Performance Analysis of Smart Grid Fault Diagnosis System Based on Edge Computing Architecture
Zhen Xiang, Liang Zhang, Xiaoyan Guo, Jian Cao, Xuan He, Qiang Zeng
Guangdong Power Grid Company (China)
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
The intelligent grid fault diagnosis faces the bottleneck of high communication delay and insufficient real-time performance, together with the lack of a comprehensive evaluation system that jointly characterizes algorithmic accuracy, system-level responsiveness, and edge-side resource behavior. This study has constructed a three-layer edge computing architecture for fault diagnosis based on cloud-edge collaboration, developed a lightweight deep learning algorithm that integrates structured pruning, INT8 quantization, and knowledge distillation, and established a multi-dimensional evaluation framework covering algorithm performance, system performance, and edge computing characteristics. Experimental results show that the edge-cloud collaboration scheme significantly reduces the end-to-end response delay while maintaining high diagnostic accuracy, and the ablation analysis on the compression pipeline confirms that the lightweight model preserves the diagnostic accuracy to within 0.3 percentage points of its full-precision counterpart at a parameter scale suitable for resource-constrained deployment, with the edge intelligence rate and resource utilization efficiency jointly improved under different load conditions. This study unifies algorithm-level, system-level, and edge-level indicators into a coherent evaluation framework, providing analytical support and practical guidance for the architectural deployment of intelligent grid edge computing systems.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
计算机 / AIIoT and Edge/Fog Computing
Smart Grid and Power Systems · Power Systems Fault Detection