摘要
针对高干热环境下智能电表退化趋势难以准确分析的问题,提出基于改进支持向量回归的智能电表测量误差评估模型,并提出优化模拟退火算法实现多个模型参数的自适应调整。首先,提出加权双核函数融合包括时间与环境因子在内的多个输入信息,利用不同应力特征的权值因子表征其对智能电表的影响。在参数设置环节,利用优化模拟退火算法提高参数选择的准确性。结合国网新疆高干热环境试验基地精确度等级为0.5级的智能电表测量误差数据展开实例分析,结果表明,该文模型在准确预测智能电表退化趋势的同时,能评估不同应力特征对其测量误差的影响程度,可以为设备的健康管理提供指导建议。
Considering the problem that the actual degradation trend prediction for smart meter is challenging in the high dry heat environment,a measurement error evaluation model based on the improved support vector regression is presented,and an optimized simulated annealing algorithm(OSA)is proposed to achieve the adaptive setting for multiple model parameters.First,a weighted dual-kernel function is proposed to integrate multiple input information including time and environment factors.The influence of different stress features on the smart meter can be described by the weight factors of different stress features.In the parameter setting stage,the OSA is utilized to improve the accuracy of kernel parameter setting.The case study is carried out using the measurement error data of the smart meter with an accuracy level of 0.5 in the state grid Xinjiang high dry heat test base.The experimental results demonstrate that the proposed model can provide a precise prediction result for degradation trend and evaluate the influence degree of different stress features on the measurement error of equipment,which can provide suggestions for the health management of smart meter.
作者
马俊
唐求
段俊峰
刘颉
韩敏
易珂宇
滕召胜
MA Jun;TANG Qiu;DUAN Junfeng;LIU Jie;HAN Min;YI Keyu;TENG Zhaosheng(College of Electrical and Information Engineering,Hunan University,Changsha 410082,Hunan Province,China)
出处
《中国电机工程学报》
EI
CSCD
北大核心
2023年第12期4581-4588,共8页
Proceedings of the CSEE
基金
国家自然科学基金项目(52077067)
湖南省自然科学基金项目(2021JJ30124)
湖南省研究生科研创新项目(CX20220395)。
关键词
智能电表
测量误差评估
改进支持向量回归
健康管理
smart meter
measurement error evaluation
improved support vector regression
health management
