中文摘要：

基于模糊理论的变压器励磁涌流识别方法面临判据重要度难以确定的问题，通常情况下依赖于专家经验的做法很难客观反映变压器保护各判据之间的相互作用关系。建立了基于Choquet积分的k可加模糊测度的计算模型，k可加模糊测度在测度的复杂性和表示能力方面做了折中，可以以较小的计算量获得能够反映判据关系的重要性测度。仿真了变压器励磁涌流和内部故障等运行状态，采用二次谐波等三个判据和得到的重要性测度实现了变压器励磁涌流模糊识别方法。仿真分析结果表明该方法得到的判据重要度能够全面反映判据之间的相互作用关系，在此基础上的励磁涌流模糊识别方法可以综合各判据的优势，改善了变压器保护的性能。

英文摘要：

The transformer inrush current identification based on fuzzy theory faces a very difficult problem that how to get the importance of the criterions. It is difficult for those methods depending on expert experience to reflect the interaction relations among the criterions of transformer protection objectively. In the paper a nonlinear model based on Choquet fuzzy integral for k-additive fuzzy measures calculation is presented, k-order additive fuzzy measures make a compromise between the complexity and expressing ability, based on which the importance of the criterions can be obtained with less computation. The inrush currents and internal faults of transformer are simulated, and three criterions including second harmonic, waveform symmetry and terminal voltage are adopted to identify inrush current. Simulation results validate that the importance of the criterions obtained by the presented method can reflect the relations among the criterions fully, based on which the inrush current fuzzy identification can integrate the advantages of the various criteria and improve the performance of transformer protection.