基于最小二乘法的电力变压器参数辨识算法

Electric Power Transformer Parameter Identification Algorithm Based on Least Square Method

根据变压器的Y/△接线方式,分别列写回路方程,消去三角形侧绕组中的环流和公共磁链的影响.这种方式无需改变三角形侧TA配置,直接利用三角形侧线电流构建了变压器绕组的电阻和电感参数辨识算法,利用带动态遗忘因子的最小二乘法,消除了最小二乘法"数据饱和"的影响.应用EMTP进行了仿真,结果表明,参数辨识算法稳定,不受负荷波动以及功率因数变化的影响,辨识精度较高.

The loop equation is established respctively from the Y/Δ connection mode, and then the circulation current of triangle side and public magnetic chain of both sides is eliminated. The result shows the less need to change the TA configuration of the triangle side. The resistance and induction of transformer＇ s winding parameter identification algorithm can be directly constructed from the line current of the triangle side and the current of the other side. The ＂data saturation＂, got from the least square method （LSM） algorithm, can be eliminated by using the LSM with dynamic forgetting factor. EMTP simulation results show that the parameter identification algorithm is stable, not affected by the load fluctuation＇ s and power factor＇ s changes. This identification result has higher precision.