欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
提出一种基于似然剖面的电力系统元件暂态模型参数可辨识性方法。可辨识性是参数辨识的前提，可辨识性分析有助于在参数辨识过程中选取合适的扰动数据，并确定可以被有效辨识的参数。参数的可辨识性由模型本身的结构特点和用于辨识的实际扰动深度、实测数据噪声大小共同决定，前者对应理论可辨识性，后者对应实际可辨识性。综合考虑这两种可辨识性，引入统计学中利用似然剖面构造置信区间的方法来进行参数可辨识性分析，通过求解一系列优化问题，得到给定模型中每个参数的置信区间和可辨识性指标。基于此，不仅可以判断参数是否具备理论可辨识性，还能直观地反映各参数的实际可辨识性。
英文摘要：
This paper presented a method of identifiability analysis on transient parameters of electrical power components based on profile likelihood. Identifiability is the premise of parameter identification. Identifiability analysis can help choose data for parameter identification, and find out the parameters that can be effectively identified. The identifiability of parameter is determined by the structure of the models, as well as the severity of the disturbance and the quality of the accuracy of measurements, corresponding to the theoretical and practical identifiability respectively. Considering these two aspects, this paper took advantage of the method of obtaining confidence interval by exploiting profile likelihood to analysis the identifiability of parameters. By solving a series of optimization problem, confidence intervals of each parameter of the models can be established, and the identifiability indices can be derived, based which we can judge the theoretical identifiability and tell how well a parameter can be practically identified with ease.
同期刊论文项目
同项目期刊论文
期刊信息
位置：