摘要
针对多不良数据辨识中存在的残差污染和残差淹没问题,提出一种多不良数据辨识方法,在应用P-Q分解法的基础上,首先选取部分量测进行状态估计,接着用剩余量测逐一替换参与估计计算的量测,并根据替换后各量测标准化残差的大小得到可疑数据,其间量测量的替换可以打破发生残差淹没的平衡,使得由于发生残差淹没而导致标准化残差合格的不良数据凸显出来,之后又通过状态估计对可疑数据进行校核,恢复受到残差污染的量测为正常量测,最终将不良数据辨识出来。此外,还给出替换和减少一维量测后计算标准化残差的简化方法,以提高计算速度。最后以某地区220kV电网为背景进行算例分析,表明该方法的有效性和可行性。
According to the “residual pollute” and “residual submerge” existing in multiple bad data identification, a method for identifying multiple bad data is presented. Based on the P-Q decoupled method, first, doing state estimation using a part of measurements, then using the remnant measurements replace the measurements which participate in state estimation one by one, and shadiness data are identified after replacement based on the value of standard residual. Meantime, measurement replace can break the balance existing in “residual submerge”, and make the bad data whose standard residual are eligible because of “residual submerge” appear. Then the shadiness data are checked via state estimation, the good data that effected by “residual pollute” can be resumed and all bad data are identified finally. In addition, some simplified formulas for computing standardization residual are given when a measurement is replaced or reduced, which can improve the computation speed. Lastly, taking one area＇s 220 kV power network as background and doing experiment analysis, it is proved that the method is effective and feasible.
出处
《电力系统自动化》
EI
CSCD
北大核心
2007年第13期52-56,62,共6页
Automation of Electric Power Systems
关键词
不良数据辨识
状态估计
量测量替换
残差检测
电力系统
bad data identification
state estimation
measurement replace
residual detection
power system
