| 周晓,李永清,张有兵.基于ELM的非侵入式电力负荷识别算法[J].高技术通讯(中文),2020,30(10):1018~1024 |
| 基于ELM的非侵入式电力负荷识别算法 |
|
| |
| DOI：doi：10.3772/j.issn.1002-0470.2020.10.004 |
| 中文关键词: 非侵入式； 负荷识别； 极限学习机(ELM)模型； 事件检测； 累积和(CUSUM) |
| 英文关键词: non-intrusive, load identification, extreme learning machine (ELM) model, event detection, cumulative sum control chart (CUSUM) |
| 基金项目: |
|
| 摘要点击次数: 2423 |
| 全文下载次数: 1720 |
| 中文摘要: |
| 电力负荷识别是需求侧管理的重要环节，为解决传统侵入式负荷监测高成本、不易安装维护的问题，以非侵入式负荷监测为背景研究电力负荷识别算法。从负荷特性出发，针对各电力负荷的暂态及稳态电气特性，提取并建立负荷特征标签。然后，采用极限学习机(ELM)神经网络模型，将输入特征非线性地映射到输入层，实现快速收敛至全局最优点。采用基于累积和(CUSUM)的双边事件检测方法，实现快速准确地检测出负荷投切事件，实时触发负荷识别。最终，以4种常用电力负荷进行实验，结果表明，所提出的负荷识别算法可准确识别出负荷类型，运算效率高，且适用于组合负荷识别。 |
| 英文摘要: |
| Power load identification is an important part of demand-side management. In order to solve the problems of high cost , difficult installation and maintenance of traditional intrusive load monitoring, the non-intrusive load monitoring is used as the background to study the power load identification algorithm. Based on the load characteristics, this paper extracts and establishes the characteristic labels of the loads according to the transient and steady-state electrical characteristics of each power load. Then, the extreme learning machine (ELM) neural network model is used to non-linearly map the input features to the input layer, and the model quickly converges to the global best. The cumulative sum control chart (CUSUM)-based bilateral event detection method is used to quickly and accurately detect load switching events and trigger load identification in real time. Finally, experiments are performed using 4 kinds of commonly used power loads. The results show that the proposed load identification algorithm can accurately identify the type of load. It has high computing efficiency and is suitable for the identification of combined loads. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
|
|
|
