AT 牵引供电系统色散特性计算与行波故障测距算法研究

项目名称： AT 牵引供电系统色散特性计算与行波故障测距算法研究

项目编号： No.51467004

项目类型： 地区科学基金项目

立项/批准年度： 2015

项目学科： 电工技术

项目作者： 陈剑云

作者单位： 华东交通大学

项目金额： 46万元

中文摘要： 对于频变参数的输电线线路，色散现象对行波故障测距精度的影响很大，如果我们预先就掌握了波速和衰减系数随频率变化的特性曲线，反而可以利用色散现象所包含的丰富信息进行更准确的故障定位。基于此设想，我们提出一套新的单端行波故障测距方案：（1）采用地模和线模信号分量的传播速度差原理作为单端检测算法基础；（2）为保证变换精度，模变换矩阵通过特别的迭代算法寻找，不采用如Clarke等简单的固定矩阵；（3）用小波包变换把地模和线模信号分解成多个子频带信号，分别找出其波头到达的时间并计算出估计距离；（4）用贝叶斯网络综合各子频带的测距结果。 研究工作以电气化铁路AT牵引供电系统为对象，重点探讨频变参数的相模变换矩阵生成算法，分析模量的传播特性，建立和训练贝叶斯网络模型。最后进行实地数据检测，用实际数据对新提出的测距算法进行验证。研究的成果不仅可用于电气化铁路牵引供电系统，也适宜于电缆输电线路。

中文关键词： 故障测距；输电线路；小波变换

英文摘要： The precision of traveling wave fault location are influenced mainly by the dispersion phenomena of transmission lines whose parameters vary with the frequency , but we could improve the location accuracy by using the dispersion phenomena which contain a lot of fault messages ,if the characteristics curves of wave propagation velocity and attenuation with the frequency are known in advance. A novel scheme method based on this conceive is proposed : (1) Use single-terminal method based on time delay between aerial mode component and zero component ; (2) The modal transformation matrices are obtained by special iterative algorithm , not using Clarke etc. normal Matrix to ensure the accuracy; (3) Decompose the mode signals into many subband signals using wavelet packets transform, got each arrving time of subband signal wavefront and calculate the estimated distance value respectively; (4 ) Synthesize all estimated distanc value to gain accuracy distance value by Bayesian Network. With the electric railway traction autotransformer feeds system as the research objective, we'll study the phase-mode transformation matrices generated iterative algorithm; analysis the mode signals propagation characteristics; Construct and train the Bayesian Network. The novel scheme will be verify by measurement data from the actual electric railway traction autotransformer feeds system. The research results would not only suit electric railway traction but also suit in case of cable transmission lines.

英文关键词： Fault location;Electric railway traction;Wavelet packet;Dispersion;Bayesian Network