Scholarlink

<p>This paper presents a neuro-symbolic artificial intelligence (NSAI) algorithm that integrates neural networks with symbolic reasoning to optimize microgrid energy management and power quality in distribution systems. The proposed hybrid architecture combines deep reinforcement learning with knowledge-based symbolic rules to address the complex challenges of modern power systems with high renewable penetration. The algorithm was evaluated using MATLAB/Simulink simulations on a modified IEEE 33-bus distribution system under five operational scenarios: normal operation, high renewable penetration (up to 70%), contingency scenarios, islanded operation, and power quality disturbances. Results demonstrate significant improvements compared to conventional methods, including a 22.4% reduction in operational costs, a 17.8% enhancement in energy efficiency, and an improvement in power quality metrics, with a 43.2% reduction in total harmonic distortion (THD). The NSAI algorithm achieved superior performance in contingency management, with faster recovery times for line outages (3.5&#xa0;min vs. 12.8&#xa0;min for rule-based control) and DER failures (2.8&#xa0;min vs. 10.5&#xa0;min), while requiring substantially less load shedding (45.3 kWh vs. 187.6 kWh). The algorithm maintained a 93.7% renewable utilization rate in high renewable penetration scenarios with only 6.3% curtailment. The neuro-symbolic interface enables bidirectional knowledge transfer and provides interpretability through rule extraction, making the approach suitable for real-time implementation with computation times of 0.42&#xa0;s per 5-min control interval. This research contributes to advancing intelligent energy management systems for next-generation distribution networks.</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

Neuro-Symbolic AI Algorithm for Microgrid Energy Management and Power Quality Optimization of Distribution Systems

  • 期刊论文

  • 发表日期：

    2026年3月28日

查看全文
  • Dinesh Kumar Nishad,
  • A. N. Tiwari,
  • Saifullah Khalid

摘要

This paper presents a neuro-symbolic artificial intelligence (NSAI) algorithm that integrates neural networks with symbolic reasoning to optimize microgrid energy management and power quality in distribution systems. The proposed hybrid architecture combines deep reinforcement learning with knowledge-based symbolic rules to address the complex challenges of modern power systems with high renewable penetration. The algorithm was evaluated using MATLAB/Simulink simulations on a modified IEEE 33-bus distribution system under five operational scenarios: normal operation, high renewable penetration (up to 70%), contingency scenarios, islanded operation, and power quality disturbances. Results demonstrate significant improvements compared to conventional methods, including a 22.4% reduction in operational costs, a 17.8% enhancement in energy efficiency, and an improvement in power quality metrics, with a 43.2% reduction in total harmonic distortion (THD). The NSAI algorithm achieved superior performance in contingency management, with faster recovery times for line outages (3.5 min vs. 12.8 min for rule-based control) and DER failures (2.8 min vs. 10.5 min), while requiring substantially less load shedding (45.3 kWh vs. 187.6 kWh). The algorithm maintained a 93.7% renewable utilization rate in high renewable penetration scenarios with only 6.3% curtailment. The neuro-symbolic interface enables bidirectional knowledge transfer and provides interpretability through rule extraction, making the approach suitable for real-time implementation with computation times of 0.42 s per 5-min control interval. This research contributes to advancing intelligent energy management systems for next-generation distribution networks.