变压器参数辨识研究及仿真

Study and Simulation on Parameter Identification of Transformer

新型变压器安全保护原理需要对变压器绕组参数进行精确辨识,传统的最小二乘法存在计算量大和收敛慢的问题。在基本微分进化(DE)算法的基础上,通过增加对控制参数F和Cr的调整、增添局部增强算子等改进措施,提出改进微分进化算法,克服了基本微分进化算法控制参数调整困难的不足,提高了算法的全局搜索能力和收敛速度,可以实现对变压器绕组参数的精确辨识。通过matlab/simulink仿真软件对算法进行仿真,表明改进算法能够正确辨识变压器绕组电阻和漏感参数,具有较好的应用前景。

Accurate winding parameter identification is an important requirement in transformer protection. The traditional least-square method has drawbacks of large calculation and slow convergence, and the basic differential evolution algorithm has difficulties to adjust control parameters. A new improved differential evolution algorithm is presented by adding adjustment of control parametersFand Cr, and adding a local enhancement operator. The new al- gorithm can overcome the difficulties of control parameter adjustment, improve the global search ability and conver- gence speed. The new algorithm can realize precise identification of transformer winding parameters. Simulation resuits by using matlab/simulink show that the algorithm can correctly identify leakage inductance and winding resistance parameters of transformer with very small error. x