摘要
在非侵入式负荷识别任务中,随着家用电器类型的不断增加,功率差距不大但V-I轨迹相似的设备很容易被分类错误。针对这些问题,本研究提出了基于颜色编码和残差神经网络的非侵入式负荷识别方法。首先,对采集到的高频电压、电流数据进行预处理;然后,再通过二值轨迹映射和HSV颜色编码将V-I轨迹转换为视觉表示,不仅在V-I轨迹中融入了丰富的电气特征,还增强了负荷特征的唯一性;最后利用PLAID公共数据集对本研究所提方法进行了验证。结果表明,本研究所提方法显著提高了识别准确率,并能够有效区分各个电器设备。
In the non-intrusive load identification task,as the types of household appliances continue to increase,devices with small power differences but similar V-I trajectories can easily be misclassified.In response to these problems,this paper proposes a non-intrusive load identification method based on color coding and residual neural network.The collected high-frequency voltage and current data were preprocessed,and then the V-I trajectory was converted into a visual representation through binary trajectory mapping and HSV color coding.This not only incorporated rich electrical features into the V-I trajectory,but also enhanced the uniqueness of load characteristics.The method was finally verified using the PLAID public data set.The results show that the method proposed in this article significantly improves the recognition accuracy and can effectively distinguish individual electrical equipments.
作者
杨苗
游文霞
刘玥
汪芯茜
YANG Miao;YOU Wenxia;LIU Yue;WANG Xinqian(College of Electrical Engineering and New Energy,China Three Gorges University,Yichang 443002,China)
出处
《电工材料》
CAS
2024年第2期94-99,共6页
Electrical Engineering Materials
基金
国网湖北省电力有限公司科技项目(SDHZ2019251)--基于大数据的配电网规划决策系统研究、开发及应用。
