CAJ | 학술논문

基于变压器时间差定位法的双曲面方程,提出一种用于变压器局部放电源定位的泰勒？遗传综合算法。该算法采用最大似然估计法建立最优方程寻求局部放电源出现最大概率的位置,通过泰勒展开将似然函数线性化,再由遗传算法全局搜索为泰勒展开提供合适的初始点,并利用定位双曲面函数方程个数多于定位所必须的最低个数时出现的冗余数据融合,不断修正定位结果,使得定位寻优过程快速。研究结果表明,冗余信息的增加能有效改善定位精度,利用遗传算法寻找合适的初始点大大提高了泰勒算法的可行性。
기우변압기시간차정위법적쌍곡면방정,제출일충용우변압기국부방전원정위적태륵？유전종합산법。해산법채용최대사연고계법건립최우방정심구국부방전원출현최대개솔적위치,통과태륵전개장사연함수선성화,재유유전산법전국수색위태륵전개제공합괄적초시점,병이용정위쌍곡면함수방정개수다우정위소필수적최저개수시출현적용여수거융합,불단수정정위결과,사득정위심우과정쾌속。연구결과표명,용여신식적증가능유효개선정위정도,이용유전산법심조합괄적초시점대대제고료태륵산법적가행성。
Based on time delay extraction by UHF sensors to locate partial discharge in power transformer, hyperbolic mathematical equations are nonlinear and hard to be solved. In this paper, Taylor-genetic algorithm is proposed to solve the hyperbolic equations, which used the maximum likelihood method to establish optimal equation to seek the greatest probability of PD source. Then made the optimal equation linear by Taylor expansion. Finally, usd genetic global searching to provide the right initial point for Taylor expansion, the location of PD source after some iteration steps was obtained. While the number of hyperboloid equations is larger than the minimum number for PD location, we can correct results by the redundant data fusion. The results show that the increase of redundant information can effectively improve location accuracy, and the use of genetic algorithms to find appropriate initial point greatly enhances the feasibility of the Taylor algorithm, making the searching process rapidly and accurately.