指数型目标函数电力系统抗差状态估计的解法与性能分析  

Solution and Performance Analysis to a Robust State Estimation Method With Exponential Objective Function

作　　者：郭烨[1] 张伯明[1] 吴文传[1] 

机构地区：[1]电力系统及发电设备控制和仿真国家重点实验室(清华大学电机系),北京市海淀区100084

出　　处：《中国电机工程学报》2011年第7期89-95,共7页Proceedings of the CSEE

基　　金：国家自然科学基金项目(50707013)~~

摘　　要：针对一个指数型目标函数电力系统抗差状态估计模型,给出一种有效的求解方法。由于指数型目标函数电力系统抗差状态估计模型的目标函数连续可微,因此可通过与加权最小二乘估计类似的牛顿法方便地进行求解,与现有状态估计程序之间具有很好的兼容性。理论分析表明新的抗差估计模型具有很强的排除不良杠杆量测的性能,并通过与加权最小二乘估计和广义M估计的数值算例比较,进行了验证。基于IEEE标准系统和2个实际省级电网的计算分析,将本文方法与现场广泛应用的含不良数据辨识环节的加权最小二乘估计程序进行比较,结果表明所提方法的抗差能力具有明显的优势。An efficient method to solve the proposed robust power system state estimation model with an exponential objective function was presented. The objective function of the estimation model is differentiable, so solution based on Newton method can be used. As its solution is similar with weighted least square estimator, the new method has good compatibility with traditional methods. The theoretical analysis shows that the proposed method has powerful ability to compress the effects of bad leverage measurements, which is also verified by comparing this method with weighted least square estimator and generalized M-estimator. IEEE standard test systems and two provincial power grids were used to compare the performance of proposed method with a widely used weighted least square estimator with a bad data identification module. The results show that the proposed method is much more robust in identifying bad data.

关 键 词：电力系统 状态估计 不良数据辨识 抗差估计 杠杆量测 

分 类 号：TM74[电气工程—电力系统及自动化]