CAJ | 학술논문

对GIS设置了四种典型缺陷, 用基于动态粒子群算法 ( DPSO) 优化的BP网络对其局放图谱进行识别, 能较高效率地判断出缺陷类型; 用灰度信息图像配准的方法对X射线数字图像进行处理, 可以直观地辨别缺陷形态和位置. 通过实验与应用得出, 先用局放图谱识别出缺陷类型, 再于可能出现故障位置进行X射线检测, 可显著提高诊断的准确度和效率, 并减少X射线单独检测因时间过长对工作人员的辐射伤害.
대GIS설치료사충전형결함, 용기우동태입자군산법 ( DPSO) 우화적BP망락대기국방도보진행식별, 능교고효솔지판단출결함류형; 용회도신식도상배준적방법대X사선수자도상진행처리, 가이직관지변별결함형태화위치. 통과실험여응용득출, 선용국방도보식별출결함류형, 재우가능출현고장위치진행X사선검측, 가현저제고진단적준학도화효솔, 병감소X사선단독검측인시간과장대공작인원적복사상해.
In this paper, one method of DPSO optimized BP network has been applied to recognize partial discharge diagram of GIS four typical defects pre-set. It can judge the defects type more efficiently. X-ray testing technology, based on gray scale image reg-istration can visually identify defect type and location. Through the analysis of the experiment and application, identifying the defects type by partial discharge map firstly, and then testing by X-ray at the possible failure locations, can improve the diagnosis efficiency and accuracy significantly, and reduce the X-ray radiation damage to the staff because of detection alone using too much time.