Direction-finding Location Algorithm of Cloud Flashes

Cloud lightning location is achieved by excluding solution with large gross errors to optimize initial solution, and joint constrained optimization of weighted integration and Gauss-Newton iterative algorithm based on the multi-station direction-finding cross-algorithm. The lighting position of each group is used as initial positioning solution, which is achieved according to elevation and information of azimuth. Initial solution is optimized through removing the solution with large gross errors by testing function of T-distribution, and then more accurate location information is obtained utilizing the weighted arithmetic. Cloud lightning location information is obtained accurately finally using Gauss-Newton iterative algorithm for constraint calculation. The algorithm is evaluated with the Monte Carlo simulation method, and then the influence of locating result is analyzed. Assuming the error of site layout is 10 m, the error of angle finding is 1°, the position precision is significantly improved using the algorithm of removing gross errors in four-station network simulation. The position precision of three-dimensiond angle of arrival loction (3D-AOA) is higher than integration solution under the same simulation conditions, which shows that the position precision is improved effectively by utilizing the weighted arithmetic and Gauss-Newton iterative algorithm. It shows that the accuracy of position is effectively improved and the deviation of four-station network is less than 500 m when the direction-finding error is 1°, and more stations lead to higher positional precision, but considering the balance of economic cost and precision, four-or five-station network is suggested. As the accuracy of direction-finding increases, the positional precision also increases. Analysis of different station network distributed shows that uniform distributed mode is better than others, the position precision of stations within a station network is clearly higher than stations out of the network. The error symmetry is convenient for analyzing data in practical application. Longer baseline leads to higher positioning accuracy of station network when the station number, station network structure and the direction-finding are fixed. Due to the sensitivity of finding system to the positioning distance, the error curve becomes less symmetrical when the baseline length reaches 100 km. The analysis on different baseline length of the station network positioning accuracy is only the theoretical result in the ideal case, a variety of factors such as instrument performance, detecting network, and hardware testing should be taken into comprehensive consideration in actual application.