基于L-M算法和声波融合的TDOA雷电定位改进方法

An Improved TDOA Lightning Location Approach Considering L-M Algorithm and Acoustics

在时差信息冗余情况下求解非线性方程时,采用到达时差定位法(TDOA)的雷电定位系统会出现求解发散问题,而我国油罐系统等雷电敏感系统多经历了雷声和电磁探测的雷电定位技术发展过程.对此,提出一种有效融合雷电声波信号和放电电磁信号的TDOA雷电定位改进方法.根据已有站点信息将目标区域划分为16个子区域,在子区域内利用L-M迭代算法实现监测目标区域的雷电定位,提升TDOA雷电定位系统的定位精度和抗误差扰动能力.结果表明:传统法的雷电定位误差均值为203.2 m;所提方法经过声波信息融合和L-M算法改进后,雷电定位误差均值减小为108.4 m,监测目标区域边缘雷电定位精度提高了51.2%.研究对改进已有雷电定位系统和有效利用声波信息具有实际意义.

In the conventional lightning location system(LLS)based on time difference of arrival(TDOA),the nonlinear equations in lightning location calculation easily get to be divergent when the time information acquired from electromagnetic sensors is redundant.The LLS setup in lightning-sensitive regions in China usually experiences a development from detecting the thunderclap signal to the electromagnetic signal,such as the LLSs in oil tank farms.Therefore,an improved TDOA lightning location approach was proposed considering the acoustic and electromagnetic information emitted from lightning discharges.The targeted lightning monitoring region was divided into 16 sub-regions according to the location of the existing detection stations.The lightning location was calculated based on the Levenberg-Marquardt(L-M)iterative algorithm,which improves the lightning location accuracy and the resistance ability to measuring errors.The results show that the average error of the traditional lightning location method is 203.2 m.In contrast,the proposed approach can reduce the lightning location error to 108.4 m by considering the acoustic information and L-M iteration algorithm.The location accuracy at the edge of the targeted area is improved by 51.2%.This research can be potentially counseled in the improvement of existing LLSs and making an effective use of acoustic information.