摘要
为确定用怎样的数据建立起的模型具有较强的泛化能力,对负荷建模中实测数据的选取进行了研究,对不同电压扰动强度的数据进行了辨识,对得到的参数进行了轨迹灵敏度分析,在分析灵敏度结果的基础上提出扰动强度大的数据辨识出的模型的泛化能力较强的观点,并通过小扰动条件下感应电动机的静态特征系数对此结论进行了解释,进一步总结了不同扰动强度数据的大致适用范围,为实测建模中数据的选取提供了借鉴。
In order to know what kind of data can be utilized to derive the model parameters with strong generalization capabilities, the research on field data selection in load modeling is carried out. The model parameters are first derived by identifying the data with different voltage disturbances, then, the trajectory sensitivity analysis is utilized. After the investigation of the parameter sensitivity, it is proposed that the models built from the data with larger disturbances have better generalization capabilities. This viewpoint is also supported by the static characteristic coefficients analysis of induction motor under small disturbances. Finally, the application of the data with different disturbances is summarized which provides the guidance for the data selection in load modeling practice.
出处
《电力系统保护与控制》
EI
CSCD
北大核心
2010年第3期43-47,共5页
Power System Protection and Control
基金
国家自然基金项目(50707009和50595410)
教育部博士学位青年教师基金(20070079014)
北京市科技新星计划
国家重点基础研究发展计划项目(973项目)(2004CB217901)
“111”引智计划(B08013)~~
关键词
负荷建模
实测数据选取
参数辨识
轨迹灵敏度分析
模型泛化
静态特性系数
load modeling
field data selection
parameter identification
trajectory sensitivity analysis
model generalization
static characteristic coefficient
