Eaton (Taiwan)

Partial discharge (PD) activities are a major factor in insulation failure, accounting for over 60% of damage to high voltage transformers. Common causes of PD are manufacturing defects, environmental stress, and high-voltage surges. In this study, the effect of PD has been examined on minor and major insulation of a Power Transformer subjected to high voltage surges. The considered test voltages are high voltage operational surges like the Standard Lightning Impulse (SLI), Standard Switching Impulse (SSI), Oscillating Lightning Impulse (OLI), and modified SLI with tail chopped or additional overshoots. The OLI and modified SLI with superimposed oscillations or tail chopped are used to simulate the real-world like stresses on the transformer. To analyze the insulation conditions, the coherence function (CF) of the winding current is calculated at the neutral end of the transformer. CF is a statistical tool that measures the linear relationship between two signals in the frequency domain. The CF values are calculated for the simulated winding current obtained when partial discharges of varying magnitude are present versus when they are absent under different high frequency test voltages. From the results, it is found that when PD occurs across the main insulation, superimposed voltages consistently show the strongest correlation, while SLI with tailchopped voltages show minimum coherence with PD activity. The effect of varying the PD magnitude is negligible for minor insulation. This preliminary simulation study establishes a foundation for future experimental research to study the interaction of PD and test voltages on winding insulation, ultimately aiming to improve transformer performance, safety, and cost-effectiveness.

The Interstellar Mapping and Acceleration Probe (IMAP; McComas et al. in Space Sci Rev 214:116, 2018a; McComas et al. in Space Sci Rev 221:100, 2025) is a NASA mission designed to study two of the most important issues in heliophysics: the interaction of the solar wind with the local interstellar medium and the acceleration of energetic particles. These two puzzles are surprisingly intertwined because particle acceleration in the inner heliosphere plays a large role in the interactions at the edge of the heliosphere. The ten instruments on IMAP, situated near the Earth-Sun L1 Lagrange point, include both in situ and remote-sensing observations and are designed to work together to explore these questions. The High-energy Ion Telescope (HIT), described herein, measures in situ ions from hydrogen through nickel at energies ranging from a few MeV/nucleon to some tens of MeV/nucleon, depending upon species. While the primary focus of HIT is Solar Energetic Particles (SEPs), HIT is also sensitive to Anomalous Cosmic Rays (ACRs) and Galactic Cosmic Rays (GCRs) that are also present at these energies. Working with the other IMAP instruments, HIT studies how the variation in SEP intensities, spectra, anisotropies, and composition inform the acceleration processes that generate these particles, and, in turn, how these high-energy particles affect the outer heliosphere. This paper details the design and operation of the HIT instrument.

In big data era, real-time monitoring and alerting are the critical for quickly detecting and reacting to significant occurrences enabling organizations to reduce damage, increase efficiency, and preserve security. Apache Kafka is a real-time distributed streaming platform to prove high-throughput, low-latency data pipelines. Monitoring Kafka with Prometheus, Grafana, and JMX Exporter is a powerful technique to maintain the health, performance and reliability of real-time data. The system focuses on the performance evaluation of real-time alerting using Grafana, Prometheus and JMX Exporter for Kafka monitoring. The system configures real-time alert rules based on Kafka metrics and evaluates reliability, responsiveness and effectiveness of alert detection in various Kafka failure cases, such as broker failures, high under-replicated partitions, and JVM utilization of resources. The experimental results demonstrate when designed correctly, the integrated system provides accurate and timely alerts with low latency, replica configuration and a heap ratio for response time. These results indicate that open-source monitoring tools can offer a reliable and affordable Kafka infrastructure alerting strategy.

Fine-grained network traffic data is essential for critical network operations including performance monitoring, forensic analysis, and intrusion detection. However, the storage requirements for long-term retention of such high-fidelity data present significant cost challenges. We present a novel machine learning model called PRNet that combines formal methods with neural representation learning to generate compact, reconstructable encodings of network traffic. Our approach employs domain-constrained autoencoders enhanced with two innovations: (1) a Knowledge Augmented Loss function that incorporates network protocol semantics, and (2) a constraint enforcement module that preserves logical relationships between features. Experimental results demonstrate significant improvements across multiple dimensions: a 16.4% average gain in reconstruction accuracy, near-perfect recovery of categorical header fields and a 39.1% reduction in reconstruction error for session-level encodings compared to packet-level approaches under 75% missing data rate. For temporal features, our constraint module eliminates timestamp violations while ensuring logical consistency, with a boost of 73.8% in Spearman correlation, and 94.2% drop in Wasserstein distance for packet interarrival times. Our approach enables efficient network data preservation while maintaining high reconstruction quality for most downstream analytical tasks, though some numerical features remain challenging.

Distribution systems have traditionally received the least attention in the bulk picture of power systems; they are yet the most "critical component" as they directly impact customers’ perception and utilities’ reputation. While transmission systems have seen abundant literature, innovations, and applications of power electronic-based devices to optimize their performance, distribution systems still require more attention toward a flexible and dynamic scheme to cope with volatile energy consumption and production. As a replacement or enhancement of the traditional ones, Solid-State Transformers can play a central role in the energy management of distribution systems towards this goal, in addition to the mere voltage conversion. This manuscript, developed within the IEEE Task Force on Solid-State Transformer integration in distribution grids, offers a critical review of the Solid-State Transformer potential, services, and technical challenges for its integration in distribution systems. Particular attention is given to scientific trends and open points in the current research that must be addressed before utilities extensively integrate the Solid-State Transformer in distribution systems.

This study presents the development of high-performance power modules with advanced 3D printed ceramics, integrated cooling channels and Cu pinfins for thermal management. Thick film printed electronics was employed for circuit metallization. SiC dies were then attached using Sinter Silver process for improved thermal performance and reliability. Thermal and electrical testing was performed to demonstrate the module’s capability.

The International Laser Ranging Service (Carabajal et al. AGU2024) includes a number of multi-technique Core Sites which can be used to inter-compare different instruments and space technologies. Core stations which have each been occupied by more than one SLR system, as well as co-located VLBI systems provide a robust basis for monitoring advances in instrument development. The vertical resolution of SLR and VLBI systems has been established by Altamimi et al, AGU2024, who show scale agreement at the 1 mm level. The vertical agreement between SLR and VLBI systems at McDonald Geodetic Observatory, Wettzell Geodetic Observatory and the Matera Center for Space Geodesy is presented. The heights of these stations are seen to show no significant motion over periods which can extend for over 30 years. This stable behavior allows the combination of the strength of independent space techniques in a rigorous definition of the Reference Frame.

Banda Myzomela Myzomela boiei (S. Müller 1843) is a small honeyeater distributed in two subspecies across three island groups in Maluku Province, Indonesia: M. b. boiei on the Banda Islands, and c.350 km to the south, M. b. annabellae on the Tanimbar Islands and Babar. We compare data from 28 museum specimens and 21 sound recordings, and report the results of 152 playback experiments, to investigate the taxonomy of these three populations. Compared to M. b. annabellae, M. b. boiei has a c.10% longer wing, bill and tarsus, more extensive black breast-band, duskier posterior underparts, and a highly divergent song; in playback experiments, M. b. boiei proved unresponsive to recordings of M. b. annabellae. However, the songs of M. b. annabellae on the Tanimbars and Babar are also highly divergent, with populations on both islands unresponsive to the song of the other. Birds on Babar are also slightly larger than annabellae on the Tanimbars (with a c.10% longer tail) and may have more scarlet on the back. We consider the vocal differences between the three populations to be highly indicative of reproductive isolation and recommend they be treated as three independent species under a modern interpretation of the Biological Species Concept. Because no nomen is available for the Babar population, we name it as a new species.

Due to limitations imposed by manufacturing processes, heat exchanger designs previously were composed of tubes and fins (and several variants of these two geometric shapes).Recent advances in additive manufacturing techniques have led to the use of multi-objective topology optimization (TopOpt) to generate organic heat exchanger designs that satisfy mission-specific requirements.TopOpt-based heat exchanger design frameworks were initially developed assuming laminar flow.However, for a multitude of applications, the gaseous flow is often turbulent.The purpose of this paper is to develop a machine-learned (ML) model that can be incorporated into a TopOpt framework to design heat exchangers.Specifically, design inputs include domain size (10 10), temperature, and flow rates of hot (80 o C, 11 L/min) and cold fluids (55 o C, 67 m/s) and constraints based on energy density, pressure loss, and manufacturability.The hot and cold fluids are an aqueous solution of 50-50 ethylene glycol and air, respectively.A ML-based turbulence model is developed by first simulating the flow and heat transfer on 10 preliminary heat exchanger designs using two-equation, realizable -based steady RANS (Reynolds-Averaged Navier-Stokes) calculations.Then, a Gaussian Process machine learning model is trained to calculate the turbulent viscosity using area density as the input variable.The model is validated by performing an RANS calculation using a 0-equation eddy viscosity model with the eddy viscosity predicted by the GP model and comparing it with the test data.The outlet temperatures of both calculations were within 2% of each other.The next step is to incorporate this model into the TopOpt framework.

A instituição familiar, ao longo da história, tem passado por transformações significativas, adaptando-se aos valores e crenças de diversas épocas e grupos sociais. Este estudo se concentrou na análise do Contrato de Namoro, um acordo estabelecido entre duas pessoas envolvidas em um relacionamento amoroso, com o propósito de evitar que a relação seja caracterizada como uma União Estável. Esse contrato é formalizado por meio de um registro em cartório, sendo seu principal objetivo, a resolução de questões patrimoniais. Além de abordar as implicações patrimoniais, este trabalho aprofundou-se na definição do namoro e de suas distinções em relação à União Estável, explorando os elementos essenciais presentes nos contratos desse tipo. Destacou-se, também, a importância de compreender as nuances que diferenciam essas modalidades de relacionamento, considerando aspectos legais e sociais. Em síntese, este estudo não apenas se aprofundou na análise do Contrato de Namoro e suas implicações patrimoniais, mas também buscou contextualizar essas práticas dentro do panorama mais amplo das transformações na instituição familiar, considerando tanto as perspectivas legais quanto as sociais.