Comprehensive Life Cycle Assessment of Hydroponic-Aquaculture Systems and Aquaponics Using Brackish and Freshwater Sources: A Comparative Environmental Viability Study
سال انتشار: 1406
نوع سند: مقاله ژورنالی
زبان: انگلیسی
مشاهده: 29
فایل این مقاله در 12 صفحه با فرمت PDF قابل دریافت می باشد
- صدور گواهی نمایه سازی
- من نویسنده این مقاله هستم
استخراج به نرم افزارهای پژوهشی:
شناسه ملی سند علمی:
JR_JREE-14-1_003
تاریخ نمایه سازی: 22 آبان 1405
چکیده مقاله:
Hybrid AC/DC microgrids require fast and reliable fault detection because distributed generation, converter-interfaced sources, sensor disturbances, and operating-mode transitions can reduce the dependability of fixed-threshold and conventional overcurrent protection. This paper presents a focused machine learning framework for high-speed fault detection using transient signal features extracted from voltage, current, DC-link, frequency, power, and state-of-charge measurements. A SCADA-like signal-level dataset was generated from ۹۶۸ operating and disturbance cases sampled at ۱۰ kHz, producing ۴,۸۴۰,۰۰۰ waveform rows and ۹,۶۸۰ window-level feature samples. Eight classes were considered: normal operation, single-line-to-ground fault, line-to-line fault, three-phase fault, harmonic disturbance, sensor anomaly, false-data injection, and islanding transition. The feature pipeline combines statistical descriptors, fast Fourier transform/total harmonic distortion (FFT/THD) quantities, four-level Daubechies-۴ discrete wavelet transform (DWT) energy and entropy features, and Hilbert transform-derived instantaneous amplitude and frequency descriptors. Models were evaluated using a leakage-free, case-wise split with ۶۸۰ training cases, ۱۴۶ validation cases, and ۱۴۲ testing cases. Decision Tree, Random Forest, and support vector machine with a radial basis function kernel (SVM-RBF) achieved ۱۰۰% accuracy and macro-F۱ on the internal test set, while k-nearest neighbors (k-NN) achieved ۹۹.۸۶% accuracy and a ۰.۹۹۵ macro-F۱ score. Ablation analysis showed that statistical, FFT/THD, DWT, and Hilbert descriptors each provide useful information, although statistical features were already highly discriminative under the controlled simulation conditions. External validation on the public GPVS-Faults dataset showed stronger generalization for tree-based models, with Random Forest reaching ۱۰۰% accuracy and Decision Tree reaching ۹۸.۰۹%. The framework is intended as a protection/energy-management-system (EMS) decision-support layer rather than a replacement for certified relay logic.
کلیدواژه ها:
نویسندگان
- -
Centre of Excellence in Water Resources Engineering, University of Engineering and Technology, Lahore ۵۴۸۹۰, Pakistan.
- -
Centre of Excellence in Water Resources Engineering, University of Engineering and Technology, Lahore ۵۴۸۹۰, Pakistan.
- -
College of Engineering, IT and Environment, Charles Darwin University, Darwin, NT ۰۸۱۰, Australia.
- -
Centre of Excellence in Water Resources Engineering, University of Engineering and Technology, Lahore ۵۴۸۹۰, Pakistan.
مراجع و منابع این مقاله:
لیست زیر مراجع و منابع استفاده شده در این مقاله را نمایش می دهد. این مراجع به صورت کاملا ماشینی و بر اساس هوش مصنوعی استخراج شده اند و لذا ممکن است دارای اشکالاتی باشند که به مرور زمان دقت استخراج این محتوا افزایش می یابد. مراجعی که مقالات مربوط به آنها در سیویلیکا نمایه شده و پیدا شده اند، به خود مقاله لینک شده اند :