FRVM – Grupos de Investigación - GECaM - Artículos

Permanent URI for this collectionhttp://48.217.138.120/handle/20.500.12272/1678

Browse

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • No Thumbnail Available
    Item
    Fault diagnosis strategy for the current source section of a field-cycling nuclear magnetic resonance instrument
    (IOP Publishing Ltd., 2025-09-12) Velez Ibarra, María Delfina; Vodanovic, Gonzalo; Laprovitta, Agustín Miguel; Peretti, Gabriela Marta; Romero, Eduardo Abel; Anoardo, Esteban
    This paper proposes a fault diagnosis strategy to address catastrophic failures in all power components of the current source of a field-cycling nuclear magnetic resonance (FC-NMR) instrument. The current source, implemented with a single power MOSFET operating in linear mode, is prone to thermal instability and degradation under high-current conditions, posing significant risks to system reliability. Due to the continuous conduction inherent in linear-mode operation, fault signatures in the MOSFET could be subtle and difficult to distinguish from normal operational variations, making diagnostic methods relying on switching transients ineffective in this context. To overcome these limitations, an active fault diagnosis framework is introduced to enhance fault detection and localization. This framework combines test signal injection with data-driven artificial intelligence classifiers. Three algorithms—ResNet, a convolutional neural network (CNN), and a nearest neighbor with dynamic time warping (NN-DTW), used as a benchmark—are evaluated using hybrid datasets derived from simulation program with integrated circuit emphasis (SPICE) simulations and experimental fault injections. The methodology employs time-domain signals measured at key circuit nodes, avoiding computationally intensive preprocessing steps. Simulation and experimental results demonstrate classification accuracies of 100% for ResNet and NN-DTW, and 95.2% for CNN, with prediction times under 20 ms for neural networks. The proposal successfully diagnoses both easy-to-detect faults, validated through simulation, and hard-to-detect faults, confirmed experimentally. The entire fault diagnosis process is completed in under 15 s, making it suitable for in-field monitoring of FC-NMR systems."