基于随机梯度算法的风电机组变压器漏感辨识

Identification of leakage inductances of power transformer based on stochastic gradient algorithm

变压器会影响风电机组LVRT特性,基于模型的变压器保护原理,需要对变压器绕组参数进行精确辨识。文章利用双绕组变压器的参数辨识方程,应用多新息遗忘随机算法,无需改变三角形侧TA配置,直接利用变压器的电压和电流辨识变压器绕组的漏感,克服了最小二乘法计算量大的局限性,实现对变压器绕组参数的精确辨识,仿真结果表明,该算法能够正确辨识变压器绕组的漏感参数。

The influence of transformer on the LVRT characteristics of the wind turbine can not be ignored. Transformer protection principle based on models needs accurately identify the parameters of the transformer winding, Parameter identification equations of three-phase three-winding transforms are deduced. Multi-innovation stochastic gradient algorithm with forgetting factor is used in the identification of the leakage inductance of transformer and there is no need to change the TA configuration of the triangle side. The method can identify the leakage inductance of the transformer winding by using the voltage and current of the transformer directly. The algorithm overcomes the limitation of the large amount of calculation of the least square method, Simulation experiments results shows the effectiveness of the algorithm.