International Journal of Mining Science and Technology (Jul 2022)

A novel robust AE/MS source location method using optimized M-estimate consensus sample

  • Yichao Rui,
  • Zilong Zhou,
  • Xin Cai,
  • Riyan Lan,
  • Congcong Zhao

Journal volume & issue
Vol. 32, no. 4
pp. 779 – 791

Abstract

Read online

Due to the complexity of the real engineering environment, the arrival measurement inevitably contains outliers and leads to serious location errors. In order to eliminate the influence of the outliers effectively, this paper proposes a novel robust AE/MS source localization method using optimized M-estimate consensus sample. First, a sample subset is selected from the entire arrival set to obtain fitting model and its parameters. Second, consensus set is determined by checking the arrivals with the fitting model instantiated by the estimated model parameters. Third, optimization process is performed to further optimize the consensus set. The above steps are iterated, and the final source coordinates are obtained by using all the elements in the optimal consensus set. The novel method is validated by a pencil-lead breaks experiment. The results indicate that the novel method has better location accuracy of less than 5 mm compared to existing methods, regardless of the presence or absence of outliers. With the increase of outlier scale and outlier ratio, the location result of the proposed method is always more stable and accurate than that of the existing methods. Mine blasting experiments further demonstrate that the new method holds good prospects for engineering applications.

Keywords