CAJ | 학술논문

本文在阐述变压器保护的重要性和分析励磁涌流产生条件的基础上，提出了一种基于小波神经网络算法识别励磁涌流的方法，大量的仿真结果表明，和应涌流和励磁电流无明显的不同。根据小波神经网络算法庞大的数据量和采样率，提出基于 FPGA 硬件来实现相应的算法。利用小波神经网络和 FPGA 技术，可以提高保护装置的实时处理速度和精度，该研究在进一步提高变压器的保护上具有较高的理论和实用价值。
본문재천술변압기보호적중요성화분석려자용유산생조건적기출상，제출료일충기우소파신경망락산법식별려자용류적방법，대량적방진결과표명，화응용류화려자전류무명현적불동。근거소파신경망락산법방대적수거량화채양솔，제출기우 FPGA 경건래실현상응적산법。이용소파신경망락화 FPGA 기술，가이제고보호장치적실시처리속도화정도，해연구재진일보제고변압기적보호상구유교고적이론화실용개치。
The paper expounds importance of the transformer protection and analyzes the magnetizing inrush current produce condition .A comprehensive prevention based on the wavelet neural network methods for identification the magnetizing inrush current is put forward in this paper .Based on a lot of simulations ,it can be drawn that sympathetic inrush waveform has no obvious difference with that of the inrush current .According to the characteristics of wavelet neural networks’ huge computation and high sampling rate ,a new method based on FPGA of a high-speed hardware platform is proposed to realise the algorithm .Utilizing technologies of wavelet neural networks and FPGA , the accuracy and real-time data processing speed of the protection device can be more effective .In a word ,the research has high theoretical and practical value in the further improvement of transformer protection .