with and without solar PV integration. Simulation results demonstrate significant improvements in power factor correction and reactive power compensation, with robust controller performance under varying irradiance levels. Comparative analysis with recent control schemes highlights the superior performance and effectiveness of the proposed method. Moreover, the total harmonic distortion values obtained across multiple test cases comply with IEEE-
519 standards, confirming the reliability and robustness of the proposed control technique for grid-connected renewable energy systems.

Improving the PQ Harmonic Detection Method for Active Power Filters in Single-Phase Power Systems

2025-12-24T09:51:53+07:00

This study presents a novel improvement to the instantaneous power theory (PQ) for harmonic detection in active power filters (APFs) applied to single-phase power systems. The method is designed to eliminate harmonic currents and improve the power factor for nonlinear loads. The key contribution of this research is the development of the PQ-based harmonic detection method that significantly reduces computation time and supports electrical systems with distorted voltage sources while preserving the filter’s performance. The proposed method was verified through real-time simulation with the OPAL-RT hardware. Test results indicate that the percentage of the total harmonic distortion of source current (%THDi) was reduced from 18.11% (before compensation) to 0.16% after applying the improved PQ methods. This value is lower than the traditional PQ method and complies with IEEE standard 519-2022. Moreover, the power factor was also improved to a unity after the compensation. In addition, the improved PQ method provided computation time reduction up to 42.76% compared to the combined PQ method. The improved PQ method achieved these results with lower computational complexity, making it a practical and reliable solution for implementation in microcontroller-based real-time control systems while optimizing the resources of the high-performance microcontroller, underscoring its suitability for resource-constrained embedded systems in practical APF applications.

2-D Interference Channel Segmentation for Modified Factor Graph-Based Detection on BPMR System

2025-10-22T14:24:05+07:00

Increasing the areal density of the bit-patterned media (BPMR) recording system can be achieved by reducing the size of the magnetic grains. However, this increases the problem of inter-track interference, also referred to as two-dimensional (2-D) interference channels, which can degrade the performance of the read channel in magnetic recording systems. To alleviate the effects of the 2-D interference channel on the BPMR system. This work proposes two methods of improved factor graph-based (FGB) detection using a segmentation of the 2-D interference channel coefficients that exploits the relationship between the main bit and its nearest neighbors. Similarly, the scheme of message-passing is a hierarchy based on levels from neighboring bits to the main bit. Simulation results show the bit error rate (BER) performance between the conventional FGB detector and the modified FGB detectors on the BPMR channel at an areal density of 3 Tb/in² with multi-track processing. The BER performance of both modified FGB detectors outperforms the conventional FGB detector.

A Hybrid Local Search and Genetic Algorithm with Enhanced Population Initialization for the Traveling Salesman Problem

2025-11-26T14:03:59+07:00

The Traveling Salesman Problem (TSP) remains a fundamental NP-hard combinatorial optimization challenge with significant theoretical and practical implications. While Genetic Algorithms (GAs) have demonstrated considerable promise for solving TSP instances, their performance is often hindered by poor initial population quality, leading to slow convergence and suboptimal solutions, particularly for large-scale problems. This paper presents a novel Hybrid Local Search and Genetic Algorithm (HLGA) that addresses these limitations through strategic population initialization. The proposed approach synergistically combines high-quality solutions generated by 2-opt and 3-opt local search heuristics with diversity-enhancing random permutations to create an enriched initial population. This hybrid initialization strategy accelerates convergence while maintaining solution space exploration capabilities. Comprehensive experiments conducted on 27 TSPLIB benchmark instances (30-200 cities) demonstrate HLGA's superior performance compared to standalone 3-opt heuristics and five state-of-the-art algorithms (GGSC-SSA, SCGA, IMODE/J2020, ACO-DSA, and RL-SA). Notably, HLGA achieves or surpasses the Best-Known Solution (BKS) for 11 instances, with an overall average relative error of 9.22%—significantly outperforming the 3-opt baseline (83.91%). These results validate that integrating targeted local search into population initialization substantially enhances both convergence speed and solution quality, establishing HLGA as a competitive and scalable approach for large-scale TSP optimization

A Development of a Portable Sugar Measuring Device for Americano Coffee Using The Electrical Conductivity Principle

2026-01-05T11:31:27+07:00

This research presents the development of a portable device for non-destructive sugar measurement in Americano coffee using electrical conductivity (EC) and temperature compensation principles. The study identifies key factors influencing the total dissolved solids (TDS) in Americano coffee, including water volume, coffee concentration, and sugar content. Experimental results show an inverse correlation between sugar content and TDS, enabling the estimation of sugar levels through EC analysis. A prototype was designed incorporating an analog EC sensor, a temperature sensor, and an Arduino Nano microcontroller. Calibration of the EC sensor achieved an accuracy within 0.60% error, and temperature compensation equations were derived for both hot and iced Americano variants to enhance precision. The device classifies sugar content into low, normal, and high levels based on WHO guidelines and demonstrated 100% accuracy across validation tests for both beverage types. This portable solution provides a rapid and user-friendly method for assessing sugar levels in coffee, with potential applications in health monitoring and dietary management.

Cervical Cancer Detection using Deep Learning and Image Processing Techniques

2025-11-13T13:13:06+07:00

Cervical cancer is the second most common cancer among women across the globe. Detecting abnormal cervical cells at an early stage is vital for prompt treatment and improved survival rates. This project focuses on developing an effective approach to identify cervical cancer in Pap smear images using modern digital image processing and deep learning techniques. The system begins by preprocessing the medical slides to improve image clarity and reduce noise and then applies segmentation methods to highlight and separate the regions of interest. The significant tasks include preprocessing methods such as resize and normalization; segmentation methods such as DeepLabV3, Otsu and Canny edge detection and feature extraction methods
such as ResNet101, ResNet152, AlexNet, Inceptionv3, and VGGNet16 to extract features of the cell, such as size, texture and shape of the cell, and shape and color of the nuclei. To distinguish between cancerous and noncancerous images, various machine learning algorithms are employed, including Decision Tree, Random Forest, Logistic Regression, and Support Vector Machine (SVM). The proposed methodology is evaluated on the SIPaKMeD dataset, with performance measured using established metrics such as precision, accuracy, recall, specificity, F1-score and harmonic mean to validate its robustness and reliability. By presenting a cost-effective, automated diagnostic framework to support pathologists in early cervical cancer detection, this study aligns with the broader objectives of healthcare innovation. It has the potential to enhance diagnostic efficiency and contribute to improved public health outcomes. Finally, the combination of feature extractor VGGNet 16 and classifier decision tree gave the highest performance.