Scholarlink

<p>Detachable smart systems are contingent on unstable Renewable Energy (RE), which requires efficient planning and control of power supply, storage in user load on a day horizon. The proposed secondary load verification follows the first assessment of RE consumption plans to ensure system-tolerant performance within power quality (PQ) standards. This case study investigates an effective load scheduling in daytime cycles for specific household demands. The initial day-load sequences for the attached equipment are determined according to the RE potential and charge level, ensuring normal operation in the accommodation of user needs. The main motivation is a day-to-day examination of algorithmically composed user load schemes in the overall two-stage strategy maximising the system resource utilisation. Hidden correlations between PQ and load parameters are detected by artificial intelligence (AI) statistics in various operational states and unexpected user activities. Deterministic physical equations cannot capture system behaviour under local RE intermittence within an explicit definition. Unconventional differential learning (DfL) neurocomputing is applied in the evaluation of user demand day-case scenarios for the high dynamical PQ in an experimental RE microgrid. DfL was compared with deep and stochastic learning in the early detection of 24&#xa0;h. PQ load disorders. The modelling times were initialised in day-training intervals of data sampling. Finally, the entire 24&#xa0;h forecast series of related input used in learning is processed to estimate the PQ targets at equal input–output times. The results imply that this procedure can be further elaborated and implemented in an online load self-scheduling and optimisation system. Parametric C + + software including monthly PQ, load, and weather measurements is available in a public repository to allow for additional experiments.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Power quality day-ahead optimisation in smart grid load using PDE component differentiate and deep learning long-term modelling applied to NWP

  • 期刊论文

  • 发表日期：

    2025年12月12日

查看全文
  • Ladislav Zjavka

摘要

Detachable smart systems are contingent on unstable Renewable Energy (RE), which requires efficient planning and control of power supply, storage in user load on a day horizon. The proposed secondary load verification follows the first assessment of RE consumption plans to ensure system-tolerant performance within power quality (PQ) standards. This case study investigates an effective load scheduling in daytime cycles for specific household demands. The initial day-load sequences for the attached equipment are determined according to the RE potential and charge level, ensuring normal operation in the accommodation of user needs. The main motivation is a day-to-day examination of algorithmically composed user load schemes in the overall two-stage strategy maximising the system resource utilisation. Hidden correlations between PQ and load parameters are detected by artificial intelligence (AI) statistics in various operational states and unexpected user activities. Deterministic physical equations cannot capture system behaviour under local RE intermittence within an explicit definition. Unconventional differential learning (DfL) neurocomputing is applied in the evaluation of user demand day-case scenarios for the high dynamical PQ in an experimental RE microgrid. DfL was compared with deep and stochastic learning in the early detection of 24 h. PQ load disorders. The modelling times were initialised in day-training intervals of data sampling. Finally, the entire 24 h forecast series of related input used in learning is processed to estimate the PQ targets at equal input–output times. The results imply that this procedure can be further elaborated and implemented in an online load self-scheduling and optimisation system. Parametric C + + software including monthly PQ, load, and weather measurements is available in a public repository to allow for additional experiments.