论文部分内容阅读
变压器铁芯磁化特性的准确建模是研究变压器直流偏磁现象的关键,在使用Jiles-Atherton(J-A)模型对变压器的磁滞回线进行建模分析时,需要对变压器直流偏磁工况下J-A模型中的5个关键参数进行准确识别。提出了人工鱼群与蛙跳混合算法对J-A模型中的关键参数进行辨识,该算法将两种仿生算法有机融合,在鱼群算法寻找到最优区域后切换至蛙跳算法进行局部搜索,兼具了人工鱼群算法前期收敛迅速与蛙跳算法局部搜索准确的优势。分别将所提混合算法及多种现有识别算法应用于数值仿真算例与变压器直流偏磁实测曲线的参数识别,结果表明基于人工鱼群与蛙跳混合算法得到的变压器磁滞回线与实测曲线吻合良好,且具有识别精度高和计算效率高的优点,验证了该算法在变压器J-A模型参数识别中的有效性,进而可以应用于对变压器直流偏磁下运行特性的准确分析。
Accurate modeling of the transformer core magnetization characteristics is the key to study the DC bias phenomenon of the transformer. When using the Jiles-Atherton (JA) model to model and analyze the hysteresis loop of the transformer, Five key parameters in the JA model are accurately identified. A hybrid algorithm of artificial fish swarm and leapfrog was proposed to identify the key parameters in JA model. The algorithm merged the two biomimetic algorithms organically. After the fish swarm algorithm found the optimal region, the algorithm switched to the leapfrog algorithm for local search. Artificial fish swarm algorithm with rapid convergence and frog leaping algorithm local search accurate advantage. The hybrid algorithm and many existing recognition algorithms are respectively applied to the parameter identification of the numerical simulation example and the measured curve of the DC bias of the transformer. The results show that the hysteresis loop of the transformer and the measured The curves have good agreement with each other and have the advantages of high recognition accuracy and high computational efficiency. The algorithm is validated in the parameter identification of transformer JA model and can be applied to the accurate analysis of the operating characteristics under the DC bias of the transformer.