A New Acoustic Emission Source Location Method Using Tri-Variate Kernel Density Estimator

Abstract

A new AE location method using tri-variate kernel density estimator is developed in this paper. Firstly, combinations of every six arrivals are obtained from a multi-sensor location system, and the preliminary location results are obtained by solving the systems of linear equations constructed by these arrival combinations. Secondly, the tri-variate scaled kernel functions at each AE source coordinate are established. The tri-variate kernel density estimator is constructed by adding and normalizing these scaled kernel functions. Finally, the extreme value of the density function is calculated and the coordinate corresponding to the extreme value is extracted as the final location result. Pencil-lead break experiments were carried out. The results verified that the proposed method was more accurate and effective than traditional methods in the location performance. Moreover, the influence of outlier scales and proportions of the proposed method were investigated by simulation tests. Results showed that the location performance of the proposed method was higher than that of traditional methods under different outlier scales and proportions.

Keywords

• Acoustic emission (AE)
• distribution model
• Kernel density estimator
• outlier
• source location