A-RQEA与TSVD并行求解的电弧电流重构反演方法  

Inversion Method of Arc Current Reconstruction Combining A-RQEA with TSVD

作　　者：赵宏晨 刘晓明[1,2] 钱小毅 ZHAO Hongchen;LIU Xiaoming;QIAN Xiaoyi(School of Electrical Engineering,Shenyang University of Technology,Shenyang 110870,China;School of Electrical Engineering,Tianjin Polytechnic University,Tianjin 300387,China)

机构地区：[1]沈阳工业大学电气工程学院,沈阳110870 [2]天津工业大学电气工程与自动化学院,天津300387

出　　处：《高电压技术》2018年第12期4020-4028,共9页High Voltage Engineering

基　　金：国家自然科学基金(51377106)~~

摘　　要：开关电器设备中电弧理论研究是输变电装备领域的关键问题,对此引入数学反演方法重构电弧形态和电流密度。根据电磁正演分析建立算子方程并离散化为代数方程组。为了克服电磁逆问题的不适定性,采用基于等位基因的实数编码量子进化算法与正则化理论并行处理的手段。通过高效精确的智能优化算法重建电弧位置和形态,再利用截断奇异值分解反演电弧电流分布,并利用L曲线准则选取正则化参数。结果表明:将非线性寻优与正则化反演结合的手段,通过磁场分布重构电弧形态和电流分布状况具有较高的精度;截断奇异值分解法(TSVD)计算电弧各条折线电流值相对误差最大为8%,最小为0.03%,反演弧芯和边缘区域电流密度符合预设分布趋势,有助于理解开关电弧的复杂物理特性。The research on the theory of arc in switchgear is the key problem in thefield of power transmission equipment. We introduce a mathematical inverse method to reconstruct the arc shape and current density.The operator equation is established and discretized into algebraic equations according to the electromagnetic forward analysis.To overcome the ill posedness of the electromagnetic inverse problem,the real-coded quantum evolutionary algorithm based on allele(A-RQEA)along with the parallel processing of regularization theory is adopted.Both the arc position and the arc shape are reconstructed by the efficient and accurate intelligent optimization,algorithm,with the current distribution inversed by means of truncated singular value decomposition(TSVD).The L-curve criterion is used to select the regularization parameter.The results show that,by means of combining the nonlinear optimization with the regularization inversion,the arc shape and the current distribution can be reconstructed via magnetic field distribution accurately.The maximum relative error of the calculated current value for each arc current line by means of TSVD is 8%,and the minimum relative error is 0.03%,the inversed current density from the core and the edge regions is consistent with the trend of preset distribution,which will help understand the complex physical characteristics for the switching arc.

关 键 词：电弧反演问题 不适定性 实数编码量子进化算法 截断奇异值分解 正则化参数 

分 类 号：TM501.2[电气工程—电器]