考虑杠杆坏数据的参数辨识和快速修正方法

当存在杠杆坏数据时，现有的参数估计方法精度低、计算复杂，因此，该文提出一种基于拉格朗日乘子法的电网参数辨识和快速修正新方法。首先，建立了基于指数型目标函数的抗差状态估计模型，根据量测的误差大小而自动改变其权重，从而克服大部分坏数据，特别是杠杆坏数据对参数辨识精度的影响；进而基于拉格朗日乘子法确定可疑参数，再基于可疑参数在多个断面下的乘子值计算其用于误差修正的平均乘子值，实现参数误差的快速修正，并减少了计算量。通过3节点简单网络及IEEE 39系统的算例仿真，对比存在杠杆坏数据情况下该方法与传统方法的参数估计精度和计算时间，验证了所提方法的有效性和正确性。

Parameter Identification and Quick Correction Method With Considering Leverage Measurement

Under the existence of bad leverage point the defects in existing parameter estimation methods, such as low accuracy and high computation complexity, will appear. A new Lagrange multiplier based method to identify power grid parameters and rapidly correct them was proposed. Firstly, an exponential objective function based robust state estimation model, which could automatically revise the weights of the measured errors according to the sizes of them, was established to reduce the affects of large portion of bad data, especially the bad leverage point, on parameter identification accuracy. Secondly, based on Lagrange multiplier method the suspicious parameters were be determined, and then according to the multiplier values of suspicious parameters at multi-sections the average multiplier values utilized for error revision were calculated to implement fast revision of parameter errors, meanwhile the calculation amount was decreased. Taking 3-bus simple network and IEEE 39-bus system for example, under the existence of bad leverage point, both parameter estimation accuracy and computation time by the proposed method and traditional method are analyzed, thus the effectiveness and correctness of the proposed method are verified.